3-Inductive-Rewiring-CTLayer.ipynb

{
  "cells": [
    {
      "attachments": {},
      "cell_type": "markdown",
      "id": "6d07b1b0",
      "metadata": {
        "id": "6d07b1b0"
      },
      "source": [
        "[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/ellisalicante/GraphRewiring-Tutorial/blob/main/3-Inductive-Rewiring-CTLayer.ipynb)\n",
        "# Inductive rewiring using CT-Layer\n",
        "Adrián Arnaiz Rodríguez, ELLIS Alicante UNIT\n",
        "\n",
        "Francisco Escolano, ELLIS Alicante UNIT\n",
        "\n",
        "Ahmed Begga, University of Valencia\n",
        "\n",
        "***Tutorial on Graph Rewiring: From Theory to Applications in Fairness***\n",
        "\n",
        "[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/ellisalicante/GraphRewiring-Tutorial/blob/main/3-Inductive-Rewiring-CTLayer.ipynb)"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "GHIUBS2KwVGH",
      "metadata": {
        "id": "GHIUBS2KwVGH"
      },
      "source": [
        "## Environment preparation"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "id": "090e809d",
      "metadata": {
        "id": "090e809d"
      },
      "outputs": [],
      "source": [
        "COLLAB_ENV = True\n",
        "if COLLAB_ENV:\n",
        "    !pip install 'scipy>=1.8'"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "id": "34c457a3",
      "metadata": {
        "id": "34c457a3"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "import torch"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "id": "f2befe8a",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "f2befe8a",
        "outputId": "65112249-8225-4b40-aa73-9c5f298617e5"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Cloning into 'GraphRewiring-Tutorial'...\n",
            "remote: Enumerating objects: 32, done.\u001b[K\n",
            "remote: Counting objects: 100% (32/32), done.\u001b[K\n",
            "remote: Compressing objects: 100% (24/24), done.\u001b[K\n",
            "remote: Total 32 (delta 11), reused 23 (delta 7), pack-reused 0\u001b[K\n",
            "Unpacking objects: 100% (32/32), done.\n",
            "Submodule 'DiffWire' (https://github.com/AdrianArnaiz/DiffWire/) registered for path 'DiffWire'\n",
            "Cloning into '/content/GraphRewiring-Tutorial/DiffWire'...\n",
            "Submodule path 'DiffWire': checked out '5aded812451639187680c2d70bf16f4d2f21ca2e'\n",
            "mv: cannot move 'GraphRewiring-Tutorial/DiffWire' to './DiffWire': Directory not empty\n",
            "mv: cannot move 'GraphRewiring-Tutorial/figs' to './figs': Directory not empty\n",
            "1.12.1+cu113\n"
          ]
        }
      ],
      "source": [
        "if COLLAB_ENV:\n",
        "  !git clone https://github.com/ellisalicante/GraphRewiring-Tutorial\n",
        "  !cd GraphRewiring-Tutorial && git submodule update --init --recursive\n",
        "  !mv GraphRewiring-Tutorial/* ./\n",
        "  !rm -rf GraphRewiring-Tutorial\n",
        "    \n",
        "  os.environ['TORCH'] = torch.__version__\n",
        "  print(torch.__version__)\n",
        "  !pip install -q torch-scatter -f https://data.pyg.org/whl/torch-${TORCH}.html\n",
        "  !pip install -q torch-sparse -f https://data.pyg.org/whl/torch-${TORCH}.html\n",
        "  !pip install -q git+https://github.com/pyg-team/pytorch_geometric.git\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "id": "MpafaIKspIgO",
      "metadata": {
        "id": "MpafaIKspIgO"
      },
      "outputs": [],
      "source": [
        "import sys\n",
        "sys.path.append(\"./DiffWire\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "id": "uoZK6-FgiUar",
      "metadata": {
        "id": "uoZK6-FgiUar"
      },
      "outputs": [],
      "source": [
        "from DiffWire.transforms import FeatureDegree\n",
        "from DiffWire.layers.utils.ein_utils import _rank3_diag, _rank3_trace\n",
        "from DiffWire.layers.MinCut_Layer import dense_mincut_pool\n",
        "\n",
        "import torch.nn.functional as F\n",
        "from torch.nn import Linear\n",
        "\n",
        "from torch_geometric.loader import DataLoader\n",
        "from torch_geometric.datasets import TUDataset\n",
        "from torch_geometric.nn import DenseGraphConv\n",
        "from torch_geometric.utils import to_dense_batch, to_dense_adj\n",
        "from torch_geometric.utils import to_networkx\n",
        "\n",
        "import networkx as nx\n",
        "\n",
        "import numpy as np\n",
        "\n",
        "from scipy.linalg import fractional_matrix_power\n",
        "from scipy.spatial.distance import pdist\n",
        "from scipy.spatial.distance import squareform\n",
        "\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.manifold import TSNE\n",
        "from sklearn.preprocessing import normalize\n",
        "\n",
        "\n",
        "import matplotlib.pyplot as plt"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "id": "d16edf02",
      "metadata": {
        "id": "d16edf02"
      },
      "outputs": [],
      "source": [
        "import warnings\n",
        "warnings.filterwarnings(\"ignore\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "id": "67ccdbfc",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "67ccdbfc",
        "outputId": "7dd3170c-aef5-49f1-85e5-30857a9ee834"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "(device(type='cuda'), True)"
            ]
          },
          "execution_count": 8,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "device=torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
        "device, torch.cuda.is_available()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "id": "yfrlFryzmQPj",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "yfrlFryzmQPj",
        "outputId": "01f385ea-b2af-4db7-b91f-dfd26269f6f7"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Wed Dec  7 15:32:24 2022       \n",
            "+-----------------------------------------------------------------------------+\n",
            "| NVIDIA-SMI 460.32.03    Driver Version: 460.32.03    CUDA Version: 11.2     |\n",
            "|-------------------------------+----------------------+----------------------+\n",
            "| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\n",
            "| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |\n",
            "|                               |                      |               MIG M. |\n",
            "|===============================+======================+======================|\n",
            "|   0  Tesla T4            Off  | 00000000:00:04.0 Off |                    0 |\n",
            "| N/A   49C    P0    26W /  70W |      3MiB / 15109MiB |      0%      Default |\n",
            "|                               |                      |                  N/A |\n",
            "+-------------------------------+----------------------+----------------------+\n",
            "                                                                               \n",
            "+-----------------------------------------------------------------------------+\n",
            "| Processes:                                                                  |\n",
            "|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |\n",
            "|        ID   ID                                                   Usage      |\n",
            "|=============================================================================|\n",
            "|  No running processes found                                                 |\n",
            "+-----------------------------------------------------------------------------+\n"
          ]
        }
      ],
      "source": [
        "!nvidia-smi"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "0b3dde99",
      "metadata": {
        "id": "0b3dde99"
      },
      "source": [
        "## Graph Classification with CT-Layer"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "aa86d2bc",
      "metadata": {
        "id": "aa86d2bc"
      },
      "source": [
        "### Train with CT-Layer"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "4bcbd767",
      "metadata": {
        "id": "4bcbd767"
      },
      "source": [
        "#### CT-Layer\n",
        "\n",
        "For using it straightforward from the **[DiffWire repository](https://github.com/AdrianArnaiz/DiffWire)**:\n",
        "```python\n",
        "from DiffWire.layers.CT_layer import dense_CT_rewiring\n",
        "```\n",
        "**However, for the sake of clarity of the tutorial, we will explain the content of that function line by line**\n",
        "\n",
        "<a href =\"https://paperswithcode.com/method/ct-layer\"> <img src=\"https://production-media.paperswithcode.com/methods/305a898a-e0a2-4d74-b8e8-c12839496577.png\" alt=\"CT Layer\" style=\"width:500px;\"/> </a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "id": "734c29cf",
      "metadata": {
        "id": "734c29cf"
      },
      "outputs": [],
      "source": [
        "def dense_CT_rewiring(x, adj, s, mask=None, EPS=1e-15):\n",
        "    \"\"\"Rewires a Graph using CT Distance (Effective Resistances) given 's' as the CT Embedding. \n",
        "    Returns the new adjacency, and the loss for the CT Embbeding (s).\n",
        "\n",
        "    Args:\n",
        "        x (dense): feature matrix: NxF\n",
        "        adj (dense): dense adjacency matrix: NxN\n",
        "        s (dense): CT Embedding: NxH (H: size of latent space)\n",
        "        mask (dense): dense mask of batches\n",
        "        EPS (float): epsilon to avoid nans\n",
        "\n",
        "    Returns:\n",
        "        adj: new adjacency = CTdist/vol(G)\n",
        "        loss: Cut Loss for CT Embedding (s)\n",
        "        ortho_loss: Loss regularization orthogonality in CT Embedding (s)\n",
        "    \"\"\"\n",
        "    x = x.unsqueeze(0) if x.dim() == 2 else x # adj torch.Size([b, N, f])\n",
        "    adj = adj.unsqueeze(0) if adj.dim() == 2 else adj # adj torch.Size([b, N, N]) \n",
        "    s = s.unsqueeze(0) if s.dim() == 2 else s # s torch.Size([b, N, k])\n",
        "    \n",
        "    s = torch.tanh(s) # torch.Size([20, N, k]) One k for each N of each graph\n",
        "    \n",
        "    # batck masking\n",
        "    (batch_size, num_nodes, _), k = x.size(), s.size(-1)\n",
        "    if mask is not None:\n",
        "        mask = mask.view(batch_size, num_nodes, 1).to(x.dtype)\n",
        "        x, s = x * mask, s * mask \n",
        "\n",
        "    \n",
        "    # Calculate CT_dist (distance matrix)\n",
        "    CT_dist = torch.cdist(s,s) # [20, N, k], [20, N, k]-> [20,N,N]\n",
        "    ## Calculate degree d_flat and degree matrix d\n",
        "    d_flat = torch.einsum('ijk->ij', adj) # torch.Size([b, N]) \n",
        "    d = _rank3_diag(d_flat)+EPS  # d torch.Size([b, N, N])\n",
        "    ## Calculate Vol (volumes): one per graph \n",
        "    vol = _rank3_trace(d) # torch.Size([20]) \n",
        "    ## Calculate out_adj as CT_dist/vol(G)\n",
        "    N = adj.size(1)\n",
        "    CT_dist = (CT_dist) / vol.unsqueeze(1).unsqueeze(1)\n",
        "    ## Mask with adjacency\n",
        "    adj = CT_dist*adj\n",
        "    \n",
        "    \n",
        "    # Losses\n",
        "    ## Calculate Laplacian L = D - A \n",
        "    L = d - adj\n",
        "    ## Calculate out_adj as A_CT = S.T*L*S\n",
        "    out_adj = torch.matmul(torch.matmul(s.transpose(1, 2), L), s) #[b, k, N]*[b, N, N]-> [b, k ,N]*[b, N, k] = [20, k, k]\n",
        "    ## Calculate CT_num \n",
        "    CT_num = _rank3_trace(out_adj) # mincut_num torch.Size([b]) one sum over each graph\n",
        "    ## Calculate CT_den \n",
        "    CT_den = _rank3_trace(\n",
        "        torch.matmul(torch.matmul(s.transpose(1, 2), d ), s))+EPS # [b, k, N]*[b, N, N]->[b, k, N]*[b, N, k] -> [b] one sum over each graph\n",
        "\n",
        "    CT_loss = CT_num / CT_den\n",
        "    CT_loss = torch.mean(CT_loss) # Mean over batch!\n",
        "    \n",
        "    ## Orthogonality regularization.\n",
        "    ss = torch.matmul(s.transpose(1, 2), s)  #[b, k, N]*[b, N, k]-> [b, k, k]\n",
        "    i_s = torch.eye(k).type_as(ss) # [k, k]\n",
        "    ortho_loss = torch.norm(\n",
        "        ss / torch.norm(ss, dim=(-1, -2), keepdim=True) -\n",
        "        i_s)\n",
        "    ortho_loss = torch.mean(ortho_loss) # Mean over batch!\n",
        "    \n",
        "    return adj, CT_loss, ortho_loss"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "86e51e58",
      "metadata": {
        "id": "86e51e58"
      },
      "source": [
        "#### GNN using CT-Layer\n",
        "\n",
        "<img src=\"https://github.com/ellisalicante/GraphRewiring-Tutorial/blob/main/figs/ctnetwork.png?raw=1\" alt=\"CT network\" style=\"width:300px;\"/> </a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "1f656163",
      "metadata": {
        "id": "1f656163"
      },
      "outputs": [],
      "source": [
        "class CTNet(torch.nn.Module):\n",
        "    def __init__(self, in_channels, out_channels, k_centers, hidden_channels=32, EPS=1e-15):\n",
        "        super(CTNet, self).__init__()\n",
        "        \n",
        "        self.EPS=EPS\n",
        "        #Message Passing Layers\n",
        "        self.conv1 = DenseGraphConv(hidden_channels, hidden_channels)\n",
        "        self.conv2 = DenseGraphConv(hidden_channels, hidden_channels)\n",
        "        \n",
        "        # Pooling for CT embedding\n",
        "        num_of_centers1 =  k_centers # k1 #order of number of nodes\n",
        "        self.pool1 = Linear(hidden_channels, num_of_centers1)\n",
        "        \n",
        "        # Pooling for MinCut Layer\n",
        "        num_of_centers2 =  16 # k2 #mincut \n",
        "        self.pool2 = Linear(hidden_channels, num_of_centers2) \n",
        "\n",
        "        # MLPs towards out \n",
        "        self.lin1 = Linear(in_channels, hidden_channels)\n",
        "        self.lin2 = Linear(hidden_channels, hidden_channels)\n",
        "        self.lin3 = Linear(hidden_channels, out_channels)\n",
        " \n",
        "\n",
        "    def forward(self, x, edge_index, batch):    # x torch.Size([N, N]),  data.batch  torch.Size([661])  \n",
        "        # Make all matrices dense\n",
        "        adj = to_dense_adj(edge_index, batch)   # adj torch.Size(B, N, N])\n",
        "        x, mask = to_dense_batch(x, batch)      \n",
        "\n",
        "        #First layer: linear MLP\n",
        "        x = self.lin1(x) \n",
        "        \n",
        "        if torch.isnan(adj).any():\n",
        "              print(\"adj nan\")\n",
        "        if torch.isnan(x).any():\n",
        "              print(\"x nan\")\n",
        "        \n",
        "        # CT REWIRING\n",
        "        s1  = self.pool1(x)\n",
        "        #adj = torch.Size([b, N, h]) --> CT Embedding\n",
        "        adj, CT_loss, ortho_loss1 = dense_CT_rewiring(x, adj, s1, mask, EPS = self.EPS) \n",
        "        #adj = torch.Size([b, N, N]) --> CT Distances\n",
        "        \n",
        "\n",
        "        # CONV1: Now on x and rewired adj: \n",
        "        x = self.conv1(x, adj) #out: x torch.Size([20, N, F'=32])\n",
        "\n",
        "        # MINCUT_POOL - Garph pooling\n",
        "        # MLP of k=16 outputs s\n",
        "        s2 = self.pool2(x) # s torch.Size([20, N, k])\n",
        "        \n",
        "        # Call to dense_cut_mincut_pool to get coarsened x, adj and the losses: k=16\n",
        "        x, adj, mincut_loss2, ortho_loss2 = dense_mincut_pool(x, adj, s2, mask, EPS=self.EPS) # out x torch.Size([20, k=16, F'=32]),  adj torch.Size([20, k2=16, k2=16])\n",
        "\n",
        "        # CONV2: Now on coarsened x and adj: \n",
        "        x = self.conv2(x, adj) #out x torch.Size([20, 16, 32])\n",
        "        \n",
        "        # Readout for each of the 20 graphs\n",
        "        x = x.sum(dim=1) \n",
        "        \n",
        "        # Final MLP for graph classification: hidden channels = 32\n",
        "        x = F.relu(self.lin2(x)) \n",
        "        x = self.lin3(x) \n",
        "        \n",
        "        #loss functions\n",
        "        CT_loss = CT_loss + ortho_loss1\n",
        "        mincut_loss = mincut_loss2 + ortho_loss2\n",
        "        \n",
        "        return F.log_softmax(x, dim=-1), CT_loss, mincut_loss\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 25,
      "id": "0a8d6f9f",
      "metadata": {
        "id": "0a8d6f9f"
      },
      "outputs": [],
      "source": [
        "def train(epoch, loader):\n",
        "    model.train()\n",
        "    loss_all = 0\n",
        "    correct = 0\n",
        "    #i = 0\n",
        "    for data in loader:\n",
        "        data = data.to(device)\n",
        "        optimizer.zero_grad()\n",
        "        out, mc_loss, o_loss = model(data.x, data.edge_index, data.batch) # data.batch  torch.Size([783])\n",
        "        loss = F.nll_loss(out, data.y.view(-1)) + mc_loss + o_loss\n",
        "        loss.backward()\n",
        "        loss_all += data.y.size(0) * loss.item()\n",
        "        optimizer.step()\n",
        "        correct += out.max(dim=1)[1].eq(data.y.view(-1)).sum().item() #accuracy in train AFTER EACH BACH\n",
        "    return loss_all / len(loader.dataset), correct / len(loader.dataset)\n",
        "\n",
        "@torch.no_grad()\n",
        "def test(loader):\n",
        "    model.eval()\n",
        "    correct = 0\n",
        "    for data in loader:\n",
        "        data = data.to(device)\n",
        "        pred, mc_loss, o_loss = model(data.x, data.edge_index, data.batch)\n",
        "        loss = F.nll_loss(pred, data.y.view(-1)) + mc_loss + o_loss\n",
        "        correct += pred.max(dim=1)[1].eq(data.y.view(-1)).sum().item()\n",
        "\n",
        "    return loss, correct / len(loader.dataset)"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "ykl_l7weqYjJ",
      "metadata": {
        "id": "ykl_l7weqYjJ"
      },
      "source": [
        "#### Load Dataset and split\n",
        "\n",
        "**Change `batch_size` depending on the availability of you memory**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "8451e83e",
      "metadata": {
        "id": "8451e83e"
      },
      "outputs": [],
      "source": [
        "#TRAIN_SPLIT = 1500\n",
        "#BATCH_SIZE = 16\n",
        "#num_of_centers = 100\n",
        "#dataset = TUDataset(root='data',name=\"REDDIT-BINARY\", transform = FeatureDegree())\n",
        "\n",
        "#TRAIN_SPLIT = 4500\n",
        "#BATCH_SIZE = 16\n",
        "#num_of_centers = 50\n",
        "#dataset = TUDataset(root='data',name=\"COLLAB\", transform = FeatureDegree())\n",
        "\n",
        "TRAIN_SPLIT = 800\n",
        "BATCH_SIZE = 16\n",
        "num_of_centers = 20\n",
        "dataset = TUDataset(root='data',name=\"IMDB-BINARY\", transform = FeatureDegree())"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "2287d420",
      "metadata": {
        "id": "2287d420"
      },
      "outputs": [],
      "source": [
        "seed = 12345\n",
        "torch.manual_seed(12345)\n",
        "stratified = True\n",
        "\n",
        "if stratified:\n",
        "  train_indices, test_indices = train_test_split(list(range(len(dataset.data.y))), test_size=0.15, \n",
        "                                                stratify=dataset.data.y,\n",
        "                                                random_state=seed, shuffle=True)\n",
        "\n",
        "  train_dataset = torch.utils.data.Subset(dataset, train_indices)\n",
        "  test_dataset = torch.utils.data.Subset(dataset, test_indices)\n",
        "\n",
        "  train_loader = DataLoader(train_dataset, batch_size=BATCH_SIZE, shuffle=True)\n",
        "  test_loader = DataLoader(test_dataset, batch_size=BATCH_SIZE, shuffle=False)  \n",
        "else:\n",
        "  dataset = dataset.shuffle()\n",
        "  train_dataset = dataset[:TRAIN_SPLIT]\n",
        "  test_dataset = dataset[TRAIN_SPLIT:]\n",
        "  train_loader = DataLoader(train_dataset, batch_size=BATCH_SIZE, shuffle=True)\n",
        "  test_loader = DataLoader(test_dataset, batch_size=BATCH_SIZE, shuffle=False)"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "MdmyLDrNqobB",
      "metadata": {
        "id": "MdmyLDrNqobB"
      },
      "source": [
        "#### Create model and train"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "4965faca",
      "metadata": {
        "id": "4965faca"
      },
      "outputs": [],
      "source": [
        "EPS = 1e-15\n",
        "epochs = 60\n",
        "\n",
        "model = CTNet(dataset.num_features, dataset.num_classes, k_centers=num_of_centers, EPS=EPS).to(device)\n",
        "optimizer = torch.optim.Adam(model.parameters(), lr=5e-4, weight_decay=1e-4)\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "b739a0cd",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "b739a0cd",
        "outputId": "5ebb139d-d7f3-4de8-efc5-d91f32c9f2c4"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 001, Train Acc: 0.5059, Test Acc: 0.4600\n",
            "Epoch: 002, Train Acc: 0.5682, Test Acc: 0.4200\n",
            "Epoch: 003, Train Acc: 0.5906, Test Acc: 0.5467\n",
            "Epoch: 004, Train Acc: 0.6082, Test Acc: 0.5733\n",
            "Epoch: 005, Train Acc: 0.5871, Test Acc: 0.6400\n",
            "Epoch: 006, Train Acc: 0.6282, Test Acc: 0.6333\n",
            "Epoch: 007, Train Acc: 0.6306, Test Acc: 0.4733\n",
            "Epoch: 008, Train Acc: 0.6082, Test Acc: 0.4600\n",
            "Epoch: 009, Train Acc: 0.6341, Test Acc: 0.7267\n",
            "Epoch: 010, Train Acc: 0.6647, Test Acc: 0.7133\n",
            "Epoch: 011, Train Acc: 0.6224, Test Acc: 0.6933\n",
            "Epoch: 012, Train Acc: 0.6518, Test Acc: 0.6400\n",
            "Epoch: 013, Train Acc: 0.6576, Test Acc: 0.6667\n",
            "Epoch: 014, Train Acc: 0.6541, Test Acc: 0.6800\n",
            "Epoch: 015, Train Acc: 0.6482, Test Acc: 0.7333\n",
            "Epoch: 016, Train Acc: 0.6541, Test Acc: 0.7000\n",
            "Epoch: 017, Train Acc: 0.5941, Test Acc: 0.4733\n",
            "Epoch: 018, Train Acc: 0.6388, Test Acc: 0.7133\n",
            "Epoch: 019, Train Acc: 0.6259, Test Acc: 0.5333\n",
            "Epoch: 020, Train Acc: 0.5953, Test Acc: 0.5067\n",
            "Epoch: 021, Train Acc: 0.6118, Test Acc: 0.6600\n",
            "Epoch: 022, Train Acc: 0.6000, Test Acc: 0.4200\n",
            "Epoch: 023, Train Acc: 0.5847, Test Acc: 0.6867\n",
            "Epoch: 024, Train Acc: 0.6306, Test Acc: 0.5000\n",
            "Epoch: 025, Train Acc: 0.6024, Test Acc: 0.7000\n",
            "Epoch: 026, Train Acc: 0.6624, Test Acc: 0.7067\n",
            "Epoch: 027, Train Acc: 0.6459, Test Acc: 0.7200\n",
            "Epoch: 028, Train Acc: 0.6447, Test Acc: 0.6667\n",
            "Epoch: 029, Train Acc: 0.6518, Test Acc: 0.7200\n",
            "Epoch: 030, Train Acc: 0.6235, Test Acc: 0.6467\n",
            "Epoch: 031, Train Acc: 0.6212, Test Acc: 0.6867\n",
            "Epoch: 032, Train Acc: 0.6682, Test Acc: 0.7067\n",
            "Epoch: 033, Train Acc: 0.6612, Test Acc: 0.7267\n",
            "Epoch: 034, Train Acc: 0.6635, Test Acc: 0.6667\n",
            "Epoch: 035, Train Acc: 0.6353, Test Acc: 0.7200\n",
            "Epoch: 036, Train Acc: 0.6494, Test Acc: 0.6867\n",
            "Epoch: 037, Train Acc: 0.6765, Test Acc: 0.7400\n",
            "Epoch: 038, Train Acc: 0.6494, Test Acc: 0.7267\n",
            "Epoch: 039, Train Acc: 0.6753, Test Acc: 0.7200\n",
            "Epoch: 040, Train Acc: 0.6447, Test Acc: 0.7267\n",
            "Epoch: 041, Train Acc: 0.6482, Test Acc: 0.7000\n",
            "Epoch: 042, Train Acc: 0.6694, Test Acc: 0.6400\n",
            "Epoch: 043, Train Acc: 0.6718, Test Acc: 0.7200\n",
            "Epoch: 044, Train Acc: 0.6682, Test Acc: 0.7333\n",
            "Epoch: 045, Train Acc: 0.6718, Test Acc: 0.7133\n",
            "Epoch: 046, Train Acc: 0.6659, Test Acc: 0.7267\n",
            "Epoch: 047, Train Acc: 0.6906, Test Acc: 0.6267\n",
            "Epoch: 048, Train Acc: 0.6729, Test Acc: 0.7333\n",
            "Epoch: 049, Train Acc: 0.6824, Test Acc: 0.7267\n",
            "Epoch: 050, Train Acc: 0.6847, Test Acc: 0.6933\n",
            "Epoch: 051, Train Acc: 0.6718, Test Acc: 0.6467\n",
            "Epoch: 052, Train Acc: 0.6753, Test Acc: 0.7333\n",
            "Epoch: 053, Train Acc: 0.6682, Test Acc: 0.7400\n",
            "Epoch: 054, Train Acc: 0.6835, Test Acc: 0.6800\n",
            "Epoch: 055, Train Acc: 0.6847, Test Acc: 0.6867\n",
            "Epoch: 056, Train Acc: 0.6835, Test Acc: 0.6600\n",
            "Epoch: 057, Train Acc: 0.6518, Test Acc: 0.7200\n",
            "Epoch: 058, Train Acc: 0.6835, Test Acc: 0.6600\n",
            "Epoch: 059, Train Acc: 0.6682, Test Acc: 0.6933\n",
            "Epoch: 060, Train Acc: 0.6965, Test Acc: 0.6733\n",
            "Epoch: 061, Train Acc: 0.6682, Test Acc: 0.7667\n",
            "Epoch: 062, Train Acc: 0.6918, Test Acc: 0.7267\n",
            "Epoch: 063, Train Acc: 0.6729, Test Acc: 0.7600\n",
            "Epoch: 064, Train Acc: 0.6988, Test Acc: 0.6933\n",
            "Epoch: 065, Train Acc: 0.6859, Test Acc: 0.7533\n",
            "Epoch: 066, Train Acc: 0.6976, Test Acc: 0.7467\n",
            "Epoch: 067, Train Acc: 0.7012, Test Acc: 0.7400\n",
            "Epoch: 068, Train Acc: 0.6776, Test Acc: 0.7333\n",
            "Epoch: 069, Train Acc: 0.6694, Test Acc: 0.7333\n",
            "Epoch: 070, Train Acc: 0.6976, Test Acc: 0.6867\n",
            "Epoch: 071, Train Acc: 0.6765, Test Acc: 0.7467\n",
            "Epoch: 072, Train Acc: 0.7024, Test Acc: 0.7333\n",
            "Epoch: 073, Train Acc: 0.6800, Test Acc: 0.7467\n",
            "Epoch: 074, Train Acc: 0.6776, Test Acc: 0.7400\n",
            "Epoch: 075, Train Acc: 0.7035, Test Acc: 0.7533\n",
            "Epoch: 076, Train Acc: 0.7094, Test Acc: 0.6333\n",
            "Epoch: 077, Train Acc: 0.6953, Test Acc: 0.6600\n",
            "Epoch: 078, Train Acc: 0.6976, Test Acc: 0.7467\n",
            "Epoch: 079, Train Acc: 0.6929, Test Acc: 0.7067\n",
            "Epoch: 080, Train Acc: 0.6706, Test Acc: 0.6533\n",
            "Epoch: 081, Train Acc: 0.6859, Test Acc: 0.7467\n",
            "Epoch: 082, Train Acc: 0.6859, Test Acc: 0.7333\n",
            "Epoch: 083, Train Acc: 0.7035, Test Acc: 0.7200\n",
            "Epoch: 084, Train Acc: 0.7012, Test Acc: 0.7400\n",
            "Epoch: 085, Train Acc: 0.7000, Test Acc: 0.7533\n",
            "Epoch: 086, Train Acc: 0.6847, Test Acc: 0.7333\n",
            "Epoch: 087, Train Acc: 0.6988, Test Acc: 0.7600\n",
            "Epoch: 088, Train Acc: 0.7047, Test Acc: 0.7600\n",
            "Epoch: 089, Train Acc: 0.6965, Test Acc: 0.7533\n",
            "Epoch: 090, Train Acc: 0.7047, Test Acc: 0.7333\n",
            "Epoch: 091, Train Acc: 0.6965, Test Acc: 0.7533\n",
            "Epoch: 092, Train Acc: 0.6788, Test Acc: 0.7667\n",
            "Epoch: 093, Train Acc: 0.6753, Test Acc: 0.7533\n",
            "Epoch: 094, Train Acc: 0.6953, Test Acc: 0.7333\n",
            "Epoch: 095, Train Acc: 0.6859, Test Acc: 0.6867\n",
            "Epoch: 096, Train Acc: 0.6906, Test Acc: 0.7533\n",
            "Epoch: 097, Train Acc: 0.6729, Test Acc: 0.7467\n",
            "Epoch: 098, Train Acc: 0.7024, Test Acc: 0.6867\n",
            "Epoch: 099, Train Acc: 0.7012, Test Acc: 0.7400\n"
          ]
        }
      ],
      "source": [
        "optimizer.zero_grad()\n",
        "for epoch in range(1, epochs):\n",
        "    train_loss, train_acc = train(epoch, train_loader)\n",
        "    #_, train_acc = test(train_loader)\n",
        "    _, test_acc = test(test_loader)\n",
        "    print(f'Epoch: {epoch:03d}, Train Acc: {train_acc:.4f}, Test Acc: {test_acc:.4f}')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "zbAIJRfHrrxJ",
      "metadata": {
        "id": "zbAIJRfHrrxJ"
      },
      "outputs": [],
      "source": []
    },
    {
      "cell_type": "markdown",
      "id": "3MKGLZMZrsZB",
      "metadata": {
        "id": "3MKGLZMZrsZB"
      },
      "source": [
        "### Inspecting CT Layer: CT Distances and CT Embedding\n",
        "\n",
        "We will analyze the predicted CT distances and CT embedding, in order to better understand the rewiring and to to one’s imagination run riot in the use of commute times embedding $\\mathbf{Z}$ and commute times distance $\\mathbf{T^{CT}}$ (i.e. effective resistance)\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 18,
      "id": "d8CPDmxDrru5",
      "metadata": {
        "id": "d8CPDmxDrru5"
      },
      "outputs": [],
      "source": [
        "#num_of_centers = 420 #REDDIT model was trained with center=average degree of the graphs in the dataset (see DiffWire paper for more details)\n",
        "#dataset = TUDataset(root='data',name=\"REDDIT-BINARY\", transform = FeatureDegree())\n",
        "\n",
        "num_of_centers = 20\n",
        "dataset = TUDataset(root='data',name=\"IMDB-BINARY\", transform = FeatureDegree())"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 53,
      "id": "CENl2yNN3T8_",
      "metadata": {
        "id": "CENl2yNN3T8_"
      },
      "outputs": [],
      "source": [
        "BATCH_SIZE = 16\n",
        "full_loader = DataLoader(dataset, batch_size=BATCH_SIZE, shuffle=False)\n",
        "indiv_loader = DataLoader(dataset, batch_size=1, shuffle=False)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 49,
      "id": "20ZKvCHhBGdD",
      "metadata": {
        "id": "20ZKvCHhBGdD"
      },
      "outputs": [],
      "source": [
        "# get cc graphs\n",
        "connected_graphs = []\n",
        "for i, g in enumerate(indiv_loader):\n",
        "  G = to_networkx(g, node_attrs=[\"x\"], to_undirected=True)\n",
        "  if nx.is_connected(G):\n",
        "    connected_graphs.append(g)\n",
        "\n",
        "del g,i,G"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 21,
      "id": "IohtuWET1dte",
      "metadata": {
        "id": "IohtuWET1dte"
      },
      "outputs": [],
      "source": [
        "#trained_model_path = \"DiffWire/trained_models/CTNet/REDDIT-BINARY_CTNet_17_05_22__08_50_iter0.pth\"\n",
        "trained_model_path = \"DiffWire/trained_models/CTNet/IMDB-BINARY_CTNet_17_05_22__08_56_iter0.pth\""
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 52,
      "id": "0HtQr2YDxR64",
      "metadata": {
        "cellView": "form",
        "id": "0HtQr2YDxR64"
      },
      "outputs": [],
      "source": [
        "#@title GNN using CT-Layer and **saving CTEmbedding and CTDist**\n",
        "\n",
        "class CTNet_latent(torch.nn.Module):\n",
        "    def __init__(self, in_channels, out_channels, k_centers, hidden_channels=32, EPS=1e-15):\n",
        "        super(CTNet_latent, self).__init__()\n",
        "        \n",
        "        self.EPS=EPS\n",
        "        #Message Passing Layers\n",
        "        self.conv1 = DenseGraphConv(hidden_channels, hidden_channels)\n",
        "        self.conv2 = DenseGraphConv(hidden_channels, hidden_channels)\n",
        "        \n",
        "        # Pooling for CT embedding\n",
        "        num_of_centers1 =  k_centers # k1 #order of number of nodes\n",
        "        self.pool1 = Linear(hidden_channels, num_of_centers1)\n",
        "        \n",
        "        # Pooling for MinCut Layer\n",
        "        num_of_centers2 =  16 # k2 #mincut \n",
        "        self.pool2 = Linear(hidden_channels, num_of_centers2) \n",
        "\n",
        "        # MLPs towards out \n",
        "        self.lin1 = Linear(in_channels, hidden_channels)\n",
        "        self.lin2 = Linear(hidden_channels, hidden_channels)\n",
        "        self.lin3 = Linear(hidden_channels, out_channels)\n",
        "\n",
        "        self.ct_embedding = None\n",
        "        self.new_adj = None\n",
        " \n",
        "\n",
        "    def forward(self, x, edge_index, batch):    # x torch.Size([N, N]),  data.batch  torch.Size([661])  \n",
        "        # Make all matrices dense\n",
        "        adj = to_dense_adj(edge_index, batch)   # adj torch.Size(B, N, N])\n",
        "        x, mask = to_dense_batch(x, batch)      \n",
        "\n",
        "        #First layer: linear MLP\n",
        "        x = self.lin1(x) \n",
        "        \n",
        "        if torch.isnan(adj).any():\n",
        "              print(\"adj nan\")\n",
        "        if torch.isnan(x).any():\n",
        "              print(\"x nan\")\n",
        "        \n",
        "        # CT REWIRING\n",
        "        s1  = self.pool1(x)\n",
        "        self.ct_embedding = torch.tanh(s1.clone().detach())\n",
        "        adj, CT_loss, ortho_loss1 = dense_CT_rewiring(x, adj, s1, mask, EPS = self.EPS) \n",
        "        self.new_adj = adj.clone()\n",
        "        \n",
        "\n",
        "        # CONV1: Now on x and rewired adj: \n",
        "        x = self.conv1(x, adj) #out: x torch.Size([20, N, F'=32])\n",
        "\n",
        "        # MINCUT_POOL - Garph pooling\n",
        "        s2 = self.pool2(x) \n",
        "        x, adj, mincut_loss2, ortho_loss2 = dense_mincut_pool(x, adj, s2, mask, EPS=self.EPS) # out x torch.Size([20, k=16, F'=32]),  adj torch.Size([20, k2=16, k2=16])\n",
        "\n",
        "        # CONV2: Now on coarsened x and adj: \n",
        "        x = self.conv2(x, adj)\n",
        "        \n",
        "        # Readout for each of the 20 graphs\n",
        "        x = x.sum(dim=1) \n",
        "        \n",
        "        # Final MLP for graph classification: hidden channels = 32\n",
        "        x = F.relu(self.lin2(x)) \n",
        "        x = self.lin3(x) \n",
        "        \n",
        "        #loss functions\n",
        "        CT_loss = CT_loss + ortho_loss1\n",
        "        mincut_loss = mincut_loss2 + ortho_loss2\n",
        "        \n",
        "        return F.log_softmax(x, dim=-1), CT_loss, mincut_loss"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "cUBYk5_H3FsO",
      "metadata": {
        "id": "cUBYk5_H3FsO"
      },
      "source": [
        "Import trained model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 33,
      "id": "RDmyFld-xR5R",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RDmyFld-xR5R",
        "outputId": "24b871fb-c938-4d65-b3f6-04d61af5040c"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "CTNet_latent(\n",
              "  (conv1): DenseGraphConv(32, 32)\n",
              "  (conv2): DenseGraphConv(32, 32)\n",
              "  (pool1): Linear(in_features=32, out_features=20, bias=True)\n",
              "  (pool2): Linear(in_features=32, out_features=16, bias=True)\n",
              "  (lin1): Linear(in_features=1, out_features=32, bias=True)\n",
              "  (lin2): Linear(in_features=32, out_features=32, bias=True)\n",
              "  (lin3): Linear(in_features=32, out_features=2, bias=True)\n",
              ")"
            ]
          },
          "execution_count": 33,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "model =  CTNet_latent(dataset.num_features, dataset.num_classes,num_of_centers).to(device)\n",
        "model.load_state_dict(torch.load(trained_model_path, map_location=torch.device(device)))\n",
        "model.to(device)\n",
        "model.eval()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 34,
      "id": "Xi8qgpjuxR3I",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Xi8qgpjuxR3I",
        "outputId": "535798e1-dda1-44f9-cd12-1f45a0355bf8"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "tensor(14.3905, device='cuda:0') 0.715\n"
          ]
        }
      ],
      "source": [
        "#test loaded model\n",
        "loss, acc = test(full_loader)\n",
        "print(loss, acc)"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "x8HvW3sRv6iH",
      "metadata": {
        "id": "x8HvW3sRv6iH"
      },
      "source": [
        "#### CT Rewiring"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 356,
      "id": "Xi8zoXR3v56_",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Xi8zoXR3v56_",
        "outputId": "82f40efe-52ad-4d19-b673-85818e8f0c33"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "(True, 1, 13)"
            ]
          },
          "execution_count": 356,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "#Select a graph\n",
        "from itertools import islice\n",
        "#graph_i = connected_graphs[-10]\n",
        "graph_i= next(islice(iter(indiv_loader),234,None)) #15 in IMBD is complete graph, (0,1, 16, 17) clearly structured\n",
        "\n",
        "G = to_networkx(graph_i, node_attrs=[\"x\"], to_undirected=True)\n",
        "degree = np.array(list(dict(G.degree).values()))\n",
        "nx.is_connected(G), len(list(nx.connected_components(G))), G.number_of_nodes()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 357,
      "id": "VY00GyxQv52f",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VY00GyxQv52f",
        "outputId": "ffa9439a-1d4a-4837-87ea-206a1a062d2c"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "correct graph classification? True\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "torch.Size([13, 13])"
            ]
          },
          "execution_count": 357,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "#predict the new adjacency\n",
        "graph_i.to(device)\n",
        "pred, mc_loss, o_loss = model(graph_i.x, graph_i.edge_index, graph_i.batch)\n",
        "loss_graph = F.nll_loss(pred,graph_i.y.view(-1)) + mc_loss + o_loss\n",
        "print('correct graph classification?', bool(pred.max(dim=-1)[1].eq(graph_i.y.view(-1)).sum().item()))\n",
        "rewired_adj = model.new_adj.detach().cpu().squeeze()\n",
        "rewired_adj.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 358,
      "id": "0fCIfSNmkm4Z",
      "metadata": {
        "id": "0fCIfSNmkm4Z"
      },
      "outputs": [],
      "source": [
        "def get_cbar(cmap, array):\n",
        "  wmin = min(array)\n",
        "  wmax = max(array)\n",
        "  return plt.cm.ScalarMappable(cmap=cmap, norm=plt.Normalize(vmin = wmin, vmax=wmax))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 358,
      "id": "f0DD6zEjlAqx",
      "metadata": {
        "id": "f0DD6zEjlAqx"
      },
      "outputs": [],
      "source": []
    },
    {
      "cell_type": "code",
      "execution_count": 359,
      "id": "1954040a",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 270
        },
        "id": "1954040a",
        "outputId": "45eaa5ad-7bce-4d26-97c0-f44f8959fa7c"
      },
      "outputs": [
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABGgAAAD9CAYAAADgbkHTAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeVwU9f/A8dcsixyCICiCoJJiHnh98zYzO7RExVswz9TMq9Q8M1OzvPIozSwt08r7DDGPn0cepXlkmrd4gIh4gKLcsMfvj41N5HDFhWXx/Xw89qE7MzvzGdB5z77n83l/FL1er0cIIYQQQgghhBBCWIzK0g0QQgghhBBCCCGEeNZJgkYIIYQQQgghhBDCwiRBI4QQQgghhBBCCGFhkqARQgghhBBCCCGEsDBJ0AghhBBCCCGEEEJYmCRohBBCCCGEEEIIISxMEjSF2LRp0+jfv7/Zt30cRVG4dOmSWfYlhBAi/w0cOJBPP/3U0s0QQghhRpMnT6ZHjx75fpwDBw5QpUqVfD/O07KWdgrxNCRBU4CWLVtGzZo1cXR0xNPTk0GDBhEXF5fj9uPHj+f77783ad9Psq0QQoiCsXLlSurVq4eTkxNeXl60atWK119/HScnJ5ycnChWrBi2trbG961atcqyD19fXxwcHHB2dsbV1ZUmTZrw7bffotPpjNt8++23fPzxx49tj6+vL7t27TLrOQohREHKuCY6OTnh6elJnz59SEhIyLTNwYMHefXVV3F2dsbFxYW2bdty9uxZ4/oqVaqwZs0a4/s//vgDRVGyLHN2dkaj0WRpQ1xcHH379sXT0xNnZ2eef/55ZsyYYVxf2B92Ptq+l156iQsXLuRpXxcvXqRLly6UKlUKFxcXatWqxdy5c9m3b58xthUvXhxFUYzvnZycuHbtWqb9TJ48GVtbW5ydnY0/06FDhxIdHf3E7SyoxJYQ+UESNAVkzpw5jB07llmzZnH//n3+/PNPIiIiaNGiBWlpaVm2zy4YCCGEsB5z585l+PDhjB8/nlu3bnHt2jUGDx7M//73PxISEkhISGD8+PEEBQUZ32/bti3bfYWGhhIfH09ERATjxo1j5syZ9OvXr4DPSAghCofQ0FASEhI4ceIEf//9N9OnTzeuO3ToEC1btqRdu3bcuHGDq1evUrt2bV588UWuXLkCQLNmzdi/f7/xM/v376dq1apZljVu3Bi1Wp3l+CNGjCAhIYFz585x//59Nm/ejJ+fXz6eceF0+fJlGjZsSLly5Th16hT3799n3bp1HDt2jNq1axtj25kzZwBDYitjWfny5bPsLygoiPj4eO7evcumTZu4efMmdevWzZSkEaKokwRNAXjw4AGTJk3iq6++4s0338TW1hZfX1/Wrl1LeHg4y5cvZ/LkyXTu3JkePXpQokQJli1bliX7+9NPP1GhQgXc3d359NNPMz0JfXjb8PBwFEXhxx9/pHz58pQqVYqpU6ca93PkyBEaN26Mq6srXl5eDB06NNskkRBCiLy5f/8+EydO5Ouvv6Zjx44UL14cW1tb2rZty6xZs/K8XxcXFwIDA1mzZg0//vgjp0+fBqBPnz5MmDABgJiYGNq0aYOrqytubm689NJL6HQ6evbsybVr12jbti1OTk58/vnnAHTp0gVPT09cXFxo1qyZ8UY6Y79DhgyhdevWODs707BhQy5fvmxcf+bMGVq0aIGbmxtlypRh2rRpAOh0OmbMmEGlSpVwd3ena9eu3L17N8/nLYQQ2fH09OSNN97gxIkTxmVjxoyhV69eDBs2DGdnZ9zc3Pjss89o1KgRkydPBrImaA4cOMDYsWOzLGvWrFm2xz169ChvvfUWJUuWRKVSUbVqVTp37mzcN0Dt2rVxcnIy9sr57rvv8PPzw83NjcDAQG7cuGHcX07XUoC0tDR69eqFs7Mz/v7+HDt2zLgu4zrr7OxM9erV2bRpk3HdpUuXePnll3FxcaFUqVIEBQXl2L69e/fi4+Nj/GxkZCQdO3akdOnSuLu7M3To0Gx/DpMmTaJJkybMnTsXLy8vwNA7aeXKlbi6umb7GVPY2tri7+/PmjVrKF26NHPmzAHI0s6ZM2fi7e2Ns7MzVapUYffu3Wzfvp1p06axZs0anJycqF27NgBLly6lWrVqODs7U7FiRRYtWmTcT8Z+58yZg4eHB15eXixdutS4Pjk5mZEjR1KhQgVcXFxo2rQpycnJAPz55580adIEV1dXateuzd69e/N83kKAJGgKxMGDB0lJSaFjx46Zljs5OREQEMDOnTsBCAkJoXPnzsTFxdG9e/dM2549e5bBgwezYsUKoqOjuX//PlFRUbke9/fff+fChQvs3r2bKVOmcO7cOQBsbGz44osviImJ4dChQ+zevZuFCxea8YyFEOLZdujQIVJSUujQoUO+7L9Bgwb4+Phw4MCBLOvmzJmDj48Pd+7c4datW0ybNg1FUfj5558pX7688cnzmDFjAGjVqhVhYWHcvn2bF154IUv8Wb16NZMmTeLevXv4+fnx0UcfARAfH8/rr7/Om2++yY0bN7h06RKvvfYaAF999RW//PIL+/bt48aNG5QsWZIhQ4bky89CCPHsun79Otu2bTP2XklKSuLgwYN06dIly7Zdu3Y13nNnJKPv3r2LTqfj2LFjBAUFERcXZ1z2xx9/5JigadSoER999BFLly4lLCws07qMJM/JkydJSEggKCiIPXv28OGHH7J27Vqio6OpUKECwcHBQO7XUoDNmzcTHBxMXFwcgYGBmZIllSpV4sCBA9y/f59JkybRo0cPY2+Tjz/+mJYtW3Lv3j2uX7/Oe++9l2P7HqbVamnTpg0VKlQgPDycqKgoY1sftWvXLmNiKj/Y2NjQrl27bGPdhQsXWLBgAUePHiU+Pp4dO3bg6+vLm2++mal36smTJwHw8PBgy5YtPHjwgKVLlzJixAiOHz9u3N/NmzeN36+WLFnCkCFDuHfvHgCjRo3ir7/+4uDBg9y9e5fPP/8clUpFVFQUrVu3ZsKECdy9e5fZs2fTqVMn7ty5k28/E1H0SYKmAMTExFCqVKlsu0h6eXkRExMDQOPGjWnfvj0qlQoHB4dM261fv562bdvStGlTihUrxpQpU1AUJdfjTpo0CQcHB2rXrk3t2rWNF6i6devSqFEj1Go1vr6+vPvuu+zbt89MZyuEECI2NjbH6765lC1bNtteKba2tkRHRxMREYGtrS0vvfRSrvGib9++ODs7Y2dnx+TJkzl58iT37983ru/QoQMNGjRArVbTvXt345PqLVu24OnpyciRI7G3tzf2sAFDTZypU6fi4+Nj3O/69etl+K4Qwizat2+Ps7Mz5cqVw8PDg08++QTAmFzJ6M3xsIfvuStUqED58uU5cOAAJ0+epHLlyjg4OPDiiy8al6WlpRmvaY/66quv6N69OwsWLKB69er4+fnlOEQVYMWKFfTt25cXXngBOzs7pk+fzqFDhwgPD8/1WgrQtGlTAgICsLGxoWfPnsb7eTD0gCxbtiwqlYqgoCAqV67MkSNHAEMsiIiI4MaNG9jb29O0aVOTfrZHjhzhxo0bzJo1i+LFi+f62djY2Gx/1uaUU6yzsbEhNTWVs2fPkp6ejq+vL5UqVcpxP61bt6ZSpUooisLLL79My5YtMyV+bG1tmThxIra2tgQEBODk5MSFCxfQ6XT88MMPzJs3D29vb2xsbGjSpAl2dnYsX76cgIAAAgICUKlUtGjRgnr16rF169Z8+VmIZ4MkaApAqVKliImJyfbGNDo6mlKlSgFQrly5HPdx48aNTOsdHR1xd3fP9bienp6Zts8ooHbx4kXatGmDp6cnJUqUYPz48caAJYQQ4um5u7vneN3PTatWrYwFFFesWJHrtlFRUbi5uWVZPnr0aPz8/GjZsiUVK1bMVLjyUVqtlnHjxlGpUiVKlCiBr68vQKaYkFMsiYyMzPFmOCIigg4dOuDq6oqrqyvVqlXDxsaGW7du5XpOQghhil9++YX4+Hj27t3L+fPnjdesjCFH2dUsefieG/4b5rR//35eeuklwJAMyVjWoEED7Ozssj2+g4MD48eP56+//iI2NpauXbvSpUuXHIdy3rhxgwoVKhjfOzk54e7uTlRUVK7XUsh6DU5JSTHGlp9++ok6deoYr7WnT582/iw+//xz9Ho9DRo0wN/fnx9++CHHYzwsMjKSChUqmPSAwd3dPU/1YVasWJFrcfyH5RTr/Pz8+PLLL5k8eTIeHh4EBwdnGjb2qG3bttGoUSPc3NxwdXVl69atmWKdu7t7pnPOiHcxMTGkpKRk+zuKiIhg3bp1xp+/q6srv//+u9TMEU9FEjQFoHHjxtjZ2bFx48ZMyzMKQmZ0Y8ztCaeXlxfXr183vk9OTiY2NjZP7Rk0aBBVq1YlLCyMBw8eMG3aNPR6fZ72JYQQIquM6/4vv/zyRJ/btm2bsYDio0ONHnb06FGioqKyfarp7OzMnDlzuHLlCps3b2bu3Lns3r0byBpnVq5cSUhICLt27eL+/fuEh4cDmBQTypUrZyy4md26bdu2ERcXZ3ylpKTg7e392P0KIYSpXn75Zfr06cOoUaMAKF68OI0bN2bdunVZtl27dm2moUMZCZoDBw4YEzQvvfSScVlOw5selfGwMzExkatXr2a7TdmyZYmIiDC+T0xMJDY2Fm9v71yvpbmJiIjgnXfeYcGCBcTGxhIXF0eNGjWM129PT0++++47bty4waJFixg8eLBJM0uVK1eOa9eumfSA4fXXX2fDhg1P3Pbu3bs/tjg+GOqZhYaGGn8/j3rrrbf4/fffiYiIQFEUxo4dC2SNdampqXTq1IlRo0Zx69Yt4uLiCAgIMCnWlSpVCnt7+0z11zKUK1eOnj17Zop1iYmJjBs37rH7FSInkqApAC4uLkyaNIn33nuP7du3k56eTnh4OF27dsXHx4eePXs+dh+dO3cmNDSUgwcPkpaWxuTJk/OcVImPj6dEiRI4OTlx/vx5vvnmmzztRwghRPZcXFyYMmUKQ4YM4ZdffiEpKYn09HS2bdtmrP2SFw8ePGDLli0EBwfTo0cPatasmWWbLVu2cOnSJfR6PS4uLtjY2KBSGcJ9mTJlMn0RiI+Px87ODnd3d5KSkhg/frzJbWnTpg3R0dF8+eWXpKamEh8fz+HDhwEYOHAgH330kfELyZ07dwgJCcnzeQshRE6GDx/Ozp07jUN/ZsyYwY8//sj8+fOJj4/n3r17TJgwgUOHDjFp0iTj55o1a8bff//N/v37efHFFwGoWbMmV69e5bfffss1QfPpp59y9OhR0tLSSElJYd68ebi6ulKlShUg67W2W7duLF26lBMnTpCamsr48eNp2LAhvr6+uV5Lc5OYmIiiKJQuXRowFMHNKBwPsG7dOuPD3ZIlS6IoSo6x4GENGjTAy8uLcePGkZiYSEpKCn/88Ue2237yySccPHiQ0aNHc/PmTcBQnLhHjx7ExcU99hxyotFoOHfuHN26dePmzZt88MEHWba5cOECe/bsITU1FXt7exwcHDKdX3h4ODqdDjAUWk5NTaV06dKo1Wq2bdvG//3f/5nUFpVKRd++ffnggw+4ceMGWq2WQ4cOkZqaSo8ePQgNDWXHjh1otVpSUlLYu3dvpofqQjwpSdAUkDFjxjBt2jRGjRpFiRIljFPS7d69O8fukw/z9/fnq6++Ijg4GC8vL5ycnPDw8DDps4+aPXs2K1euxNnZmXfeeSdLcTAhhBBPb+TIkcydO5fPPvuM0qVLU65cORYsWED79u2feF9t27Y11luYOnUqH3zwQaYZJh4WFhbG66+/jpOTE40bN2bw4MG88sorAHz44Yd89tlnuLq6Mnv2bHr16kWFChXw9vamevXqNGrUyOQ2OTs7s3PnTkJDQ/H09KRy5cr89ttvAAwbNozAwEBatmyJs7MzjRo1MukLhxBCPKnSpUvTq1cvpkyZAhiGKe3YsYONGzfi5eVFhQoV+Pvvv/n999+pXLmy8XPPP/88pUuXxtPT0zjjkEqlokGDBjx48IAmTZrkeExFUXj77bcpVaoUZcuWZefOnfz66684OTkBhtlVe/fujaurK2vXruX111/n008/pVOnTnh5eXH58mVWr14N5H4tzU316tUZOXIkjRs3pkyZMpw6dcqYaAJDT8uGDRvi5OREYGAg8+bNo2LFitm272E2NjaEhoZy6dIlypcvj4+Pj3EmqkdVqlTJWEvH398fFxcXOnXqRL169XB2dn7sOTwqY+aljBkL3d3d+euvvyhbtmyWbVNTUxk3bhylSpXC09OT27dvG6dbzygS7e7uzgsvvICzszPz58+na9eulCxZkpUrVxIYGGhyu2bPnk3NmjWpX78+bm5ujB07Fp1OR7ly5QgJCWHatGnGOD9r1ixjYkiIvFD0MrbFKiUkJODq6kpYWBjPPfecpZsjhBBCCCGEEEKIpyA9aKxIaGgoSUlJJCYmMmrUKGrWrGks6CiEEEIIIYQQQgjrJQkaKxISEkLZsmUpW7YsYWFhrF69+rFTbQvRt29fPDw8qFGjhnHZ6NGjqVq1KrVq1aJDhw5PNU7Y0rI7v4edP3/eWLB19uzZmdZt376dKlWq4Ofnl2mmm6tXr9KwYUP8/PwICgoiLS0tX89BCCEsKbvr6Lp16/D390elUnHs2DELtk4IIYSlFfU4kdN3ggzLli2jdOnS1KlThzp16vD999/nW1skQWNFvv/+e+Li4rh//z67d+82FiITIjd9+vRh+/btmZa1aNGC06dP888///D8888bx+xao+zO72Fubm7Mnz/fOMNDBq1Wy5AhQ9i2bRtnz55l1apVnD17FoCxY8cyYsQILl26RMmSJVmyZEm+noMQQlhSdtfRGjVqsHHjRpNnshFCCFF0FeU4kdt3gocFBQVx4sQJTpw4Qf/+/fOtPZKgEaKIa9asGW5ubpmWtWzZErVaDUCjRo2sutp8duf3MA8PD+rXr4+trW2m5UeOHMHPz4+KFStSrFgxgoODCQkJQa/Xs2fPHjp37gxA7969n3iqZCGEsCbZXUerVasmD4KEEEIARTtO5PSdwFLUFjuyECJHdVXFeaDXmrRtseqVcHBwML4fMGAAAwYMMPlYP/zwQ4HP5PWm8iYxxJi0bXL15Kc6v5xERUVRrlw543sfHx8OHz5MbGwsrq6uxgSWj48PUVFRT308IYQwp4KME5bwJHGi1Bulcu1JKYQQzyKJE//J7ftETt8JHrVhwwb279/P888/zxdffJHpM+YkCRohCqEHei1fqiuYtO0kB4c8j/ucOnUqarWa7t275+nzeRVDDMcwrc31HOpZ/bhWIYQwt4KKE5byRHEipl4+t0YIIayPxIn/PO33ibZt29KtWzfs7OxYtGgRvXv3Zs+ePXneX24kQSNEYaSAYpu/BaCXLVvGli1b2L17t0WKTestXN/a29ubyMhI4/vr16/j7e2Nu7s7cXFxaDQa1Gq1cbkQQhQqBRAnLM3ScUIIIayaxAmT5PSd4GHu7u7Gv/fv358xY8Y8/YFzIAkaIQohRaVg42CTb/vfvn07n3/+Ofv27cPR0THfjpMbnYUrYNWvX5+wsDCuXr2Kt7c3q1evZuXKlSiKwiuvvML69esJDg7mxx9/pF27dpZtrBBCPCK/40RhYOk4IYQQ1kzihGly+k7wsOjoaLy8vADYvHkz1apVe/oD50ASNEIURgqo1ObJeHfr1o29e/cSExODj48Pn3zyCdOnTyc1NZUWLVoAhkLB3377rVmOZwo95nsymt35paenAzBw4EBu3rxJvXr1ePDgASqVii+//JKzZ89SokQJFixYwBtvvIFWq6Vv3774+/sDMHPmTIKDg5kwYQL/+9//6Nevn3kaK4QQ5pLPccLNzY333nuPO3fu0Lp1a+rUqcOOHTvMcjxTmDNOCCHEM0nihEnUanW23wkmTpxIvXr1CAwMZP78+WzevBm1Wo2bmxvLli17+gPnQNHr9fp827sQIk+eL+bAQg8/k7Yd52lndWNG66rq8afatDY3riU1aIQQ4lESJ/4jcUIIIbKSOPEfa4oT0oNGiEJIURSzZbwLK3kyKoQQeSdxQgghRG4kTlgnSdAIURhJUS8hhBC5kTghhBAiNxInrJIkaIQojMw4ZrSwKooXVCGEKDASJ4QQQuRG4oRVkgSNEIWQooBNsaI7fYVekdk5hBDiaUicEEIIkRuJE9ZJEjSi0MioV60oRTAV+sQUFFXR/jkUxYy3EEIUHIkTQgghciNxwhpJgkZYVEqallt304h5kIZWZ1hmowK3EsXwLFkMBzsbyzbQUhRQbExNCVvnRGxF8YIqhBAFRuKEEEKI3EicsEqSoBEWodXquXQjiQdJGh6d6F2rgztxacTcT8PJ3gY/H0dsTb64FA0KoLIx8YqjlQuqEEI8ayROCCGEyI3ECeskCRpR4DRaPWfCE0hL1+Waq9XrISFZy5mrCfj7OmGrfoaSNAqmd0nU5m9T8oOeonlBFUKIAiNxQgghRG4kTlglSdCIAqXX67kQmfjY5IxxeyBNo+f8tURqPOf0zNSnURTF9KJe6VZ4RS2iRb2EEPlHp9NzOyYVtY1CKXc7SzfH4iROCCFEZhqNjtsxqTjY21DStZilm2NxEieskyRoRIGKT9KSnKp94lGOqek64hI1lHSyzZd2FUaKqghecR5SFDPeQgjz0+v1bNp2g6WrIkhK1qLXgaeHHe/1r0Tjeu6Wbp5FSZwQQghD6YSf1l1jzS+RaDR6tDo9FSsUZ8S7ftSo6mLp5lmUxAnrU7R/Y88gvV5PXEI6V6KTuBCZyMXriVy7lUxyauHIikbfTUWXhyGOOj1Ex6aav0GF1b9dEk15WaOMLommvIQQz7bvloez8Icr3ItLJzVVR1q6jmtRyUyYfpa9f9yxdPMsR+KExAkhBACffXGeFeuvkZCoJSVVR3q6nguXEhg24R9Onbtv6eZZjsQJq4wT0oOmiNDr9dy8m0r03TR0On2WJMituDQc7WzwKW2PS3HL/No1Wh0PkjR5/nxiipa0dB3FbJ+FvKJielEvK2VtF0shRMG7F5fG6k2RpKVnzeynpumY800YzRqXQmWlN5dPR+KEEEJcuprA/kMxpKbpsqxLTdUx99tLLJ1X1wItKwwkTlijZ+GbbpGn0+m5EJlEVEwqGm3W5AwYCu4mpmgJu57IzXuW6YmSlq7naUrIqBRI1WS9+BZFShHPeGNitrsoXnSFEKbb/2dMrsmX1DQd58LiC7BFhYfECYkTQgjYuf826bl8PwiPTCT2XloBtqjwkDhhnXFCetBYOb3eMF11fHLW6aqzo9PD9dspqFUKpVwKpnhWeno6169fJzLqNraulbBR572OjP7ZyM8AMmZUCCGSkrVocpn6U1EgOaVwDOG1BIkTQohnXUKCBl0u3w/UNgrJyVooWXBtKkwkTlgfSdBYubhEDQ8STUvOZNDpIfxmMiWdbbExQ8Y0MTGRiIiITK9r164Z/37r1i08PT15oV5DRk/+Gps8HkeP4SL7TFAUbIrwUC49RbPquhDWIuxKAqs2RXLuYjxOxdW0a+VFy+ZlCt0Q0mqVnbFVK2g02Qe59HQdlSoUL+BWFRISJ4QQ+ejkmThWbbpOeGQS7iWL0amNNy83LoVNIbsXr1Xdhf/bezvHZL2iKHiUfkZn/pM4YZUkQWPlomPzVnQX4O6DdEo/Zgo6vV7P3bt3syRgHn4lJCRQvnx5KlSoYHy98cYbxr97e3tja2uLXq/nxKV40nN5GpobBQUHuyL4vzAbGV0Si7KimPEWwhqE/l80Xy6+RHq6zvjU8eq1RDZsieLrGf/D0SGvaXTzq+3vQik3OyJvJsMjT0htbRWaNiz1zE6las440bdvX7Zs2YKHhwenT58G4O7duwQFBREeHo6vry9r166lZMmCfQQtcUIIy/h+xVVWb7pOapoOvR6u30jm4uV4ft3pwsyJNQvVA9PmL5Zm3jeXSM5mnV0xFR1bly10Dx8KisQJ6yQJGiuWkqYjMY9du3V6uBGbipuzDdHR0Tn2fomIiECtVmdKvlSoUIEmTZoY/+7h4YFiQnEZRVHwdCtGVMyTJ5UUBcqULGbScYqKIt0l0QrHgwpRFNy8ncIXiy6R9kgxxZRUHRGRSSz++QrDB1S2UOuyUhSFcp/W4sa4E2gSNCjJWhQF7O1sqOTryLj3q1i6iRZlrjjRp08fhg4dSq9evYzLZsyYwWuvvca4ceOYMWMGM2bMYObMmWY5nkkkTghhEafPP2D1puukpGaOE8kpOk6euc+mX6PoEuhjodZlZRehYt6ftRlW/STpio5ktQ5FMSRnGvyvJP17PGfpJlqUxAnrIwkaK5aYokHB0L0rL5JTNTiXKIFLiRKZki/+/v4EBAQY37u4uJitzaVd7YiKyVuRYo+Sz9BTUulBI4TIB5t3RKPPIUOelq5ny86bDO3nV2iejv4AnPKwp9riBngfjsX2r7vY2iq82tSDOjVcnqmkfRZmjBPNmjUjPDw807KQkBD27t0LQO/evWnevHnB3nhj3jixfft2hg0bhlarpX///owbNy7T+tTUVHr16sVff/2Fu7s7a9aswdfXl9jYWDp37szRo0fp06cPCxYsyLLvwMBArly5YnyqLIQ1W7f5erYzIoEhmb/6l+uFJ0FzB2gFlROd2LSvEXs63eFE/fsUd7Sh5StlqOrnbOkWWpbECaskCRorpn3KgrkqlcLt2zE4FXcwT4NMoLZRqFTWkYuR8Sgq07rRK+h5ztORYuoi3KMkCyuuqG4ic11Qs+tymek4ej3Dhg1j69atODo6smzZMl544QV+++03RowYYdzu/PnzrF69mvbt29OnTx/27dtnTE4uW7aMOnXqmKfBQljQ1WuJpOdQzwVAp9WTkKDB1SXvxdzN5f+Ab4FVQDu1ihMvlkb9YmkLt6owyd84cevWLby8vADw9PTk1q1b+XasnJgrTmi1WoYMGcLOnTvx8fGhfv36BAYGUr16deM2S5YsoWTJkly6dInVq1czduxY1qxZg729PZ9++imnT5/ONsZs3LgRJycn8zRUiEIgIjIp19qWsXcLyYxIyUAgEAV0Bs9Bnf8AACAASURBVLuVNrSa7EkrP08LN6wwkThhjZ6lb7xFjo0KeKp/lAqODvZmao3proefZ+YnI9Bq0tFo0nPeUK9Do0ln7c8LcHEsgv/7cqEooFLbmPSyRhlFvUx5PU6fPn3Yvn17juu3bdtGWFgYYWFhLF68mEGDBgHwyiuvcOLECU6cOMGePXtwdHSkZcuWxs/NmjXLuF6SM6Ko8PSwN8SOXDg6Wv66chL4APgFWAIMRZ4oPepJ4sSdO3eoV6+e8bV48eInPJZS4L2VzBknjhw5gp+fHxUrVqRYsWIEBwcTEhKSaZuQkBB69+4NQOfOndm9ezd6vZ7ixYvTtGlT7O2z3i8lJCQwd+5cJkyYYI5TFqJQeFxBXWfnQnA11gI9gMPAYiAE6AL4WbJRhY/ECdPjRGFiZc0VD3Ows8n7+CbA1kZBVcC9NKKiomjdujXtAl5lzJBgdm3bCOhRqUCjSSc9LRWVAlqthotnj1OzYglOH/+DUaNGFWg7CwOVjWLSyyr9O2bUlNfjNGvWDDc3txzXh4SE0KtXLxRFoVGjRsTFxREdHZ1pm/Xr19OqVSscHR2f9syEKNTavemFOofeiGobeKVpaYsXU7wOvIWh50xxYBPQx5INKsRMjROlS5fm2LFjxteAAQMeu+8yZcoYr5XR0dF4eHjk9+lk9gRx4nFfLKKioihXrpzxvY+PD1FRUTluo1arcXFxITY2Ntcmfvzxx4wcOVJihyhSOrXxxt4++zhQzFahQ6uyBdyibIwGNgJzgUggHhiX6yeeWRInrK+XjSRorJijnQ12xfL2K0xNTeGHRXP4/PPPuXPnjplblr0HDx7QunVrBg8eTEREBA/i7nDqyE7qVnbhOU8HDu/bwqL5U/H1dKCSB7zXvzMpifdZtWoV27ZtY+nSpQXSzkJBMXRJNOVlrQrqgmrKjfnq1avp1q1bpmUfffQRtWrVYsSIEaSm5q1ukhCFjW+54gS198H+kRnxbG0VSroWY8jblSzUMoMHQHtgHlAT+AZ4Gyj4vp5WIJ/jRGBgID/++CMAP/74I+3atTNn601iapzIyxeLp3XixAkuX75Mhw4d8v1YQhSkBv8ryUsNS2WJE3bFVPiUdSC4Q7kcPllA5gNfAO8DA4AvgVaAdHbOSuKEJGhEwfNysyMv/6fs7e3p0aUN586do3LlynTr1o19+/ahz23Q6VNIT0+na9euNGzYkKCgIObPn8/t27f58MMPsbFRcC9RjNPHD7D8h69wL2GLR2l3OnTowJIlS3B1dSUkJISxY8fy559/5kv7CiNFpTLp9Th9+/bFw8ODGjVqGJetW7cOf39/VCoVx44dy8/TyJG5now+rejoaE6dOsUbb7xhXDZ9+nTOnz/P0aNHuXv3boEXPBMiPw3o+RwTR1UjvbIzOnsVJV1tCQr0Ydn8erhZsBh7OoYe6u8Br2MoL7AMGGixFhV+5ooT3bp1o3Hjxly4cAEfHx+WLFnCuHHj2LlzJ5UrV2bXrl1ZiuoWBHPdeHt7exMZGWl8f/36dby9vXPcRqPRcP/+fdzd3XPc56FDhzh27Bi+vr40bdqUixcv0rx58zydpxCFiaIofPxBVUbWrsxz8Y7Ya1R4lLSjd3AFFs1+AUcHCw6D/QUYjiGTPxdDJfk7wIeWa1JhJ3HC+hI0hWAQoXgabiVsuXkvleQUncmjndLT0/D1csa7Sn0aNlzK3Llz+fnnnxk0aBB6vZ6BAwfSq1cvs81jr9frGTRoEDY2Nnz99dd06dKFZs2akZiYyAsvvGDcTqvVYmtrS2JiIk5OTgwZMoSOHTsyevRoqlWrxpIlS+jcuTNHjhyhbNlC0L0yHylmrLqe3bR4NWrUYOPGjbz77rtmOcaT0mP6xTLjyWhePe7GfO3atXTo0AFb2/+KomYUPLOzs+Ptt99m9uzZeT6+EIVR8Ual8GpUiivAQUs3BsM1YSDQBOj977KlQGeghKUaVciZM06sWrUq2+W7d+82y/7z4knixOPUr1+fsLAwrl69ire3N6tXr2blypWZtsl4Ety4cWPWr1/Pq6++mms9hUGDBhlrmoWHh9OmTRvjbCZCWDuVSqHVaU9aHfc0DB9KAgpuTpHsHcYw/rUBsALQAbOAF4GXLNiuQkzihHWSHjRWTqUoVC3nhF0xFabUZUpNTeH33aFMmzQSnc4wDVTJkiV5//33OXPmDIsWLeLw4cM899xzvP322xw+fPipe9VMmzaN48ePs2bNGnbt2sXp06c5deoUH36YOd2t0WhwcHAgPj4egLp161K2bFm2bNkCQNu2bRk0aBAdO3YkJSXlqdpkDczVJTG7Gi3VqlWjSpUq+dV0kxRUxjswMJCffvoJvV7Pn3/+iYuLizEBA4aA8+jwpozxtHq9nl9++SVT7yMhioJQoDlgnjT80/sMw732xH/fa4CvgWEWa5F1kKGwpsUJtVrNggULeOONN6hWrRpdu3bF39+fiRMnsnnzZgD69etHbGwsfn5+zJ07lxkzZhg/7+vrywcffMCyZcvw8fHh7Nmz+XXKQhQOacB2wBeww/LJmctAW8AL2Aw4YihUdg2pPfMYEiesL4kjPWiKALWNgr+vExG3krn7IJ3k5GTs7DNfSVOSk0hKSmTH5uXYq1L5+++/GTx4MAsXLkT1b7c2RVFo1qwZzZo14/bt2yxbtozu3bvj7OzMu+++a/z7k1i+fDnfffcdBw8exNbWlvfff58uXbrw22+/8fLLL2fa9uEETcYX6KFDh7JgwQLjmMbx48dz8uRJBg4cyNKlSwu8WniBURSTZ2jKGAKUYcCAAQUy/v6pKOarqN6tWzf27t1LTEwMPj4+fPLJJ6SnG2YHGzhwIAEBAWzduhU/Pz8cHR0z1TIKDw8nMjIyy7/F7t27c+fOHfR6PXXq1OHbb781T2OFKCR+xTBsf6+F2wHwE3AA2MJ/ExOuBl4DZFLtXDxBnLBKZowTAAEBAQQEBGRaNmXKFOPf7e3tWbduXbafDQ8Pz3Xfvr6+2U7BLYTVOoChKJgncNvCbYkFAjDM3LQV8MCQ0Z+BoVhZa8s1rdCTOGGVJEFTRNioFCp6OeJkE8PIz6fy3oiPuHk7Bl/fCtjZqoi8cJIJ40ZQvnx5/vnnH5YtW8Znn33GgAEDWLx4sTFJk8HDw4MxY8YwatQodu/ezbfffsuHH35IUFAQAwcONGna4d9++40PPviA3377jbJlyzJ9+nSqVavGjh07mDRpUpbkyqM9aMAw1eXIkSM5f/48VatWRVEUli5dSpMmTZg/fz7DhhXV56uKSeNB4emHAFmKubLZOXW5zKAoCl9//XW263x9fbMUDAbYs2ePWdomRGEUhSH42wCuFm7LHgz1HncDGdVvdBhKC/xiqUZZDdPjhLWytqeeQhQZoRh6zhTHsl0tU4B2QASwC8jo/L0ZOIdhqJNcJ3IhccIaFe3f2DNox45tJNy9TnLsBWZPHkytis5UKVec1m80JyYmhtOnT/PBBx8wYMAAVqxYwaVLl+jbty9arTbb/alUKlq0aMGGDRs4c+YM3t7eBAYG0qhRI5YtW0ZSUlK2nztz5gxBQUGsXr0af39/IiMjmTNnDp06dSI1NZU2bdpk+YxWq8XR0TFTgsbOzo7+/fuzcOFC47LixYsTEhLC9OnTLTruMd8pimkvK5QxZrSodUkUwhr8CrQB7mHZ++7TGAoCbwJcHlq+BagFlLdEo6yNxAmJE0KYmx5DguY1DPVnLBUodBiKkv2Boatl03+X64HpQEWgq2WaZlUkTlhdnJAETRETGhpK27ZtuXfvXqYiv/b29vTv358aNWqwadMmOnTowPDhw9myZQuRkZH07t0bjUaT677Lli3Lxx9/zJUrV/joo49Yt24d5cuXZ9iwYZnGY0dHR9O6dWvmzJnDq6++CsDIkSMZOnQoS5cuZdy4cVl67IChB82jCRqAd999l+XLl2da7uvry6pVq+jevTtXrlzJ08+qMMso6iVjRi3dSiGKnlAMQ/ktmaC5AQRjePj58IStemAmMNYSjbIyEickTgiRL84DV7B8oPgQWAt8TuZEzG/AEWA0MhbkMSROWGeckARNEZKWlsbOnTtp3bo1cXFxWWZhGjhwIEePHiU+Pp4XX3yRq1evsnz5ckJDQ7l9+zY9e/Z8bJIGDMX22rZty6+//sqxY8dwdnbmtdde4+WXX2bp0qW0bt2afv360bNnTwB27drFsWPHaN68OVevXiUoKCjb/WYkaB48eJBpebly5XjllVdYvnx5puWvvPIKEyZMoF27diQkJDzJj8oq5Oe0eJs2bcLHx4dDhw7RunXrTFNMFwgTL6bWdkEVorBLAi5hGLYfh2Xuu+MxzJA6B3h0sOw+oAxQraAbZaXMFScKJYkTQlhG6L9/tsaQoLHEWNhvMCRmBgOjHlk3HUNtnD4F3CYrJXHC+uKElf42RHb2799PlSpVKFOmDPfu3cPVNfMVtUKFCjRr1ozmzZszceJEli9fbuwRExISQlxcHG+99ZaxwKopfH19+eyzz7h27RqDBw9m9OjRnDt3jvj4eC5fvkxaWhrvvfceX375JV988QWjR4/ONJ3xwzQaDcWLF8/Sgwb+Kxb86IxSQ4YMoUGDBvTp08c4K1WRoJiW7TYl471q1Sqio6NJT0/n+vXr9OvXjw4dOnD9+nVSU1O5desWO3bsKICTyqwoXlCFKOx2A69iGLJviftuDRAEvAtklxaejkzIYTIzxonCSuKEEBYQiiF7Xg7LZPK3AEMxjMWdR+YaM8cw1KIZAdgXcLuskcQJq4wTkqApQjKGNwFZhjhlGDp0KHv37sXT05NDhw4xa9YsgoODAdi0aRNJSUkEBQWRlpb2RMdWq9Xs3buXunXrcuzYMXQ6HY0aNaJGjRo4ODjg4+PD4cOH6du3b4770Gq1OSZomjdvDsDevXszLVcUhYULFxIVFcXUqVOfqM2FnUptY9LLGukxVF035SWEMJ+M4U1Q8D3X9cAQ4AWgXzbr/8KQwGlQgG2ydhInJE4IYVaxwEEMgUJHwSdojmHI4v8Pw3R+jw5hmo7hycLAAmyTlZM4YX1xwsqaK3Ki1+sJDQ01Ft/NKUHz2muvkZ6eTnBwMJMnT6ZLly7UqlWLkSNHYm9vz4YNG9DpdHTp0oXU1FSTjz9r1iwOHjzIunXr8Pf3Z/bs2Rw+fJjo6GhUKhVNmzalevXq3LlzJ8d9aDQanJycsk3QKIrCkCFDsp2Nx87Ojo0bN7J48WJCQkJMbnNhJmNGrTPjLURhpscwhKj5v+8LOkEzE8MQq09zWD8DQ8kBYRqJExInhDC7rRgSM20xTLOtp+ACRTiGXjOlMfSiKf7I+vMYqsoPAUoUUJusnMQJ64wTkqApIs6dO4dGo6FWrVoA2daggf8SHbt37zZOVf3NN9+wfft2Nm3ahJ2dHWvXrkWtVtOxY0dSUlIee+zVq1ezYMECfv31V0qU+O+K+fHHHzNs2DDWrVuHra0tfn5+1KlTh8DAQLZu3Zpl5qjchjgB9OzZkz179hAZGZllnZeXFxs2bKB///6cOXPmsW0u/BRQqUx7WaMiOmZUiMLsOODPf73CC/LB6EpgJ7CE7GdEvQBEYpg0RJhK4oTECSHMLBRDfZe6GIIEFMxY2HtAAJAKbPu3DY+aiSGADSuA9hQZEiesMU5Y6W9DPCpjeJPy7zRp2dWgydC7d2927drFe++9x5w5c9BoNKxatYqBAwcSGRlJsWLFWL16NcWLF6dDhw4kJyfneNwDBw7w/vvvs2XLFnx8fIzL9+3bx++//86HH37I7NmzGTx4MIsXLyYyMpJ27doxefJkKlasyNSpU7l58yZgSNA4OzvnmKBxdnamR48eLFq0KNv1DRo0YM6cObRv35579+6Z9HMrzBRFMellrYriBVWIwuzh4U1QcDVo9gGzgfVAsRy2+RwYQ/bJG5EziRMSJ4QwmzRgB4ZeLCoMQQLyP5OfCnQALgO/kH2V+GvAcqA/hh42wmQSJ6wvTkiCpoh4uP4M5DzECaBEiRJ069aNXbt20bVrV6ZNm0bDhg0ZMWIE3bt3R6vVYmtry8qVK3F1daVdu3YkJSVl2c/58+fp3LkzK1asMPbcAUhPT2fo0KHMnTuXhIQEVq1axfDhwwEoXrw4/fr148iRI2zcuJGIiAiqVatGly5diIuLy7UHDcDgwYP5/vvvcxx+1atXL9q0aUNwcLBJM1IVWkoRr7pO0bygClGY/YrhAWWGghjidA4YhKFXek7Huo6h7ED7fG5LkSNxQuKEEOZ0AMOwpjb/vi+IBI0e6Ishk78UeDmH7eb8++ejMzqJ3EmcsMo4Yb2/DWEUExPDqVOnjIV0IfcEDWDs0TJu3DiWLVtGREQEY8aMoVixYsZiu2q1mp9//pkyZcrQtm1bEhMTjZ+/desWAQEBzJgxgxYtWmTa98KFC/Hy8qJjx458+eWXdOvWjTJlymRpQ926dVm8eDHh4eG88sorREZGMm/ePP755x9iY2OzbXfVqlWpWbMm69evz/HcZs2ahVar5cMPrbiagaKgqG1MelmjolrUS4jCKgpDrcWHr8TJgEM+HvMm0BX4GaiQy3ZzMEzIIf/dn5DECYkTQphTKGAHvP7v+4wETX52tZyAYQzsNOCtHLa5A3wHdAfK52NbiiKJE1YZJ6ysuSI727Zt49VXX8Xe/r/55nKqQZPB39+fatWqcfDgQYYMGcLEiRNRqVT89NNPLFy4kN9//x0wJGmWLVtG+fLlCQgIICEhgcTERNq2bUvPnj15++23M+335s2bfPbZZ8yfP58HDx6wePFiRo8enWv7XVxcGDx4MBUrVmTAgAHcvXuXSpUq0bNnT/74448sU2tnTLmdE7VazZo1a9i4cSMrVqzI9diFWZEu6lVEx4wKUVj9yn8PRQtCIoYeMTMxlDLISSyGHvU53ZeL3EmckDghhFnoMSRoXuO/4rwZNWjyqwfNdxgSM+8A43LZ7isgBRibT+0o4iROWF+ckARNEfDo8CbIvQZNhoxEx6hRo9ixYwcnT56kbNmyLFmyhO7du3P37l0AbGxsWLJkCc8//zxvvPEGXbt2pVq1akyePDnLPseOHUvfvn2pWrUqCxcuJCAgAF9fX5POQ6PRULduXby8vLh8+TIvvPACffv2pWbNmixYsID79+8D0KZNG27cuMFff/2V477c3d0JCQlh+PDhHDt2zKTjFyYKCoqiMullrYriBVWIwurR+jNaIL+el2mAYOBtMg+pys5XGGZLzak2jciZxAmJE0KYzTngClkLlUH+JGi2YRj/+iawkJwLkMVjCBTtyb42jciVxAnrjBPW+9sQAKSlpbFz505at25tXJaSkoJOp8PBIffO64GBgURERHDlyhU++ugj45Cg1q1b06FDB9555x1j7xWVSsW3337L/fv3+f3335k9e3aWglJ//PEHe/bs4eOPPyY5OZl58+YxblxuKfHMtFotJUqUID4+Hnd3d0aMGMH58+f56quvOHDgAL6+vvTv35+///6bQYMGZTvl9sNq1KjBokWL6Nixo7EQMYBerycxRcu9+HTuxqfzIEmDTqfPZU8WoAAqxbSXlSqKF1QhCqMk4BJQ86Fl9wGXfDiWHsMEG/7Au4/ZNgFYjaHmo8gDiRMSJ4Qwly3//vlwV8t7GL4pOpv5WH9jGP9aC1iLYfxtThZh6MljxVULLErihMlxYvv27VSpUgU/Pz9mzJiR43YbNmxAUZR87QAgCRort3//fqpUqZKpxktG/ZnHVeRWq9UMHDiQr7/+mnfffZfz58/z22+/ATBz5kyuXLnC4sWLjdvPmzcPvV5P165dadOmDXFxccZ1Wq2WoUOHMmvWLJycnFi6dCkNGzbE39/f5HPRaDS4uLhkKhKsKAqvvPIKa9as4fz58/j5+dGlSxdWrVrFmjVriIiIyHWfHTt2pG/fvnTu3JnklFRu3Uvl5JV4zkUkcDk6iSvRSVy8nsjxsAdE3EomNU1ncnvzW1Eu6qVHbryFKCi7MfRaf/i/U35NsT0HuIuh1/rjLAZ6AI750I5nhTnjxLx586hRowb+/v58+eWX+dzyx5M4IUQBCgXqAD4PLcuY6s+c/8euAa0xBKAt5J78SQXmYghg9c3YhmeMxInHxwmtVsuQIUPYtm0bZ8+eZdWqVZw9ezbLdvHx8cybN4+GDRvmzwn9yzq/3Qmj7IY3xcXFPXZ4U4Z33nmH9evXk5CQwNSpUxk7dix6vR47OztWrVrFhAkTOHPmDOvXr2fpjz+zNmQPoyd+yUdTv2HVL/sIj7qHVqfn22+/xdXVlaCgIDQaDbNmzXriIr3ZJWgeVqZMGcaNG8fly5eZOXMm7u7uVKtWjSFDhnDq1Kkc9ztx4kSqVq/FkTN3iLydQlq6Hp0edLqHXnq4fS+NU1fjuROX/QxRBa1IjxlFbryFKCihZK0/kx9TbK/FcL+9lMffXKRiKD8wxMxteNaYK06cPn2a7777jiNHjnDy5Em2bNnCpUuXCuAMcidxQogCEAscJPPwJjB/Jj8Ow7jXRGArUPYx2/8IRCO9Z56SxInHx4kjR47g5+dHxYoVKVasGMHBwYSEhGTZ7uOPP2bs2LGZ6r7mB0nQWDG9Xp9j/ZncCgQ/zMPDg9atW7N06VKCgoLQarXGGZKqVq3KzJkzGTx0ODcfqFmyeg9JWnviEjV4+lSibqPmXLudwl8X47gZp2f+VwtQFIXVq1fj6+tLo0aNnuh8NBoNTk5OpKen5zpFtkql4s0332TDhg2UKlWKUqVK0apVK1588UV+/vlnkpOTM22fptEzcOR07ByKk9tIJj2GRE3ErRTLJ2kUBcXGxqSXVVKKZtV1IQobPYbZS5s/stzcU2z/gaHXzEbAlNuWnzF8F8jvab6LNDPGiXPnztGwYUMcHR1Rq9W8/PLLbNy4sQBOIhcSJ4QoGFsBHVkTNOYMFGlAJ+AChkBR4zHba4DPMfScedVMbXgWSZwwvu7cuUO9evWMr4dHiURFRVGuXDnjex8fH6KiojId6vjx40RGRmYqK5JfJKxZsXPnzqHRaKhZs2am5U+SoAFDseCFCxcChqFN48ePJz09HYCGL77KlDk/Uq/Ry6CoeGRCJewdHAEVAe26obUvR0qqhhkzZuRpimutVoutrS1OTk459qJ5WP369fHy8qJevXqEh4czZswYVq5cSfny5Rk5ciQXLlxAr9dzITIRnd6Q2DFFRpImOU37xOdgViqVaS8rJU9Ghch/xzHUg3k0aWLO++6LGCbh2Ai4mbC9FpiPYWrtwuxaVBJrN19nbch1rkQkWro52TMxTuR2YwqGmm0HDhwgNjaWpKQktm7dSmRkpIVO6j8SJ4QoAKGAJ1mn3DNXoNBjCBJ7gCUYhiw9znrgMobeM4X4//jFy/Gs/iWS9aFR3LiZ/PgPWILECfQKlC5dmmPHjhlfAwYMMPkYOp2ODz74gDlz5uTjmfwnt7JMopDL6D3zaK2ZJ03QNGzYkJIlS7J9+3YCAgJ47rnn+P7772nbvguxyfY4OOZebBjARq0mTaPj+IUYXFxL0qJFiyc+H41Gg1qtxtnZmfj4eJPOYciQIXz99de0bduWdu3a0a5dO65cucJ3331Hs2bNCOwQTL/3JqKyebJ/6jo93Lqbhq/n4889PyiK8tgaQtYsY8yoECJ/PTp7UwZz9Vy/DXTG0BO9oomf2QA0AbzMcPz8kJqmY/Kssxw+fg/0oEeP6ieFmtVKMO2jGjg6FI6ei08SJzJuTHNSrVo1xo4dS8uWLSlevDh16tTBxsI9NCVOCFEA0oDtQBBZH9vfA8qb4RifAD/9+2cvE7bXAzMwzNrUzgzHzwcPEtIZO+U0F68koNPqUVSwcOkVXm5Sio+GV0GtLhwPUCVOmMbb2ztTsun69et4e3sb38fHx3P69GmaN28OwM2bNwkMDGTz5s3Uq1fv6RvwiMLxr0fkSXbDm+DJatCA4T/vkCFDWLBgAQAzZsxgypQpHDtzHXv7J0lQKOhRMXna/DwlFzQaDTY2NsYEjSm6du3K8ePHuXjxonFZxYoVmT59OpGRkXTv+x7kceq4mPtpaC05u5P0oJGbcyGe0q8Y6jE+yhw1aJKADsBn/Fe/MSVFy6atUbw97Bhd3znMlNnnCLuSYPyMHpgFjH7KY+enmV9d4Mjxe6Sl6UhL15Geric1Tcc/Z+8zcWbWooEWZcY40a9fP/766y/2799PyZIlef755/O58Y8ncUKIfHYAw1TW+ZXJX4ohMfM28LFhUXyChuUbrtFzyFGCBxxm9sKLXL/xUO+TbcBJYCyF9pvq2E9Pcz4sntRUHekaPWlpetLSdez/M4b531+2dPMykzjx2DhRv359wsLCuHr1KmlpaaxevZrAwEDjehcXF2JiYggPDyc8PJxGjRrlW3IGCu0/e/E4MTExnDp1ypjJe9iT9qABCA4O5ujRo1y6dInatWvzfLVauJfyfOJ22Razw9XjOXSPjoUywaM9aExhb29P//79jUO0MrXF1hbHEqVNHtqUhQLxSTnXwslvRbpIsIkXU1NuvPv27YuHhwc1amQ/oFmv1/P+++/j5+dHrVq1OH78uHGdjY0NderUoU6dOpkuxFevXqVhw4b4+fkRFBREWlraU5+yEAUtCrAFPLJZ97Q917VAd6AbkPE/Jz5BQ78Rf/H1D1cIu5LIjZsp7Dpwm4Fj/mbrrpsA7AAqA5We4tj56e69NPb+cSfbGf3S0vUcPxWX+YuEhZkzTty+fRuAa9eusXHjRt566638bPrjmTFOCCFyEArYkXXYkZ6nDxQ7gQFACwzTZStwOyaV7oOPsHRlBFevJXE9OoXQ/4umz/vHOHL8ruFz0zH03LHwJSgnYVcTCLucQLom63ed1FQdW3beJCHRct8fHiVx4vFxQq1Ws2DBAt544w2qVatG165d8ff3z7II4gAAIABJREFUZ+LEiWzevLlgzuXh9hT4EYVZbN26lddeey3bKtL37t3L1C3LFA4ODvTt25dvvvkGvV5Pl+7volbb5qltSUlJXI5IobLvkyV4tFqtyQkajUZDSkoKycnJBAYG0qpVK4KDg1EUxbg8NU1DGb8mqFR57H6nh3SthXrQKEqee/5YC3PdVPfp04ehQ4fSq1f2/Wa3bdtGWFgYYWFhHD58mEGDBnH48GHA8O/+xIkTWT4zduxYRowYQXBwMAMHDmTJkiUMGjTIPA0WooDk1HsGnv6+eySGJMvQh5Z9/cNlbtxMyXTTqtMZblhnfxNGgxdKMsPNjvlPcdz8dvr8A9S2KtLSs69BplLBiTNx+JS1zPDXTMwcJzp16kRsbCy2trZ8/fXXT9QTN79I8kWIfKTHkKB5DSj+yLokIJ28B4p/MBQFro6hnsy/Xyk+m3ue+/fT0T6UA9dqQavVMWHGWUIHNcbudxtDobK8fQ3Jd3+fisu1h72tWuFcWDz16xSCMvgSJ0wWEBBAQEBApmVTpkzJdtu9e/ea56A5kASNldqyZQtt2jw6capBXFxcjr0JcjNw4EBq1KhBuXLlWDlwIhpd3v7FOzgW56flP+HvVxoXFxeSk5ONSZOH/3x0WXp6Ol26dOHUqVMMHz4cZ2fnHD+r0+lwcHDA3t4eBwcH0tPT6dSpE97e3sblrq5uDP+48VP1E7PkvaHVztBkAj3mm3mjWbNmhIeH57g+JCSEXr16oSgKjRo1Ii4ujujoaLy8sq+Aodfr2bNnDytXrgSgd+/eTJ48WRI0wuqEAlNzWBdH3oc4fYmhd86ah5alpunYue92tk8UAdDDNztv4hRUgVp5PG5BUKsVckvLK4BtIaktAOaNEwcOHDDbvszBnHFCCJGNc8AVsh9zGvfvn3kJFNcxTKddAsOTghKGxXdiUzl94UGm5MzD9HrYNzuGlqXLQL88HLeAqG0UVIoCuUQLW3XhyS5LnLA+kqCxQmlpaezcuZOvvvoq2/X37t3LU0bz5MmTaLVa+vTp82+2Ne+9R0q4uvL222/ToEED3NzcjImUR/90dXXFwcGBYsWKsWzZMnr16kV8fDz169enffv2OX5OrVZnqnOze/duhg8fzuHDh43L9Xo9xy4+yDLzlMkUKGapG3EFsNbhSyYyNeOdUVU+w4ABA56o8npOU+d5eXmRkpJCvXr1UKvVjBs3jvbt2xMbG4urqytqtTrT9kJYkyTgElAzh/V57UGzEUOR3/8jc+477n4auZUeS0vXsedaEp/l4ZgF5TiwqIYrSRp9jsl5rVZPgxcKwVNRkDghhHg6of/+md3z3nv//vmkl7sHGLpuPsBQ38bnv1XRt1KwVSvkNGo8JUXLjbAUGMb/s3feYU1dbxz/hC2Ke6CgAsU9ita6sSp117211j0q1PWz7lq1dY+6xVW1at114FbUilq31tW6EAeiVQRFdpL7++NAZCQhQBjB+3mePEDuzb0nKOc9+Z73/b5gm8r7ZhJqwP7zQqjX6veZqVgub+YMKCXkOGGSyAKNCXL69GnKlStHsWLFtB5PiwfNhQsXGDhwIPPmzWPNmjU07zQ0XWPs0rkTuXjLvHnzWLduHS4u+vt7xMTE4OnpSYcOHfD19cXR0ZHatWsbfL/GjRujVCrx8/OjQYMGAERHR/Pm5WPyFnTA3CL1/9UVKLCzzaosFgUKEzYATpFU+Aak5CqfHh4/foyDgwP+/v40btyYKlWqkC9fvgy5l4xMZuKLyFrX9WeWFoHmL4TXoy+QtMAnr52l3pRvcwsFsfY21EvlPTOaWITotATIA3yXy5z9nUpyc9dTiE68zWtjbcZXTYtTIJ9VFoxUG3KckJGRSQf7gWokElE0pEWgiQU6A7eBg8CniQ8XLmilO8sSsMGMwliBZyrumUm8A54D5kDtdzZ8+V9RTuR9RbR58jjRv6cTVpbZZW6W44QpkoP/xXIuuro3xZNagebhw4e0a9eOdevWMXToUCIiIogMD035hTowU4CtjTleXl6MGzeORo0a8eDBA72vifefAVJlEhxPwk5UT548Yfz48ZQqVYqtvy3HLA3KsUIBxQpaZW2ra4XCsIeJklnmj/pa58V/dXFxoWHDhly7do1ChQoRGhqKUqlMdr6MjKmgq712PGGAXSqu9wCRcb4LKKzleC4bc+rVLKRzVaE2UzC0SeqN5zOKl8BPQFXgLLAW0TjkOuDfrTT9upUmVy5zctuKh421GZ3bODJsQDazN5bjRI5cnMvIZDjBwDm0Z89A6gUaCfgWkV65Cmia/JQS9rlwyW+LQodGo1ZBozZF0t9i0IgEAneB90B5oMzfYFYfxv5bjrZ1i2NtZUZuW3Ns4+LF4G+c6dw6m60Z5ThhcnFCzqAxMSRJwsfHhz179ug8JzQ01GCBJjg4mJYtW/LDDz9oPG08PT3ZuXkV3ft/T1q7TBeyE85eQ4YMwcLCgkaNGuHr66uzHVt8i22AvHnzplqgkSQJFxcXRo0axbFjx+jduzdnz56lTJky3Hn8nrdh0anKojFTQLH8WbhLqsCkW2gbQmZNlm3atGHp0qV069aNCxcukC9fPooXL05ISAi2trZYW1vz+vVrzp49y5gxY1AoFDRq1IidO3fSrVs3NmzYQNu2bTNnsDIyRkAC/gS9ZrwShntsvUZ4Pa4FXHWcEw5cG+SK1Z13KN/Hoo79EDysrM0I71qa7vZZb6x7CfF7+RsYCFxA2COoEcbHzwEfhQKrTqXo2caRf++/Q62G8mXsyGWTzXzB5DghIyOTVg4iJj5dSn5qPWimI4LED0A/Hef8CpN2lmdw/WtESyqUcdOXQgFWkhlj7pQl99qs/2iqQmxKAJQANHLLGYSgZQcWJxUMq+DKgAgn/n34HgtzBRXK2GGZbTJn4pDjhEmS9X8FMqnizp07KJVKqlTR5SxguAdNVFQUbdu2pW3btgwd+qGkqXfv3jg7O9Oj//9IfZKVRJF81omyVgYMGIC5uTmNGzfm2LFjVKhQIdmr4ltsg8igCQoKMuhuERERbN68mSVLlhAbG0udOnWoU6cOs2bNAiA2NpYFP/+PDr1GULhocQz5SGKmgLKOubPYCFJhNFOvfv36sX//fooWLcqtW7cAePPmDV27diUgIAAnJye2b9+e6rK49GBMU6/u3btz6tQpXr9+jaOjI1OnTiU2NhYQAmHLli05ePAgrq6u2Nrasm7dOgD++ecfBg8ejJmZGWq1mnHjxlGxYkUAZs+eTbdu3Zg0aRLVqlWjf/9s7FYnI5OEq0AlIHmPP0FqxJlIoD3wI1BHxzkvgXZA74LWdF1Sg+Z7nlLE9yUxUSpcnXPzrmtpvqpeMMtM12OAHcAyoCDwHbCBD9FNCQxAlG1tRqSwA1hbmfFppWy0lZsM48WJ7EhONX+UkckW+AD2wGc6jqcmg2YTQpjphaiDTYoU9/xUcGqSmw2darB5wVNOlPsPpblEFZd89Flbisod8oH2/g2ZwlvgBSI2lCHJJ6CDQCegJKJ9eCnxtK2tBdWryHEiq8ipcUIWaEyM/fv307p1a52lN0qlkoiICOzs9Cevq9VqevfujYODg0bMiCd//vx07dqVw7vX06BFD6ytdS3zEyNJat6GhqB+G0xp+88THevbty8WFhZ4eHhw7NgxKlWqlGzcCQWad+/e6b1XQEAAy5YtY/369dSpU4f58+fz5Zdf8s8//+Dh4cG0adNQqVR07doVpVKJe/VS3LgfjEphhY2N9l1cMwWYmSko62hLnlxZ/KehwGht8bS1oZ41axYeHh6MGzeOWbNmMWvWLGbPnm2U+xmEEdMNt2zZov9WCgXLli1L9nzdunW5efOm1te4uLhw8eJFo4xPRiazSam86T3CbyUl1MA3iOyZDjrOuYtYs85EbCyeyWeJY28XdvUWvmMvgC+B9YYM3MgEAd6IblMtgN9IngEUCXRDWCVMJWs796UaI8aJbIkJpqXLyJgEMcBhoCu692HjBZqUtIcTiIyZRsAakk+iscBgYB3QB1gF9u1s+F9wGf63pYxopf09ouRKWzepTOApwlg/L1BO2wmbEWOvivi9FcmskRkBOU6YJDn4XyxnkpL/TGhoKPny5cMshXS2cePG8fz5czZs2KD13Lp16zJ54miC/K8abP5taWGOOiyANq1bsXHjxmTHe/Xqxbx58/jyyy+5ceNGomNJBRptJU6SJOHr60u7du2oUaMGkiRx4cIF9u3bR5MmTVAoFFSsWJEKFSqwceNGmjVrhp2dHXv37sXK0oK+Xb8kKvg+ZqpwYmNiMFOApFahUiqxtTbDpbgtbq52WS/OAKCIU4wMeKRAgwYNKFiwYKLn9u7dS+/evQGRMaWvZC6jyIk1ozIy2YEDiCYaujDUIHgMIr17hI7jZxHZNb/ywcZgBcKGIJ5fgGFk3m6QhDAz7oEQZeyBy3HjSCrOvEV0gm0MTMPExBnAmHEiuyLHCRmZDOA0wohMn5IfglAs9CVf3Eao92URbutJnQHeIYLDOkQa5q+IFtyHEDWmlnH38UaIRZlo8aVEbDDcRWRWlkNH8s4S4GugPnAS0xJnADlOmGackAUaE+L169fcvHmThg0b6jzHEP+ZFStWsHfvXvbs2YONTfLsmJ07dzJ69GiR5RIbRtmSubEyl4iKjNAYp8YTHRVJTEw0+XJbUNkpD82bNubkyZNMmTKFcePGoVKpEp3fo0cPFi1aRLNmzbh+/brmeZVKpfGgSSrQvH//nhUrVlC5cmWGDx9OixYtePz4MfPmzdPaHerrr79m+PDhVK1alY0bN2JpacmiRYsoV64cTT3cOX/yD35b9gNVXewoYPWWfl2+oLKzHQXzWmKWTUyyFIBCYWbQI74Ndfxj1apVKV7/5cuXFC8uQpG9vT0vX77M4HeUnJw4ocrIZDWBiDVvUT3nhJLypuhS4CGwQMfxncBQRBOQ+HzJV8ANhOARf5+9QO8UR51+ohBlS7WBWQhD42sIsUhbttB/QBPExu/wTBhfRpCaOGGqGDNOHD58mHLlyuHq6poscxhE58euXbvi6upKrVq1CAgIAIRXX6NGjciTJw9eXl6a8yMiImjVqhXly5enUqVKjBs3zhhvWUYm49mPqIH9Us85KQWKIITCbYso/0l67nOgAaLt3xpgCmLSWhX3dWDceUsRaZ2Z9OcTghBlAhC6Ujkgt7YTJcSYhwFtEaJSNumcnRrkOGGanyeyQ6qAjIEcPHgQDw8PraJKPCn5z+zfv59p06Zx5swZChUqlOz4L7/8wvz58zl69Cj3799nyZIldOnShT8PbuDho2dUql6fMuWrYGZmztvQN/jfu8ma5fNZvmwx5Up+AUClSpW4cOECnTp1on379mzevDlRyVWXLl0wNzenefPmHDx4kOrVq2vNoHn48CHLli3jt99+o0GDBixdupSGDRvq7awUEBDAzJkzMTc3p1+/fpiZmREUFMScOXM4f/48AH5+frRs2RIrSzNcnEryyP8B0dHRWFtb6/8HyEwUGKxmp7cNtUKhyPRuVRKmN1nKyJgCB9DdlCOelDJo9gFbEGX22jZPf0Fslp4AEkaRX4G+fNj5WYbwdsnImfUZImtnJ9AG2Ao4p/CaAMR6+2f0byBne1IRJ0wRY8YJlUqFp6cnx44dw9HRkc8//5w2bdpofMcA1q5dS4ECBXjw4AFbt25l7NixbNu2DRsbG3766Sdu3bql8XGLZ/To0TRq1IiYmBg8PDw4dOgQLVq0MM6gZWQyAglRB9sYIa7oQl+geI9I0wwG/ND4sWi4jUhhDEGIQc3jno9GGAm3RrT2DgcWxV2rairfRyp5gihpzY+OMqaEqBHK/VJEadNqTPcTsxwnTBLTlcs+QlIqbwL9LbYvX75M37592bNnD598kjiPUK1WM3LkSNasWcO5c+dwc3OjXbt2PHz4kL///htvb286dWhNHTdnureuQ/fWtSlVCIrlt0CBiqVLlya6XuHChTl69CjFixenbt26PHr0KNHxjh07snLlSlq0aMGlS5c0Ao1arebGjRvcvn2b2rVrY2VlxZUrV/jjjz9o1KiRXiHh9u3buLu7M2zYMMaMGcPy5csBUc41cOBAXF1dkSQJPz8/3N3dAbCwsKBUqVL4+/vr/b1mPgpRM2rIIw0UK1ZMY8QcFBRE0aL69tszhpyoeMvIZDU+pE+guQRMAnaTfO2uBkYiSpuOklicUSM8XuKzZSKAjQjrAWMjIT4TdEGIMk6IbJm5pCzO3EF8FliCiYszQEbHieyAseLExYsXcXV1xcXFBSsrK7p168bevXsTnZOw9LdTp074+voiSRK5c+emfv36yTbHbG1tadSoEQBWVlZUr16dZ8+eGeeNy8hkFP8A/qQ8AeoKFEpEOdINhPt6tSTHTwH1EN4zp/kgzoAILK/4UAe7BiHyjDd49KkiFpEt8y9QGCHMFDPkRV8jxJn/IXYeTFWcAeQ4YZqfJ0z6v9zHRExMDMeOHUsmhCRFl0ATEBBA27ZtWb16NbVq1Up0LDIykl69evH69WvOnDmjeb2lpSVDhgxh0qRJmJub4+7ujkKh4JNPPuG///6jWrVqODs7M3jwYAIDAwkMDMTBQdOMDisrK7y9vVm2bBl169Zl27ZtNGjQQHO8bdu2mJub06pVK+bMmUNYWBgVKlTA3NwcKysrnjx5Qq5chrVlvXDhAm3btmX+/Pn07NmT//77j3LlytG5c2eOHz/Ov//+C4C/vz8KhQJn5w/LeFdXVx48eKC1u1SWkoGu623atGHDhg2MGzcua9pIK3Km67qMTFYSgWgNqrvHnyAE7ZnrjxACyx6Sl0hFIhp0lESY7iadnY4CNfgg2qxFCCj67epTRyTwOyJjpjSiG9MXGO4dcwGR4bOZ5J8pTJYc3J0jNXEivtQ3nkGDBjFo0CDNz4GBgZQsWVLzs6OjIxcuXEh0jYTnWFhYkC9fPoKDgylcuHCK9w8NDcXHx4fhw021YE7mo8En7mtKSn4oyY27JMALUdK0EpElk5DfEZPsJ4iSoNJJjq+IO9YEYVQ8D3BHCDpG5A1CB7JAlDEZ/Nk8AuF6fwjhfD82NS/OxshxwuSQBRoT4fTp05QvX55ixfRrv9o8aEJCQmjZsiVjx46lXbt2iY4FBwfTpk0bSpcuzZEjR5KV+QwcOJCffvqJn3/+GYVCgUql4vnz54SEhPDixQvs7e2pW7cuCoWClStXMm3atESvVygUeHl5acSS6dOnM2DAAM3xsmXLUrt2bfr160euXLnYuXMnFStWxNXV1WBx5tixY/To0YP169fTqpWwxixatCitWrViwIABzJ49W1Nidfr0aRo0aJAoEydeoMlWKBRGU7O1taEeN24cXbp0Ye3atZQuXZrt27cb5V6pwdTUbBmZ7I4v4EHK68lQkm+MvkEY/q5CLGgTEhx3rD0ig0YbK/iwCRob9/Npg0adMo+B5QjhqAOivCppRn1KHEOMfS+ifWqOwIhxIrtiaJxIb6lvelAqlXTv3p1hw4Zp9cWTkclW+CAUascUztOWQTMHIcyMBwYleF4CZsc93wAxWSd97W1EUJiDqN/YjKhRTdk20WACEH5k8aa/qSIUIVqdQ7zHQfpPNxnkOGGSyAKNiWBIeRMk96CJjo6mffv2NG/enGHDhiU619/fnxYtWtC+fXtmzJiht/NTvNnvzp07cXBwoGPHjkydOpUVK1bQrl07Dhw4wKpVq5g4caJWL5cmTZrg5+dH69atuXnzJh4eHixfvpxr164xYMAAateuzeTJk5Ekibx58xIWFoYkSSl6o+zYsQMvLy92795N/fr1Ex0rVaoUr169onv37prnEpY3xePq6qrJsMlWGKlmVFcbal9fX6NcPy3k1JpRGZmsxAfREjslQgCHBD9HIYSPCYhGFQnxRwgzPyA2FrXxBOEHGZ+b+TvQFJFSnlYkRKb8krjrDwWuA4bJ9onZCcwAjpD4fecIZG8Bg3BwcODp06ean589e5Yo4zfhOY6OjiiVSt6+favVqy8pgwYNokyZMowYoavfmYxMNuE1os3dJAPOTSrQbEEY+XZHGHjFo0QY6a5AlD5tQLvxmHfc830BFULQcSNxCVQaiEHEKRAJO2mJEQTFjeMfYDu6g52pIscJkyNnS2o5BEmSUiXQxGfQqNVq+vXrR6FChZg3b16i8y5dukT9+vUZPnw4s2bN0inOrF27lubNm/Prr7+iUqmYOXMm48ePZ8KECezYsYN79+7Rtm1bzp49S8WKFdm1a5fOsRUrVozevXuzevVqevXqRZs2bXj8+DHTp0+nWbNmODs706lTJ86ePYuFhQVRUVF63+uqVasYMWIER48eTSbOhIaG8uuvv+Li4sKRI0c0z2sTaMqUKZP9MmhArhk1wZpRGZmsQgL+BBoacG7CdbcasV5uhShJSsglxIbiMvSvV1cjGnIo4q73C6J0Py2EIzYvqyPW86PixtGPtC28VyM6UR0jB4ozIMcJA+PE559/zv3793n06BExMTFs3bqVNm3aJDonvvQXxGZU48aNU9wkmjRpEm/fvmXhwoVpfo8yMpnGIcQknVJ5Uwyi3Cc+UJxGmOU2QLTMjp9SwhHq/gpgDEKd1ybOvEeYlHVGKPd7EOYw40hzCVFw3CWeIbJlypNGccYfsTPxEOGyn9PEGZDjhAl+njDdf42PiDt37qBSqahSJSVngcQCzQ8//IC/vz+bNm1KJMDs37+fli1b4u3tzdChQ3VeS6VSsWrVKn788Ufy5MnDjBkzUKvVtGrVikKFCjF69GgmTpxIiRIlKFeuHO7u7ixbtkzr+IcOHYqzszM3b97kyJEj9OnTh0WLFvHkyRNApAgXLFiQ7du306VLF6ytrRO12k6IJEnMnDmTWbNm8eeff/Lpp58mO2fatGm0bt2aMWPGaHx7Xrx4QXBwsGgfnoBsWeIEcWmJBjxMEQMnU1ObUGVksoqrQCUM65iUsHvqREQ6+Ogk5+xHiCK7SZ5Vk5BYxIZjj7if9yLElZI6X6Ed/7gxVAeeIrKBtsXdOy3TgIRot72T5IbGOQo5ThgUJywsLFi6dCnNmjWjQoUKdOnShUqVKjF58mT27dsHQP/+/QkODsbV1ZUFCxYkasXt5OTEqFGjWL9+PY6Ojty5c4dnz54xffp07ty5Q/Xq1XFzc2PNmjUZ9duQkUk/PoA98FkK54XGfc2PcNhtB7ggAkJ8kPkPaIQQNZYiMmJ0farcArwDhiAm55kIf5tUiiESwivtbtz35eKGleYZ7ibC/yYUUSPcJK0XyubIccLkPk/IJU4mgI+PD1999ZVBrZBDQ0PJnz8/q1evZvv27Zw7dy6Rl4u3tzdTp05l//79ycyCk3Lw4EHs7e357LPP8PLyYsyYMSxZskQzjuHDh1O2bFkuXLhA+/btefToEU+fPuXq1at8+umn7N+/nyVLlnD79m0GDRrErVu3KFGiBADu7u6sXr0ad3d3Nm3aRO7cuTE3N6dhw4bs2rWLhg0bcuDAAfr27ZtoTJIk8f3333PkyBHOnDmjuV5C7ty5w8aNG7l9+zZ2dnZ8//33PHjwgGvXrlGvXr1k2UJOTk4EBgYSExODlZVVir/jTEGhAD0lZzkBU5ssZWSyMz4Y3pUoPoNmJXALseZO+Oe4ErFJegyxltfHXuBLIA8fbAg2GDgOCTiOKGN6CXgiMudt9L3IwOuOQXjX7CNj23xnKUaOE7/88gtr1qxBoVBQpUoV1q1bl6xzUWZjzDjRsmVLWrZsmei5hL55NjY27NixQ+trAwICtD4vSZLRxicjk6HEAIcRZUgpTRshcV/NEEbAlghj4IJxz9+Lez4IYQqmr8+EhMiwqQLURUz6VxDeMwZ610YjhBkFomufUeb0c4jUUVtEhlAl/aebLHKcMElkgcYE2L9/P5MmGVIwKjJoHjx4wJIlS/Dz86NIkSKAKHeaOHEiu3bt4syZM8nabGvD29ubb78VvfCcnJwIDQ1N1CUhV65cTJkyhbFjx+Lt7U3jxo3p378/Q4YM4dWrVxQtWpTvvvuOzp07a/WlGThwIGXLlqVr165069YN8ziXcXd3d5ydnfnf//5HsWLFNAsqpVLJwIEDuXv3Ln/++ScFCxZMdk1Jkhg+fDiTJk3StI7u168fK1asIDY2NlEXqXgsLS0pWbIkjx49oly5VNuKZRxmOdd1XSJnuq7LyGQVB+IehhACXEaIML58WAhIiIyaG4g1dB4DrrUCWBT3/UmE72RKs2h8tvsqxJp4AsK/xhhrLCXC29ESsWmbc2fROIwUJwIDA1m8eDF37twhV65cdOnSha1bt9KnTx+jXD8tyHFCRsaInAbCMEzJjxdoFiIyZf4EnOOeOwe0QUzYJ/lgPqaLS8A1hNO7ApHaWAL4JuVhvEaUMlkjSpiMxmFEaZYjIsXSyZgXz4bIccLkkAWabM7r16+5efMmDRs2NOj8wMBA5s+fz/79+ylTRvSqiImJoV+/fjx8+JCzZ89qRBt9PHr0iAsXLrBz504AFi5cSOPGjVm9ejVz587VnNe7d2/mz5/PiRMniIyM5JdffiE2Npb9+/fTpEnKuYJffPEFf/31F40bNyYmJobY2FgsLS2xt7fn+++/p2/fvqxZs4YmTZrQrVs3oqKiOHbsGLlz59Z6vb179/L8+fNEpVvffvstNWrUoHjx4qxevVrr6+LLnLKNQJPTM2hMMN1QRia7EogQJJK2xtZFECJT5TgQP5PGIEqa8iDsAQxZHNxFlDhVjvt5JmLtrYv7CD+bw0A3hB1CcQPHbAhRCP/KCsB0ckZ3VL0YOU4olUoiIyOxtLQkIiJCa4ZqpiLHCRkZ4+GDSE/80oBzX8d9fYgICPF7s7sR9ayOiAk8aRtubaxABJavgYvACUR7bR1pMPFlTLEIuxqjr8q3Ar0QgeswoL85rukjxwmTJAeTkupeAAAgAElEQVR/AswZHDx4EA8PD4PSx548ecK9e/eYOnUqdevWBUTJU/PmzYmIiMDX19cgcQaEAe8333xDrly5uHnzJleuXGHRokWsX7+eiIgIQPyR7tu3D3Nzc4YPH07lypXp27cvXbt25e+//zb4PTo7O7NgwQJiYmJo2rQpwcHB2NnZUbJkSfbv30///v2pUaMGNjY27Nu3T6c4ExUVxahRo1i0aBGWlpaa552cnDQGgdWrV9f62mzpQ5OTa0bJmTWjMjJZwQFS9nyM5zFC0NnBh/Klt0BLRDbLCgzfufFGWAqAyMhRkNzaQI1Yx7cEegM1ERk6UzCuOPMOka1eH9Gx6aOZOgyME69evaJGjRqax6pViXvbOjg4MHr0aEqVKkXx4sXJly8fTZs2zaI39QE5TsjIGAEJIdB4IEp6Ujp3cdz3E/iQcbMY0SbQDZFFY4g48wYhiHwN2CFU/AJobWEdibC7uceHTEyje4etQAhMdRFtAnO6OBOPHCdMLk7IAk02x9DuTaGhobRs2RIrKytNW+mnT59Sv359qlSpwo4dO7C1TWlWFkRHR/Prr78yePBgAGbNmsWIESOoWLEitWrVYvXq1cyaNYtPPvmE+fPnM2HCBD7//HM8PDw4duwYXl5eLF++XNOa2xBsbGz47LPPqFWrFjVr1kStVhMWFoaTkxOFCxfG39+fDh066PWImT9/Pm5ubnz5ZfLtgXr16mFpaZlIuElImTJluH//vsHjzRRk13WTm1BlZLICQ/1nQhEts+0RWSYgDHkbIxp0jMdwYSMCYUkQ39Z7Ztzr43mHWM+7IUqNpiLW9D0AYzt9vUK09e5F2rtHmSwGxokiRYpw+fJlzWPQoMSfkEJCQti7dy+PHj3i+fPnhIeHs2nTpix6Ux+Q44SMjBH4B5GWYoiSH9/2DoQxmBoxsQ5HlDb5Aobt9QpDsiiEkn8HkY3zHUKsieM/hDDzElHGVA7jxwgk4CdgKOJ3cBjIZ+ybZGPkOGFyccJ0P919BMTExHDs2LFkpnbazuvYsSMNGzYkOjqa/Pnz8/fff1O3bl369evHwoULNf4uhvDHH39QuXJlypUrh7+/P0eOHGHIkCFcv36dmJgY/ve///Hvv/+ye/duzp49S/fu3Zk7dy7r1q0jLCyMvHnzUqhQIQ4fPmzwPZVKJZaWlsyaNYspU6Zw+vRpDh8+TP369enYsSPnzp1j+PDhbNu2Tevrnz59yoIFC5g/f77W4+/fv8fa2pqzZ89qPZ79MmgMzJ6RM2hkZD5qIoAHfCgz0kU0QkwZxofOqTeA5sBcxAZnatiOaOxhjVhcByFafP8LeCFsCd4hyqh+Az5P5fUN5Qmi8cZ4hMj0cWG8OHH8+HGcnZ0pUqQIlpaWdOjQgXPnzmXCe9CPHCdkZIyAT9zXlASaHYh2evEBJReiHnUBYmLfRcoZOPFIiDTLOsCnCAd5W2CYOPQAUSZrgRBmnAy8bKpRAyOByQgVfxdp7MdtqshxwhTjhCzQZGNOnz5N+fLlKVZMdw6eJEkMHDiQPHnyMG3aNHLnzs3Jkydp0qQJ8+fPZ9SoUQZ1f0rIihUrNObAs2bNokGDBrRq1YrWrVvToEEDSpYsyeDBgxOVC9WrV49q1arh5OTE7t278fLy0rS3NgSlUomFhUis79WrFw0bNmTDhg2UL1+eadOm4ebmxtGjRxk5ciS///57stePGTMGT09PnJ2dtV7/zJkzdOnSReeYsptAIylAMjc36GGKSAph6mXIQ0ZGRje+iKx1fbO8BAxAWA80Qwg0xxF+LVsRGTSpZSUwOO77mXHXaAYMBNwR4s8kDPfFSQv/IBqJLER/E5GcijHjRKlSpTh//jwRERFIkoSvry8VKlRI8XUZiRwnZGSMhA9QDVE7pIuzCAGjHiJY2CAEnR0IFX8xqXNdP4moV/oWUVv7O0R4wr+FxNOlEdkyydt9GJFYhHK/CBgBrEcYtn1EyHHCNOOEbBKcjTGkvGnKlCn8888/nDp1ilevXmFpacnXX3/Nrl27cHd3T/U9b9++zYMHD6hTpw5jx45lzZo11K5dm1GjRtGuXTssLCzInTs3S5cupU6dOoleO2PGDOrVq0dUVBRnz57l+++/5/79+xqzYn0kFGguXryIn58fTZs25enTp/Tt25eVK1dSpUoVjh8/TpMmTVAqlXzzjbCA9/Pz4+zZs6xdu1brtSMjI7l27Rrbtm2jcuXKBAUFUbx4YvcDZ2dnnj17pjEpznoUJl2+ZAimpmbLyGRHfPhQZqSLyYgNw3GIDJd3cd8fQf96XRdXgfyIhfUPiGyabogN0mppuF5auITwtNlIct+bjwfjxYlatWrRqVMnqlevjoWFBdWqVUuW3p4VyHFCRiadvAb+QijmuriHULlLAXsRpUCxCFPfrYjW3KllBcJEpjO8nAah5SDXSCN3Y9JHJGLcPojypol8ROZkCZHjhCmSsz8BmjCSJKUo0Kxbt46NGzfi4+NDrly5mD17NmFhYZw6dSpN4gzA1KlTKVKkCJUrV+bQoUN07dqVc+fO0alTJ42A0qdPHw4ePMjLly8TvbZixYq0b9+eO3fuEBwcrGlvbQgqlQpzc3N8fX356quv6NSpE1WqVMHPz4/379/TqFEjXr58ScWKFfH19WXChAmsW7cOlUrFd999x9y5c3V67Fy8eJHKlSvj4OBAt27dkpleAVhZWVGiRAkCAgJS9wvLSGQPmhw56crIGAsJ0Tm1oZ5z1gJXEB1OQWSqv0Bk3qRFnAFhwishMtd9EQ051pF54owv0BfRUOTjFWfiMGKcmDp1Kv/++y+3bt1i48aNWFvraLOSichxQkYmnRxClPno+jjxCpGKaBZ3bgBCpJEQXjRpEWeCQNoDD0bD3XCw/hXK1YJSxnSG18dbRErnfkTwm8RHKs7EIccJk4sTpvvpLodz584dVCoVVapU0Xr82LFjjB8/nkOHDlGwYEEGDRrE8ePHqVGjRqrTzWJiYvj999+pWbMmu3btolWrVly5coXAwEBmz56d7PwCBQrQuXNnrS2rf/rpJ9RqNRs2bGDIkCFs2LCB8PDwFMegVCoJCgqie/fu7Ny5k1q1ahEWFkbu3LnZvn07TZs2pWbNmly/fp3y5ctz4sQJJk+eTO/evcmXLx9dunTReW0/Pz+NYOXp6cnKlSuJjY1Ndl72K3NSGPQwVYw1ofbr14+iRYtSubJ2Fw5Jkhg2bBiurq5UrVqVq1evAnD9+nXq1KlDpUqVqFq1aiJ/oz59+uDs7Iybmxtubm5cv37dKO9ZRsaYXEV0XtK1PDqCKEXahlhrDwL8gc6k3h9RhfB3/AKxhu+P6JYaitaGHBnGHwiLhENkQPtVE0SOE6a38JaRyVR8EM7w2pqYRiKMf58D+xCmMF/EHfsUaJD624UDd3fDfWdw6gTlFkD+/4Axqb9WmniJeA9/Ab8jSqw+cuQ4YXpxIksEmtfB0aze9IjWX5/Dvc2fNGx3ml6elzh84iXRMeqsGFK2Iz57Rpt/zI0bN+jZsyc7duzAwcGBNm3aEBgYyI8//qjXryYpL168YOrUqTg5ObFmzRpq1KhBy5YtmTFjBlu2bKF169aUKlVK62s9PT3x9vZOJnQ4ODjQqlUrli1bhpOTE+7u7mzevDnFsfj6+nLp0iWOHDlCgwYNsLOzIywsDAAzMzOmTJnC3LlzadKkCX/88Qdly5Zl9+7dbNmyhfr16+v12Uko0FSuXFnz2qRkK4FGocjRGTQSxptQ+/Tpo9eQ+tChQ9y/f5/79++zatUqjb+Sra0tv/32G7dv3+bw4cOMGDGC0NBQzevmzp3L9evXuX79Om5ubnrH0KtXL4Oek5ExJj7o9ny8jmi8sQexcdgWsUbvQ+pal74B5gBVEF2b6iAMeTsjNiaHknkl/WsRVgjHgJKZdM9sjRwnTGbhLccImSwhBtGx6CuSf+JTAT2BCwgh4yZCrCkLlAEcUnerIIS+80YJ5WZCWSewKAosQ9ThZoaiHgDUB+4jAmS3TLhndkeOE1kWJ9Iz72f6v8bB4y/oMugiW/54SsjbWCQJlCqJR08imL/iPp37nyfgacoZFzkdHx8fvvoq+dL72bNnfPXVVyxevBhXV1e++OILHB0d2bdvH9HR0RQoUEDL1T4gSRLnz5+nZ8+eVKhQgaCgII4ePYqvry/nz5/H09OTiIgIFi9ezNixY3Ve59NPP8XZ2Zm9e/cmO7ZkyRKCgoL466+/8PT0ZNmyZUiSpPNac+bMwcfHh1atWlGtmkiSTyjQxNOlSxcOHz7M8OHD+fnnn1m/fj09evTg999/Z8mSJVqvrVQqOX/+PPXr19c8p8vAONu12s7JXZwMnEwNmVAbNGhAwYK6beb27t3LN998g0KhoHbt2oSGhhIUFETZsmU1/kglSpSgaNGivHr1Kk1v5/bt24l+VqlUXLlyJU3XkpExlANAKy3PP0Wsu7cigrwHor32T4jMb/1RQvA3wli4PkLgOYPIxjmMyJ55h/CO7Jeud2A4cxHv5yhQOJPuaRJkozgRERHBTz/9xMCBAwG4f/8++/fvT/sFjRgnsho5RshkCaeBMLSXN41G1IkuQKRjDkK0xDsNvEcYjaWAGqGF3EU0aCoHlDwAPENkrqxABJ3x6XoXhnELYXAcjHDBb54J9zQV5DiRJXEiPfN+pgo0x0+/ZP6K+8TEqImJTf6BPTJKRcjbWL4dc50X/0Vl5tCyFa9fv+bWrVs0bNgw0fPv3r2jVatWeHp68umnn1KnTh3at2/PqlWrsLCwICQkhPz5tc+o0dHRbNy4kZo1a9KzZ08+++wz/P398fb2pnLlyly8eJG3b9/StGlT1q5dS7169VIslfLy8mLZsmXJnndwcKBChQp4eXnh4eFBdHQ0Z86cSXaeJEmMGTOGDRs28P333yf6kK1NoAH47LPPuHjxItu2bWPNmjXMnDmTkydP8ssvv/DLL78kO//vv/+mZMmSFCr0Yc+4bdu2+Pv7c+PGjUTnZqsMGhQ5uosTGO66/urVK2rUqKF5aPMQ0kdgYCAlS37Yb3d0dCQwMDDRORcvXiQmJoZPPvlE89zEiROpWrUqI0eOJDo6Wuu1Z86ciZ2dHTdu3CBv3rzkzZsXOzs7ihYtStu2H2NfGePy9HkEC7zv03/kFUb+cIOTZ1+hVOkWez8mAhGZK0m7JL1FiDFLEA03mgA/IrorAYSgW6BRAjsR2eH/Q6zpbwLfIwyB/RAbq8URYs03GN5xNa1IwFjgPMJOwC6D72daZK840bdvX6ytrfnrr78AsRaYNEmfM2nKmHp3DjlGZDz3/d8zfeG/9B9xhbE/3eLC1Td6NwU/KnwQ3Zi+TPL8IkT7u+8Q4szPCOV9H5AH/YECofncRbTKdkEIM5qyWW9E9s2XwC9AU7SXVxmT84hyrHhjtjr6T/+4kONEZscJY8z7mdbFKTpGzZyl91MsYZIkCA9XsmTNQ6ZPqJRJo8s6YkPf8cbvEjHBoUhKFZYF8uIbcBcPDw9sbGw+nBcbS+fOnalbty61a9emYcOGzJ07V9PJCCAkJCRZBk1gYCDe3t6sXr2aqlWrMnnyZFq2bIl5kj9Eb29vBg8ejEqlYt68eezcuTPFsbdv356RI0dy69atZP4fw4cPZ+TIkZosmqVLlyYyLlYqlQwZMoRbt25x+vRptm7dqjEhBt0CDYC9vT0FChSgatWqtG/fnj179nDq1CkaN26MSqVi9OjRmnNPnz6dzDDZ0tKSwYMHs2zZMlauXKl5PlsJNApMNt3QEOJTEg2hSJEiXL58OcPGEhQURK9evdiwYQNmZuJ3PnPmTOzt7YmJiWHQoEHMnj2byZMnJ3vt+PHjNY+ZM2dm2Bg/Rk74/cf0hXdRqdQoVeK5W/++5ROnPCz6uSrW1qYrTqaV+A8dCoWCAyQvb4pFlB4NQ4g3HUne5UjbuvsVsDru3IaI0iVt0XcFYlM0ClFudD7N78QwVMAQxHyxndR1eP0oyGZx4uHDh2zbto0tW7YAoow0PR+UUxMnsityjMhYtu55yupNAcTGqlGr4e7D91y9EUKt6gWZNrYiZmYm/h8oDWjiBAoh0HiQWEnfDYxEBJDbCDOxaXww0lWhM9XyOUKcyYOOiiV/hPnZj8AmhB9MRmfPHEXsShRH1L86Z/D9TA05TmQ6xpj3M02gOXnmFYb+/tUSnLscTMjbGArks8rYgWURb6//w6OFvxK06whmVpZIKhVIoDA3xy48nF4uDrw8cJKizRuAmRlDhgzB0tISd3d3OnfuzObNm2nSpEmia4aEhFChQgUkSeLs2bMsWbKEo0eP0rNnT06dOkX58tqb271584bdu3czZ84cfv/9d8qUKcPnn3+e4nuwsrJi0KBBLFu2LFm3pvbt2zN8+HC+//57Dh48yI8//sjz588pUaIEUVFR9OzZk7CwMI4fP06ePHkStdkGIdC8e/dO63137tzJ27dvuXz5MvPnz6dWrVr88ccfGpEmNjaW8eNFRPDz86NTp07JrjFw4EAqVKjArFmzNKKWi4sLT548STaWrELKRhNqRpBZE6qDgwNPnz7V/Pzs2TMcHERxdXxW2vTp06ldu7bmnPg27NbW1vTt25d58+bpvcfMmTMJDAzk8ePHKJVKzfMNGqTBYU+G12+imb7wbjJBPzJKzT3/MNZsDsCz3yc6Xp1zkCSJ0At/4//Lr7w+dgZlRCRIYG6bi5hGtWgyqj9Svc9QKBQaE2B3xIbpMIRnjFOSa4bwIXP9KiLT5hJi8/QvdGe1v0Ss5RsBq4B2es41BtFAD8AVmMXH3YBDH9kpTlhZWREZGanxhHv48GG6O3xkt4V3WpFjhPG59zCM1RsDtMaJ81fesO/Ic9q1SKWRigkiqdW8PvEX/vPWEHL+GqrIKBQKM8xz5aLYOw+ce/YlH3EZ8ecRE6sb8AiRBrMe6J3ggvFL7ziBRg08jPtaHCihbzArEbUZfRDBojYfTIczgu3A10BFRP2tfQbey4SR40TWkJ55P9M+hR44HkRklMrg8y3MFZy79IZWX+asv7bYd++50smT0AvXUUfHIKnUqKMSl09YADx4xrVe/8PCLg/X29Xk2rVrdOzYkXHjxnH8+HGqVq2a7NqvX7/mxo0bVK9enffv3+Pl5cWqVavIl09/v44NGzbQqlUrChUqxOzZs1m8eLHB72fQoEFUqlSJWbNmJbpPkSJFqFmzJo8fP+bPP//UtLf+3//+R7t27ShUqBA+Pj6aP0ptAo22DJqIiAhGjx7Nb7/9hqWlJePGjaNChQq0atWKxYsXc+rUKTw8PFAqlUyaNIkzZ86waNGiZNext7enRYsWrF+/npEjRwLiw7i9vT2PHz9OVOqSNZiwv4whZGI9aJs2bVi6dCndunXjwoUL5MuXj+LFixMTE0P79u355ptvkol4QUFBFC9eHEmS2LNnj84OUfGMGzeOrVu3UrFiRU12mkKhkBffacTnSJBOQT8mRmLv3iAGv3fBoqRCpFI7AMXIxIiW8bw8eIo7o6YT/eIVqshoUH/4EKJ6H07p/ScJPvEXJ4sUpMLc8axq1wQFkAuRAeOLKEtKyhuEEDMGyI0QctaScr3zrwi/GRVC1DmezvenjzCgA6I8K7Maf5gm2StOTJ06lebNm/P06VN69uzJ2bNnWb9+fdovaCL+MoYgxwjjs23vM2KV2rPyo6LV/P7Hsxwv0Dzd8Ad3f/gFZdh7VO8jNM9LqFC+f89zhQ9BC4+Qe38pKo35gULDagoTrxcIn5mDiIk2ISHiy7siwvjXDPgEAzwxohGBoi3CtOwRoowqo/6GVyJSOushMoUycsfApJHjRFaRnnk/05azoW+TtzXWR3SMin/+fUzNT80pXLiw3i49pkJMcAhn63cl6mkQ6uiYFM9XhYWjDAvHecU+Gjf7lK1bt3Lu3DkcHR0TnffkyRNWrFjBnj17qFKlCjNmzKBZs2aacg19SJKEt7c3a9euZe/eveTOnRsPDw+D31OJEiVo1qwZGzZsYNiwYYmOtW/fnoMHDzJ+/Hi2bNlCs2bN8PHxoUaNGixfvjxRmZVSqUz0sy6BZs6cOdSpU4cvvvggybdt2xZnZ2fatm1Lz5498fX1pUmTJrx48YJcuXIl8h9JiJeXF71792b48OGa31V8mVPWCzTZS/HOCIw1oXbv3p1Tp07x+vVrHB0dmTp1qqa72JAhQ2jZsiUHDx7E1dUVW1tb1q1bB8D27ds5ffo0wcHBmuCwfv163Nzc6NmzJ69evUKSJNzc3PD29tY7ht27d3P37t107wLICO4/CicmVnc5rDJWIuz7WArEJMiwNEOINA5aHiUSfJ+XbJ+O4b9wHXcnL0QdqduLzUySUIVHEhkeyJVvRhM9cSi2YwdzCbGRaJPk/JeIzJdjiPX5WgxvqqFClD+dRRgDNyDjNipfI7xvBiCyemT0k53iRJMmTahevTrnz59HkiQWLVpE4cLps3TOrgvv1CLHCOPzMCA8oW6djJdB0SKTQ1tMKIpJ10xKksSd0TN5umYbqgjdcUJSqJGiogi7dY+LvQdSxWIqjlHtRBz0Q7TSTsKz9xBeFuwKp7Lx0k7EBD4IYUBcEd1tBtODBMwEJiJc8reT8WZoJo4cJ7KG9Mz7mSbQ5LJJ3UyoVinZ8vsGZv64jfDwcIoXL46DgwMODg6UKFFC833C53LlypVBo08/qqhoLjTtQ+Tj50ixhotVCsBGUlD32E1Gn9uBfZw4I0kSf/75J0uWLOHkyZP06tWLChUqsHjxYurWrWvw9U+cOIGVlZXG22b8+PGpFsM8PT0ZMGAAXl5eiUShdu3a8fPPP1OpUiWOHj1KWFgYJUuWxNvbO9k9VCpVogyaPHnyEBERgVqt1lwzICCAJUuWcP369WRjqFq1KhcvXqRDhw7cvn2b/fv3U79+fQoWLIgkSVrfU506dbCzs+Po0aM0by7s3suUKcODBw9o1qxZqn4HRkehADMTXj2kgITxDLvi61h1oVAotJpZf/3113z99ddaX3PixIlUjcHFxYXY2Fh58W0kihaywswMnYtvCQlbZZLwpUZs9wUB+iyLcqNbvIl/2JN5vaOT8GTtdu7+qF+cSYoiMoqaM1bwyC43i4d+nWin8yIi4+VvhFGwC6LraVIBRx9HgFqIDcp5iHV4RvAMIc78gMigkUmBbBYnJEni0KFD+Pv7M3nyZJ48ecLFixepWbNm2q5H9jYATg1yjDA+hQta8+CR7q6vudXmIpUwCKEyJ8QcMc+nJOhnU1fyez8u5Oma7XrFmaSopShuxv6IZVE7ivl5QIK9SxXC8BegRDA43kOkY6YGb0RNahSio9JvGL8VjRrhWr8A0apwHVkWq00GOU5kGemZ9zNNoKlfqxAPA8JTNAmOx9LKik3b5uJUcjmRkZE8f/6c58+fExgYqHlcunRJ831QUBC5c+fWKt4kFHGKFCliUGaJsQlYvon39wNSJc4kxFpS4D9qJnkPrmHz5s0sWbKE2NhYvLy8WL9+PXZ2dpQvXz7FNttJ8fb25ttvv+XkyZOEhYXRrl27VI+tfv362NjYaDJX4nFycqJkyZI0btyY8ePH07VrV548eaJVLEla4mRmZoatrS3v378nb968AIwePZoRI0bozIgpUqQIvr6+fPvtt7Rr1w43Nzdu3LihMWlKel+FQqExMI4XaFxdXbNFq21hepVNJWEjkV0V79Tw3XffoVAosLW1xc3NDQ8Pj0QTcWrKBWU+0LpZcXyOvtAaL8zUUO9lIazjI3IuRPshQ6fWcOBe3EMXCsQOq7bFe8IFfH6Mmo0T/vAJt0dOT5U4E49lRCTlx80lvGFtrCq6sgMhxBRENOrYgFgrp1acAWEO/AMiG74CGePBeBdharwI4WkpkzLZLU4MHToUMzMzTpw4weTJk7Gzs6Njx45cunQpzdc09Tghx4iMo33LEvx9O5TIqORxwlKtoPWj4lAaWANUQTjcBsY9En5/F2GU+1bLTexIWdDP5PLaN2cu82jR+lSJM/GoieLam1E0znUSKwryFlHtZIbo0GcGok016O3ilIybiLKmucBsxO+9W6qHpx8lIrVyAyKoLSSTexGbJnKcyHyMMe9n2pTSullx1m19bPD5797cY8ZPm5g4cSKlS5fmk08+0Vt2IkkSwcHBiQScwMBALl++zL59+zQ/v3v3Dnt7e63iTcKfbW2Nly8nqdU8+mVdmhbdGtRqXp+/Rg0HJ1zdazN//ny+/PLLRKJDaGhoqgSaoKAgfH19Wbt2LR06dGDs2LFpEq8UCgVeXl4sXbo0mXFxzZo1mTNnDtWqVaNy5cqcPn2av//+m08/TZxXqc2YN2/evISFhZE3b15OnDjBlStX2Lhxo96xWFlZsWbNGhYtWsTo0aM1nZpiY2OZN29eMpGme/fujB07Fn9/f1xcXHB1deXkyZOp/h1kCNkoJdHoZOOa0dRQo0YNQLR/b9OmTRaPJufwiVMevmpqz4FjL4iK/rD4tjCDPJIl391JEAsi4742RRgd5iPx4jvhIxjDkBA1QS8RTrq6yIX+3VcHhKuigV73AUt/Q0pgJJda1LGx7Fy0nlkrf6YloizJNc1XEzxG/Bo+R5Q2rdB/epq4AvRCrLtTtqeXSUQ2ihMXLlzg6tWrVKtWDYACBQoQE5NyObdOckCckGNExlGnRkGqVy3Alb9DEsUJSwsFRQtb83XTUkIwaI7wKZkCtEG3qB5OcvEm4eMkIhsn6RRthsjG0SbeJIwJRiqvfTBnpfAlSyMSEjd/3UmeMYPIi5YypjgPmlR5ungD1kBZhMnZUoyb2RKFEHz2AlMROwYmPjdkKnKcyFSMMe9nmkBTIJ8VHVo5sPfQ80QTqTasrcyYObMpRw4EUL16dTp16sSECRMoXbq0ztcoFAoKFy5M4cKFk334T0h0dHSiTJz4769evZroOQ9wMcUAACAASURBVBsbG53lVPGPIkWKJGtXrY3XvudQJjDvSisKSWJNh77UXTs72TFJkggJCSF/fsNn1DVr1tClSxfu3r3LvXv36NGjR5rH1qNHD8aNG0dAQABOTk4AnDp1ih07dpAnTx42bdpE/fr1GThwIMuWLWPVqlWJXq9UKpOlgMX70CiVSoYNG8aCBQsMKmNTKBR07NiRKVOmMHnyZH744Qc2bNjAiBEjWLhwYSKRxtbWlr59+7JixQrmzp2rKXHKDkhGjD6LFi1i9erVSJLEwIEDGTFihNGunVay24SaFnr37p3ySTJpYsQgVyqWzcuG7Y959jwSaytzmjYsSt9upSn8j7VYpCV0qj0a92iDaPE5UMtFoxALbG2L74SLckPXvpGIvPCUpoz4bBw9C3hVriiertuFFJt2gQalisJbfbg4fzz58uROfpjUB/1VCEuBM4gNVf122annFOAF/AFo7zMoow9jxon0YmlpiUql0sTYV69epTtj2dTjhBwjMg4zMwUzJlTij4OBbN39jFfB0eS2taB1E3t6dSmNXR4LGIIw3JqBMMOtjxBqGpP8A35uhMBQVs9N1cArdMeQh8BpPogcSa+fkqCfQnlt1POXBJ84j8FtcbW9hchoghetp/roASi0/X3Gj93Q/d73iN2ArgiH+qIIR3lj8Q4R1/9E1Ot6GfHaHwlynMhcjDHvZ2rPC8++LoS+jeXPc6+0ijQK1FhZWzBpZDnc6xTFvc7PjBw5kgULFlC9enU6d+7MhAkTKFWqVJrHYG1tjbOzM87OupO048WOpNk4169f58CBAxpRJyQkhGLFiuksp4r/PmDpRlTvddfJGoqZWuLdrqNIa2YlywSJjIzEzMwMGxvDkteVSiWrVq3Cx8eHadOmMXr0aKys0t7SPHfu3PTu3Rtvb29mzZrFnj17GDRoEDt37mTAgAFERkbSqVMnQkJC2LFjB7Nnz06U7ZPUgwY+CDQrVqygePHiqSq/8vPzo3HjxsyYMYPWrVvTtGlTzp8/j5eXF0uWLEk0GXz77bfUrFmTqVOn4uLiQkBAACqVyiDxLeNQGM3U69atW6xevZqLFy9iZWVF8+bN+eqrr3B1Te/eevrIbhNqeqhSpUqyv8l8+fJRo0YNJk2aRKFChbJoZKaLQqGgWaNiNGtULPnB+gi32zOIxbZvgmP74h5tEUJNtQTHbBD1OfpqdCREqyNd4k3841Uq3sx/cY9ruk95YXVM3DudWCoUvN1+kHz9Oic7FopIMDKUGITfzFWgM2LT0pjsQfzzHQTSHtU/ZowXJ4zBsGHDaN++Pf/99x8TJ05k586d/Pzzz+m6Zk6JE3KMyBjMzRV0bu1I59aO2k+wBoYiBIN4oeZLwB0x+TQidZkY8Wb0xYDqes6LQHdJVSDCoPc5yUtzFXwwu9ci6D89sCsVg9WNOjKK1yf+osiX9ZIfDEV8Okyu8WtnM6L1XmOEKfMMUu9fo4v/gBbAjbj7pH0f+SNGjhNZRXrm/UwVaMzMFEwaWY4GdQqzaecT/r0firmZRHRUNNY2toS9ucC6DZ64On9w5SpUqBDTp0/XCDXVqlWjS5cujB8/Pl1CjT4UCgUFCxakYMGCVKlSRed5MTExBAUFJRJxnj9/zo0bNxI9t0hZnBJG+lWrY2JQvQ/Hwi5PoudDQkJSVd504MABHB0dsba25uzZs2zatCndYxs6dCh16tTBxcWFH3/8kUOHDvHZZ5/Rvn179uzZw+TJk6lcuTJffPEF69atY9SoUZrXaitxsrOz4+nTp0ybNo0///wzVebFfn5+uLu7U758eS5cuECXLl3ImzcvV65cYejQoSxfvlwj0ri4uFC7dm22bt1Kv379KFq0KE+ePNEr4mUKRppQ//nnH2rVqqUp2/viiy/4448/GDMm6xrYiprYLLu90WnRogXm5uaaLLStW7cSERGBvb09ffr0wcfHJ4tHmEOpj8ii8UNk1CQUavbGPdohhBo3A6+pAArFParqOS8aUbyvbSc14YI8UtcFEvNe+QiVIiLdaduq8Eje33uk9VgIqbMV2IOoHLuPeBt10je0RKxHZMUfA4oY8bofHdlk4a1Wq3F2dmbOnDn4+voiSRJ79uyhQoUKab5mTooTcozIYmwAT0RruDWILkAeiLrNKQihxpjYIupL9e2DqRFdj3SVVD0GzpGoNDfM7B5qs3SUg8QhxSqJePAYtAk0IRjuryYh6l4/BfYjyriGpnt4gseIAPQUEctbGum6HyNynMgS0jPvZ6pAA0L8+KJOYb6oU5hmLbrQvmMflixZyIZf5zNwwDYe3quGq3PyDjqFCxdmxowZjBo1ivnz52uEmgkTJug0jc1orKysKF26tN7SK0mSOF7KnZgXqdlu1YOFOVFv3pIniUATGhqaqvKmFStW8O233zJ79my+++47o3juuLq6UqhQISZMmMDZs2cpV05UtrZr147vvvuOKVOm4OnpyZUrV1i+fDkjRozQiCS6BJqVK1fSs2dPKlasmKqx+Pn50b+/aNJasGBBDh06xMiRIzl27BhXrlxh0KBBrFq1SnN/Ly8vxo8fT9++fTWttrNSoJEUCtRGcl2vXLkyEydOJDg4mFy5cnHw4EFNfWSWoci+rutp4fjx41y9+sGspEqVKlSvXp2rV68aRfyUSQF3hFBzGiHUJGzCtSfukVqhJiWsEUaIuqd/sXIIRX85VSDwH8QSarSa+pjX2vLrxVBSI9B4IzLKfwLGpX9YGhYABxAVaXmNeN2PDWPGifRiZmaGp6cn165do3x5IxWr5aA4IceIbIINokRmALAaIdQ0Br5ACDUNM3EsZohyoKLoj0tRaGJG7LB3okNSOlHHKFGeey9qVovFjSHeIyc1Sv4FRIvAqYjf31hSl6api38Q4kwYIlDUN8I1P1LkOGE4hw8fZvjw4ahUKgYMGMC4cYlXPt7e3ixbtgxzc3Py5MnDqlWr9H4+Tc+8n6Wh7/7dyzRu4IqjvYJX/z3TGM3qo3DhwsycOZO7d++SP39+3NzcGDp0KE+fPs2kUacOhUKBRR7jGQ7HREbh8IkzhQsXpnLlyjRu3Jju3bszdepUwsPD+fXXXzlw4ACXLl3iyZMnREcnN1Pw9/fnypUr1K5dGx8fHzw9PdM9LkmSGD9+POHh4ZQoUYKyZT8U8datW5fnz5/z6NEjRo8ezZUrV7CysuLIkSOac5RKZbKSotjYWM6fP8+UKVNSNZbg4GCePHmCm9uHiGdpacnSpUsZMWIEjx8/5vLly/Tr1w+VSvRebNq0KeHh4Zy/fJPBI6dikc+FW4/CuPc0nJCwWKR01PumGYXCoMerV6+oUaOG5pHU36dChQqMHTuWpk2b0rx5c9zc3NJUvvXy5Uv69+9PixYtALhz5w5r165N89uTFIY9TAGVSsXFixc1P1+6dEnzfyup8CiTgTRAZNGcIvmO6B5EuVMHxIIyM1DwwbilGSLN/gfEjuM+hDvuCyAaLL8znlTxW/68uCIal3ggssJHIgSXIERJ0WXExqQuu51/EBu81oA/Yq2cXiRgAiLh6QCyOGMUDIwTKXH37l3c3Nw0j7x587Jw4cJUDcXDw4Ndu3YZNV7mlDghx4hshg2iE5A/onXcPUTMaITwOslO2AAugDtYlM2T0tkGYaaywHyTrdjcKIvImLFFbDgcQpTwDkBM2IuArQiT5DuIjJ54l4oViE5XDxHBwhj2hhcR44pFbLrI4kz6keNEinFCpVLh6enJoUOHuHPnDlu2bOHOnTuJzunRowc3b97k+vXrjBkzJlEliK5rpnXez7KoEBUVxfPnz3F2dqZUqVI8efKEXr16MXbsWB49epRi9sL/2TvvsCbP741/EmQpSsWBAzdOtHVr1aqtWvesto5WResAV6t119VWqzha9657fR11/CyOOureWgcO3IqKCCgCCmT8/jgkJJBAAihoc1/Xe4nvePImkOc8733ucx8dUaNT1FSsWJGOHTsycuRIPDzM1KJmEJw98omUMB2gVCopVb4CAdeuERISgpOTE46Ojnoz3b///pvnz58THBxMcHAwT58+JWvWrLi7u+Pu7k7evHm5ffs2pUqVom/fvtStW5dr167pj7m4WD/5q9VqfHx8uHDhAmfOnKFWrVqcOnWKGjVqAGBnZ0erVq3YunUr33//PaNGjWLp0qXMmTOHpk2bolJrcc6WA+es2dFotSgVCrRaLefPn6dly5ZWKYMAjhw5wscff2zyj9/Hx4fSpUvTqVMnXr9+Tbdu3Vi2bBlhkRr++N8/qBVKipWUlHh0jIboGA0Rr1QoUJDfzYH8uRytKrVKCyytGc2TJw9nzpxJ9pyePXvqFUWjRo1K1Xeke/fueHt7M3HiRABKlSrFV199pR/XWrwLi2pLsWTJEnr06EFkZCRarZYcOXKwZMkSoqKiGDlyZEbf3n8P9RAVzT9IVu+gwbE/47d2iKImuTKmtwV7cC5fALtszqijLKyJMgONsyMFinpQDFGIhyLrbieEC4pGyoqeIt2ZniJeM3ZAbiSR6o5wR+URtXpr4G78MUttCRJDHT9WHLCRDFx8vGdIL2+B0qVLc+HCBUBiesGCBWnbtq1VYyxcuJAZM2aQJUsWnJyc0Gq1KBQKIiIiUn1f70ucsMWITAonYCBiKq9T1NSP3yYgpH8mQjbPIijss6TNTB5Q2mXBySmfdK7SQYMEiFdI+v5PpPW42sQAWZAy4KdIp8JVSJnTahKCiG7LjeUT/t+I2jUvEqjMN++1wQrY4kTKOHXqFJ6enhQvXhyAjh07sm3bNiOFTI4cCWmlqKioFJ8H0zLvZ9gaSVdCYm9vT5EiRbh37x5Zs2ale/fuzJ8/Hz8/P4vGyZMnD5MnT2bIkCFMmzaNjz76iE6dOjFixIhMQ9QU7t2R52cuoU5jJyeFnZLCHVtyerkfGo2G+/fvExAQQEBAANu2bSMmJoYdO3aQPXt2vLy8qFOnDmXLlqVw4cLkzJmTmJgYgoKC6Nu3L+3bt2ft2rXUq1eP7777jqdPnxIcHIxCodCTOYakTuJ97u7uuLq6Ehsby9dff014eDj79u0je/bs+Pj4MGfOHD1BA9C2bVumTJnC999/T58+fZgzZy5O2d05ey0UDVlo+/X32NnZcfZGBG7Z7Tl/UmoTy5e3vmeIzn/GHD777DOOHj1KixYtOHz4MKs3/U2FyrVwcDKtdNJoALQ8Co0h8pUaT4+sKN84SaNIV9f1p0+f6r11tmzZwokTJ6we49mzZ3z55Zf8+uuvgLC/qTVSzsw1o6lBtWrVuHTpEi9evADEBEyHL7/8MqNuy4Z6SNbvIELGHDI4tiV++wIYS4YTNfnbNuHKwLQZ5QFk0cKAr5rzQ/z/wxE1zBUkCRqGKGiyA+UQv0wvpNWqC2KJcA/YjCj/tyLrb1+E0IkmQRikW38nXo/rNp1iPgZpo10Y6Xr7Hn31MxjpGyd02LdvHyVKlEi2fNsUXr58ma738T7FCVuMyORwJoGoWQRMRuLHpwhRY35J+VZRqPsX3Jm1Is0EDS4K8jyqKwRMABIgAuK3wwiTrpNXuiDG+gURUiYb0mnqFBIUYpAvawDoA48hFBiz/6a2vEhWYAASjHYjJsk2pANscUIHXcWBDr1796Z3794ABAUFGVmmeHh4cPLkySRjzJ07lxkzZhAbG8v+/fuTHDdEWub9DCNorl+/rvcoKVy4sL7cxcfHh5o1azJ+/HirfFHy5MnDlClTjIgaXevnggULvpH3YCnytW7Iv71Hp3kchYMDxb7zBkRJU7RoUYoWLUqzZs1wcHAgMDCQWbNm8eDBAwICArhy5QqnT59mxYoVBAQE4OzsTK5cufjggw8IDg6mcePG/PHHH+TOnRuQMqXIyEi9+kanwAkODiYgIIADBw4YHYuJiUGhUODs7EzNmjUZNGgQ7u7uuLi4sGXLFlq3bk2ZMmVwd3fn008/pXPnzjx9+hQnFzdW/HmUmJgY1PF/gvYOjvH3AKERcbgXq8a85du5dHK36Q8jGRw+fDhFgs/T05OTJ0/yx9r/o3iZimgsUNpptBARreLO41eUKJB+ZWvmkJ6u61988QWhoaHY29szd+5cq1VJIJ26QkND9YzxiRMnjCYba/E+LLxXr17N119/zYwZM0weT0n+aMNbQn1ETXOQpETN5vitPULUmPeFf3M4APaDc5Av5nMeK3aiVSTtcmgRFAryNKmLYx43/a6cQK34LRTp4todWZfriJuDSHfUO4jaxgFZI19BSqOGIwobHTSIn01wou08CcqcYBKSr7okqwswGtOkjhsZXHP9jsLSOJHcwjQx1q9fT6dOnay+F8Naex1cXV0pUqRIqst43vU4YYsR7xicgUFAb2AhQtTURdjq8WQ4UZMtT1FyfFCW508upHoMhYM9hXt9hZ2jQ4IHTn2DE/IAzZGSXEPi5jwysesHQsiaF0j58E8IC++IlEglDhC64KAjdiLN3OAD5PNOjv3X7U+tpPM/BlucEFhScZAS+vXrR79+/Vi7di2//PILK1asSHJOesz7GUbQXLt2TU/QFClShPv37wNQokQJqlevru+oYy3y5s2Ln58fP/zwA1OnTqVChQp06dIlw4ia0NBQJkyYQGxcKM3tsqFUp3LhDTgUzodrRdNO17ouTgqFgsKFC1O4cGGaNGmiP67VagkKCqJx48bUrVuXzZs34+XlhaenJw4ODnh5eVGuXDnKlSun/7lWrVpm5VuhoaE0adKE4sWL89133/Hs2TMjUsfd3Z1hw4bh5OREcHAwERER2NnZ0aP3AEb8NBsHB0eyZrU3+14dnZwpUNgTl+wfoNZosVNa9u2Liori8uXLVK9ePcVzHZ1dqPNZC6wpg9RoIfxlHJGvVLg4v8GvjwKL6kEtxeHDh9M8xowZM2jVqhW3bt2idu3ahISEsGnTptQN9o74BqSEqCjRB6d3RsCGN4T6CBtxECFqDL8Wm+K39vHHrBfvWY8bwFDEjwYoTg+e2O1Fy+tUDad2dqLQYPMlh+GALoK4AjXjN0NEIN2amiLWAzGIet0p/tpyiOKmHFDSYDxTCAVaIp6czTAmb64jPJlu3R6GED9OJL8ed0eeHzKyROr2vShWb7rPmQvhKJUK6tfOTae2hXDP4/R2b8SKOGHpwjQ2Npbt27frlZLWwNfXl3Pnzum7X166dIny5cvz4sUL5s+fz+efW+lk9B7ECVuMeEfhjPip6IiaKQhR0wAhat62L0pc/H2MhxJhvbjg+ANqVerKYRVxSorU7WL6oBYJFAWQ95u4xCsUYfY3ArMQRj8KIW9axp+TlYRgUQ6oGv9vUYyZ/mgkAExDMgReSOlxOAmB4TLiK2fa914IGnOBwpyk8y3h8rUIVm28z5XrETg6KGnawJ32LTz4wNX8s88bgS1OWISCBQsa+dk+fPgwWd6gY8eO+Pj4mDyWHvN+hipo6tevD4iC5t69BI+W/v37M3r0aLy9vVPt95E3b16mTp1qRNR8/fXXjBgxggIF3rxuLjY2lnnz5jFp0iQ6dOjAuHMHCPjcm9ePQ3Q1M1ZBbW/HmEeXeLl1K23atElyPDw8PFmZmUKhIDw8nOfPn1OhQgW0Wi1r165Fq9Xy+PFjveLm4sWLrF+/nitXrqBUKk0SNyqVisaNG9OiRQsmT55s8nfUvXt3WrVqxY0bN8iSJQtxcXEsX7mGMlWa4BCvlkkJdnZZyJrdlfVbD3Lqn61mS6+cnBIWxSdOnKBSpUo4OzunOH5wWKxV5IwOGi08CYvBs+Cb+/poUaBRZA7XdR0qV67MP//8w/Xr19FqtZQuXRp7+9QHmvehO0efPn0AGDduXAbfiQ0WQ4HI1usj5U/jgCMGx3VETYf4Y15v4B7CkGzjXMBAqe5qV45ilbty99oq1K+sW3zbZXXmWc8vafFxJVYg6+HE0HVPTQ43kHW0EvgR0OV5IoFrSEL1BPAH4gvpAJTBmLjxRNbWLRGPyfbxY5S04H28ImnC9T5wNDSGxw+jeZ7DnmdFs6FWKLAnIQGcHKljWcSxDCfPhTFq0hXi4jT6UL71r8f47wtmvl8lihd5eyndNxEn/P39qVy5Mu7u7lZfW6BAAZYuXYqXl3xpAgICGDt2LH5+frRr1876hTfvfpywxYh3HFkRGWEfxGV9CqKiaYgQNSa6VKcrtEhN6g/IBPwZuE9vQN6pnxK8Yz+aV9aR+XaOzpS2H0zWNgVFHfQdxtLFKET2aK6LUy6EnPot/mcV0ApYToIkU6e42Y+w+zoYsvy6bSfS9rwjsAIJKKYQizHDb/izbrsJQWdf8fjVa/K8cqRIVCKluyOmg4WpfWmUdO7c+5gZC28SG6vRP2es3fyA7bses+S3KuTNnZ5RKXnY4oRlqFatGoGBgdy5c4eCBQuyfv161q5da3ROYGAgJUvKSmbnzp36nxMjPeb9DCVodG/Aw8ODJ0+e6FstN27cmAEDBnDixAk+/vjjNL2Ou7s706ZNY+jQoUydOpXy5cvrzYjfBFGj1WrZtm0bQ4cOpWTJkhw8eFBvMJRj32qO1upA3IsIsEJJo3R2ovJyP6bly06XLl3YvXs306dPNyoBe/78uVHXIlNYsGAB3t7ezJo1i127dgFC3BQoUIACBQrQsGFDo/ehK226cuUKAQEBbNq0iYsXLxIREUHhwoV5/vw5s2bN0hM3+fPn15M1lSpVonDhwuzYsYO2bdtib29P3YZtePYizuL3DeDo6EyJMh/xIjiQ+/duExgYaFR69fTpU5ycnPSkTWhoKE5OTkyYMMEkoePi4oJCoUCj0RLyItaqezFEeKQKlVpDFrs3t3pMzxKn9EB0dDQzZszg3r17LF68mMDAQK5fv06LFi2sHut98hYAuHHjBj4+PgQHB3P58mUuXrzI9u3b+fHHHzP61mwwBwUio/4UWUiOA44aHN9IAlEzlvQhauKQTOEEkmYEWwBToXTpwcQ1f07Qnh2oFZaRNHZZncnX9nMaTxtBWWSt2w9J/hp+zSxpsz0f8EZKkc4a7HdBEqFVE50fhahhApDOUCvjf34A1EASoFoSFDfm1t86OCMEUVHdPb+I4+cZV7lx6TkO9krsNVrKuTowalBpvCp8kERJ/xR5RjD8fwyy1s5Nyqr55GzyY+M0jJ0SQEyMcfxWqbWoX6kZ5xfAqrnVUniH6Yv0jhPr1q1LlWwdZB7ULboBypUrx7Vr1/Smi9bifYoTthjxjiMrwlb3JYGoqQM0QoiaWm/gNS8BQxCz3FKI0rKFrNsrLvfjTDtfQvedRhNnWZxQZnWi2HfdKTawq3RoGoJ4vaxAal8hIS4lx+QHAdsQdc0BYCTGtbSGMKyl1RE3h4A1BufYxZ/THWOmvwQJT6kOgEf8ZgKPnrxiwrSrBN6JwiGLgjiVloI5nRnXoAwl1C5JCZ1HJNTkmrLzsUOkmsl55piRdIa/iGX6AiFnDBEbp+V5RBxT595g6ri3W0ttixMpI0uWLMyZM4fGjRujVqvp0aMHXl5ejB07lqpVq9KqVSvmzJnD33//jb29PTlz5jRZ3mSItMz7GULQaLVaIw8ae3t78ubNS1BQEEWKFEGpVOLr68ucOXPSTNDooCNqdIqa8uXL07VrV4YPH07+/PnT5TXOnz/P4MGDCQkJ0f+SDZGtRGE+Of0nOz5qisPrOOxVpqzRE6B0dkRhZ0fVTXPJ3aAW+YELFy7g4+ND1apVWbduHR999BEgCprkfEUiIyNZt24dw4YNo3Llynz4YfKOmAqFgnz58pEvXz4+++wz/ftr1qwZY8eOpVKlSnriZuvWrVy5coW4uDi92qZcuXJ88sknTJ8+nTZt2qDVwsvXCuwdUlqeJ4Wd0o4WX3xD4bxJVTFardaoa1W/fv2oUaMGKpWKc+fOGXnmBAdL8ay7uzvlP6zEsJ/m4eiYstLGFJQKeBWrIbvzGyRoMpmVpre3N1WqVOH48eOAyAE7dOiQKoIG3p+FN0CvXr2YOnWqnnT+8MMP6dy5s23x/S5AgcjWPyMpUaMF/oeQNV8iRI0paUpK0AI7kHKmG4mOVQBmIBlZQBGhoPy/P5FVU5xA5SzIAmq1aYN5O5esoNXiOcqHEkN7o1AoqB5/+/2BXUhyMlf8+eEkT9CEA8eR6q5OJE9W6JANqBy/gax5uyDrfldkPX4BWAsEIh93aYzX4aUwrXKJjdPQd+g5noTEoFJpiY2TmPn49Wt+mHCJuZMrUsYzO5YUL6sRdX7ipOtxjNfuuqYmhibIuvX4y9OhqMyoLrVaeBz8mtv3ot66iia9EBUVxd69e1m4cGGqrvfy8sLHx4eOHTsCsGHDBsqVK0dMTEyq1ZbpGSd27drFoEGDUKvVfPvtt4wYMcLoeExMDF27duXs2bPkypWLDRs2ULRoUUJDQ2nfvj2nT5+me/fuzJkzR3/N2bNn6d69O69evaJZs2bMnDnTpKrYFiPeE+iIGkNFTW2EqJmA1IemFcFIrFmCTKIzEWLIYPmstLen2q8LuFZpOnft1qBwUpjtAGiXPRsKOzvKTRtJoW7tZOcWxAz5e8QgfxniO6MjaJILFEuQetQARImauE7WEKZqaV8jiY//Q9Q3JRGC5hiwzuA8nRmaTm1jKNE0mE5eRMTRe8h5IiLj0GggNj73ejs4Ct+tF1g+qyr53c2Un+pKuhIHhsQqnevx/5oSLCmQIBsfKPbkDEZhJhet0cCZf8OJeBlHjuxvr9TJFicsQ7NmzWjWrJnRvp9++kn/88yZM60aLy3zfoYQNE+fPsXOzk5vTgsJPjS6Mh1vb29+/vlngoODUyWhMod8+fIxffp0hg4dip+fH15eXnTr1o1hw4almqh59OgRP/74I/7+/owfP56ePXuaNTo6efsGQ7OFsX/m7zyau4bIq7d4/eoVDrovj50dds5OvFTFENWgMp2X/o6DWwLx4urqypo1a1i9ejUNGzbkb1TGWwAAIABJREFUxx9/ZODAgXoPGnNYs2YNdevWZfHixaxcudLq93jo0CHat2/PggULaNdOJnhdiZoOISEh+q5SAQEBXLp0iePHj+Pi4sLX3j5802soDo7W1+hrgZDncSYJGrVajVKpJHv27KhUKm7fvo2fnx9KpZLIyEiioqKIiorS/xweHs6zZ89wcM6BJg1+QAAaS5yFUw1FplPQ3Lp1iw0bNrBunUTQrFmzok1NjRi8F94ChoiOjk7ie5RaszMbMgiGRM0+hKg5Fn9MC2xAyJqvgDFYTtRcQDKVic3+8wK/IEaMhurj4aB4pKAE3hR16szjAf7cmrqYKO6iVNhDNtDExpG1WCFKDO1Fga+aY+dsPK9mR5Kia5Ek5+z4t5VSidNKhJhZhHHVl6U4BPggfJYuP5ZY1/ka4agCEHXN/+L/r0WIGsN1+IMjIYSGx6IywYrExGhYuOIOv/1sWfstOxLKoVLKXerW7ImTrtdCYoiJMx837OwUBIe8fosETfrGCZ0RfGqxfPly5s2bx++//w5A7dq1mTZtGvb29hw4cMD6AdMxTqjVavr168fevXvx8PCgWrVqtGrVyqiF6tKlS8mZMyc3b95k/fr1DB8+nA0bNuDk5MTPP//M5cuXuXz5stG4Pj4+LF68mBo1atCsWTN27dpF06ZNk7y+LUa8Z8iGzOt9EdmhH6Ie+RxR1KSGqHmNlA5Niv95IBJr3EycqwFFHzvK5hhGyVO+PNy7nds/Lub18xCULvag1aKJjSXHR2UpMbQX7q0aoDR8+FUgJNMnyKTfAumgpPORMfc4oULakXshE7i1jxMvkTba+xHiaWCi44a1tDrFzSkk/upgj1Gw2BIXRHS0yqR7xOsYNas33Wdov1Km70eBfL5uJG+oBhIYXmK6vMqA0Al6/JqYXObjhH0WJaHhsW+RoLHFiYxCWub9DIkOhuoZHXSttnXtkXPmzEmHDh1YvHjxG8kw5MuXjxkzZiQhaoYPH06+fPlSHgD54KdPn87vv/9Or169uH79ulGPdFPn9+rViznz51GiZUtKdP2C89v8WdDnO77v2RtNbBwOedxwrVSOa44avu3Vi64fJB1PoVDwzTffUKtWLTp37szu3bsJCQkxS9BotVrmz59P06ZNCQsLo3Zt64pmd+zYQc+ePVm3bh0NGjQwe16ePHmoV68e9erV0+8bO3Ysjx49olv3HiizpH4yilOpqVatWhLSRaVSkS1bNlxcXLCzs0Or1TJp0iSyZcum32/4c6FChShbtiwfuOXF0TFtNaCWGhenFto33srbOjg4OPDq1St9dvDWrVtp+gwz64SaGuTOnZtbt27pP5tNmzalmzLPhrcMBaJmaQD8jRA1x+OPaZFe1RuQOqIxmF/UPUZMXJbFX6eDI7KwH4EwKYY4hGRl42Hn54hHrjZ4qNsQRyRxlZ7DFrDP6Yq9a+KLk6Iz8pzwDaKSf4l5VYwWWIrwRU1JUN1Yih3I291JQnmSKTghCdvEtEoMorDRdXzdAtw/GIz9a/ML3bMXw9FotCjTeS42XLOXAe4C/kBwXic09kqUZtSvarXWfKb2DSEzxQlnZ2d8fX1p0aJFkjWei4sleqykSK84cerUKTw9PfUy+o4dO7Jt2zYjgmbbtm2MHz8egPbt29O/f3+0Wi3ZsmWjTp063Lx502jMx48fExERQc2aIg/o2rUrW7duNUnQ2GLEe4psiD+MD1K+qiNqGiNETXLqEh10CYARwD2gdfw4ZjgFiH+tY8BKyFLChaIlOlNkfydUpyOI2xeBIosd9m6uZHFJgSwuB5xEypR+R29Yb5ag2YGUOGkQ2WQjC96fDiGIW/x5xJvmaxPnpFRLa0jcnAM2wd91nxKb3XSyUK2Gf449M0/QWAMFYjScA1Hx6F8EOI0EiQfg8cwZxw+UxNiZjl1xKg253KyvJkgLbHEiY5CWeT/DCJoyZcoY7StcuLC+k5MO/fr1o3nz5owYMeKNZRry58/Pb7/9xrBhw5gyZQrlypWje/fuDBs2zCxRo9FoWLt2LaNGjeLjjz/mzJkzFCtWLMXXGj9+PFWqVKFly5b6fVejwgmv+xFlJg4xOje3VouzszN///23WcOkEiVKcOTIEcaPH8/u3bv5999/TZYunThxgsjISHbu3MmUKVNSvE9DrFy5kmHDhrFz506qVbO+tr5Pnz5UqFCBH3+aTkiE8XOKNVAoFMydOzcJ6eLk5KT/w586dSr3799n9uzZZsd58eIFf//9N3v27qVAyY9xzpq6vyuNFpwc3pyJr1aR+UyCJ0yYQJMmTXjw4AFdunTh6NGjLF++PFVjaXn3zR8NMXfuXHr37s21a9coWLAgxYoVY82aNSlfaEPmhQJZfDZE2I1xiDsuyB/wOoSs6YjI0XUh7RUwHTFgjMIYHeP3m/Jzf4X4AuhQB8nOjpH/2uOCfXWX5NkPEyiGNK36GTHbvYXx2lKHf+LfwjKEZLEGKxG/472IQiU1cERKqwybZw1UaUnakDMBWq3Mxek9lbxGuDJ/REzlhpBWk6q6MUqpMNkdVqGAgvmdKVroLZY3ZbI4sX37doYOHUpsbCx37tzhwoULjB07lu3bt6d8sQlYEydSag8bFBREoUKF9P/38PDg5MmTRmMYnpMlSxZcXV0JDQ01UnsnPt/DI8EUw8PDg6CgIJPn2mLEe45sSAmrjqiZirDjTRCipoaZ604gZUYnELnhH4jcMTncQ8icJhiRHIqLCuwrumJf3NW6e3dClDufIzEKpPyoEkm7Hi1A2PvHSBcnSx+MHyDx9B6wFVHsWIPEtbQ6vAJVH635Lk+AKo1qeZMIRup4/YE9iPG/AqgBjb92Z+GlOya9beyUUL2SGzlc3mJ5ky1OZBjSMu9nyFtKTkFjiI8++ohixYqxbdu2N35P+fPn5/fff+fy5cuo1WrKlSvHkCFD9L4lOhw5coSaNWsye/Zs1q9fz4YNGywiZ86cOcOKFSuYNWuW0f6LFy+aJFUUCgX9+/c3qnU2BXt7eyZOnIiDgwMjR45k8ODBxMTEGJ2zYMEC6tWrh52dnVHr7ZTw+++/8+OPP3LgwIFUkTMgPiUNGjTg+LEjaeoabadUUL16dby8vChSpAi5c+fG2dnZqNb70KFD1K1r3BNQq9Vy4cIFJk+eTL169fDw8GDx4sWUK1sW16yp77iX08WeLHZvWEGDwqLtbUCj0RAeHs6WLVtYvnw5nTp14syZM0nK3KyBVmHZ9i6gePHi/P3334SEhHDt2jWOHDmSbFc1G94hKJCF6zHE1MVwoa0jasohcpWpSM38GIzJmZok1Neb+7OYgEhIQNiKJUiEvmhwjmXVPEmQBeGX8iLq8hUkJcvnI6r1apj1YTSJ3xFSZw+pJ2dMIQKIreqGxtH8MqV4kWzpNg/fQkimFshz0iakPOwIQnANByrbK/llZDmcHJUY+sPbZ1GQLWsWJgxLSSOf/shMcWLChAmcOnVK74dXsWJF7ty5k6YxLY0Tuvawus2QnMkMsMWI/whcgGHAHcSf5gwy/zdDlCo63CNB4ngXIWbOkDI5o0VKk0DIEt1XOxqJH6mMEYCw0Lpc8TgkWDwzOH4TmegdEHVPWwvHvYb49DxGSI3U2RYmxStgOVS/5UYWMxyMQgEVy6fUu9ACqBBztx8RdU8+oBtSqtUCie0hwHFw/cmeYQNL4eioNHrucXBQ4pbTgaH9LOlnmL6wxYmMeZ5Iy7yfIQTNtWvXkhA0phQ0gEUkRXqiQIECzJw5k8uXL6NSqShXrhw//PADp06dokOHDnTu3JnvvvuO48ePU6uWZbbtcXFx9OzZk+nTp5M3r/ES9uLFi3qj38To3Lkzx44d4+7duymOr1KpuHDhAnfv3qVGjRpcvXoVgNDQULZv387ly5cZMWKERW3LtVotY8aMYcGCBRw5coSyZdO26Ozfvz8z/H5Gk1q/EiBnCrWaGo2Go0eP8sknnxAeHs7GjRvp0aOH3sj20aNHjBgxguDgYL1RYNni+VLF0CgVkD/Xm2+Rp1UoLdreBpRKJX5+fuTKlYvmzZvTokULs1lFi2DhZGrJhNqjRw/y5s1L+fLlTR7XarUMHDgQT09PPvzwQ86dS8jJr1ixgpIlS1KyZEkjN/azZ89SoUIFPD09GThwYIpeOyVKlKBLly6sWrXK5Dxmw3sABSJbP455omYYkinUoXD8/mMk70lwDphm8P9xCNED8K/B/jQsvl8iHoaHEYVMZ6SrE8ATRD2+E3kLlkCL8FAHkC6wVuZszeIRQoZUBwo0ykf2LKbnOCdHJb2+KZrq13mF/BoHIZ40veP3TUH8Khchzx+Ji4yrfpSTP2ZWoWmDfOTN7Ui+vI582dqD1fOqvlX1jA6ZKU7Y29vj6mr8l2DJmsMs0jFOFCxYkAcPEr6cDx8+pGDBgmbPUalUvHjxgly5zBf7FSxYkIcPHyY7pg62GPEfgyFRMxnxUamJkP3eiFzxT2QSDYzfZ4nIYRVCciRWYgYgZUemHycsh656cwYyQX6IyAgBFsbf42Nkkrbkfs8gPjcxiEyzbvKnW4RQ4Cfk/ftCR60HWRxMz3EO9kq6f5VKIvQxkn34CunUVAf4FVEc/YK0OXyMZDw6YlQX3PhTd+ZNrkj9WrnJ7eaAR34nun9VmBVzqpLb7e212NbBFicyhqBJy7yfqRU0AG3btuX69etJjNneNHREzZEjRzhw4AA1a9bkyZMnHDp0iM6dO6NUWv7R+fn5UbBgQbp06ZLkmLmyJBAT1m7dujF//vxkx9d1cMqdOzebN2/G19eXTz75hEWLFrFs2TJq1KhBWFgY7du3T/FedUZ6/v7+HD58mMKFC1v2JpNB9uzZCbxxjX/+3onGlItXClAoIF8yE5pGo2Hjxo0olUrat29PkSJFWL58OZUrV+bQoUMEBgYya9YsmjZtatSa3NFBSb6cDlhjX6BUgFsOe7I5vXm5YGZivAEaNmzItGnTePDgAWFhYfottUivCbV79+76tvGm4O/vT2BgIIGBgSxatAgfHx8AwsLCmDBhAidPnuTUqVNMmDCB8HDRyeqMH3XXJTc+QEBAAH369CE0NJShQ4dSokQJ2ra1NL1kwzsFQ6JmGaYNHEFkGDuQhVtyf8dxQE8SFsYVEU8DENm2IeGThs6cug5OORG7g8+R9eYxxHumFlI9ZUmlvhpp430f2Iy0xk4rriIfQxOkqccFYKyLPXN//YhcOR3I6mxHFjtwclLi4KDEp3sx6lS3nCTWImbEs5BEcWWkU2wD5DPYh3zsXqTM2xcumJURA0uzZVlNNi2tiU/34hmy6IbMFSe8vLxYu3YtarWawMBABgwYYHEiyxzSK05Uq1aNwMBA7ty5Q2xsLOvXr6dVq1ZG57Rq1UpP1G/atInPPvss2QeH/PnzkyNHDk6cOIFWq2XlypW0bt3a5Lm2GPEfhQtCZtwE2iG+ZsuRuLERIRostd0IRsqhagO+iY7piPy0KGggwUn+e0Tx44qUJg1BVD45EYmlKf+YxDgAfIqUJx0hqWO8tbiDGBkXRpIYNYB/oOBRZ/wmlMclm50+Tjg72eHspGTM4DKUKZmyXxsgsfgwMAop7yqAmLIdRtj6DYii6AgwGgkiyTwKlvbMzs8jvNi64mPWL6pB1y+LvNXSJkPY4kTGEDRpmfffugdNbGwsDx48oESJEkb7dQoarVZrFBAdHBzo3bs38+bNY968eW/tPlUqFYsXL2bChAm0aNGCxYsXs2zZMipXrkzPnj0ZOnRoEjWMKVy9epXffvuNc+fOJQn0z549IyoqKlkSxMfHh48//pjx48fj7Gx6Gfz8+XO9QbBCoaB379588skndOrUiZs3b1KmTBmGDRuGnV3ypEJsbCxdu3bl6dOn7N+/P1nD45Sg1WrZu3cvU6dO5erVqzRo0IAbl0/waaMWWNP8SKWKI5tTliSESFhYGHv27MHf35/du3ej1Wpxd3dnzJgx1K1b1+xnlRgeeZxQaeDR05fYOyS/wFYqwDVbForlS4/HkeShzYRdnDZsEBv9uXPn6vcpFApu376dqvHSa7KsW7dusiqzbdu20bVrVxQKBTVr1uT58+c8fvyYgwcP0qhRI9zc5Am7UaNG7Nq1i/r161ts/KiDnZ0d9vb22NnZoVQqyZs3r0Xzgw3vKCKQThu/AbFmzrmALPK6ILJoc8zHtPhzQTKSS0loIXrJ4LyipEmm8pwE30cFkrCtjRgI30YSurNMX2qEWKArkB+YQ9qyPFpENe6HlPAPRRqEGI7pWcyFLctqcup8GHfuR+Oaw566NXOT3SXl5Us08ozgH/9vQYQAmoG833ekgtIsMlucmD17NhMnTsTR0ZFOnTrRuHFjxowZk6Yx0ytOZMmShTlz5tC4cWPUajU9evTAy8uLsWPHUrVqVVq1akXPnj355ptv8PT0xM3NjfXr1+uvL1q0KBEREcTGxrJ161b27NlDuXLlmDdvnr7NdtOmTc3GCVuM+A/jb6Q19yVERVMJedhvibS2HofUlqaEgUiXI135qyEuIkRI8TTeq2Grv4qISmQwMmnq8BtGLb9NYiuiPCmJKH5MC8ssw1mkfHgjEiO7kMCmx6PyhznZsaoWR0+F8ij4NbndHKhbMzdOKSVTgxClkD/ye3oR/xq1kRjfFFElvcPBwhYnMg5pmfffOkFz69YtChUqhIOD8bc7R44c2NvbExYWlkRS2rt3b7y8vPj111+TyKLeBHbt2sWQIUNwd3dn165dVKwotG/lypUZPnw4kydPpkyZMnz77bf88MMPZj9sjUbDt99+y4QJE0ySMJcuXeLDDz9MNkPj6elJtWrV2LBhA927dzd5jk5BY4iyZcsyceJEvv76a86fP5+iEiYqKoovvvgCZ2dn/vrrL5ycUteJIi4ujo0bN+Ln54dKpWLo0KF06tSJ2NhYihQpwuCho4lSOVtM0kRHReKeTYFGk4OzZ8/i7++Pv78/V65coV69ejRt2pRx48YxevRoPv/8cxo3bmzV/SoUCi6d2su2vw7Sb8g4NBqS3JtSCUqFggK5HHDP6Zg2KZ4VeJvqGEuQ1hpRQ2ixfEJNyfwxJZgyhwwKCkp2v6XGjzrkyJGDChUqMHjwYHr16pWsLN6GdxgqhEAZg9SbG6IhCQYvp+P3aRBJ+hpkUTkGWbDqcA3xntHhB4xNENOpvAkSFDSGKIUkd4chzTHMiYF0iALaI8qbUaR+zapG1CvTEeX4cGQ9bA52dgo+rpqLj6sm/73SIqVa/vHbQySB2xQpXXr7BUhvHpkpTmTNmpWJEycyceLEdBnPmjhhCZo1a0azZs2M9v3000/6n52cnNi4caPJa80lAapWrWqRwtsWI/6DuIawzv+HEOz/QyZQBTIhzUYmweqIj8k4knYv0mFr/PW/kGBIb4iLiMIyrc/hiQNFVsTr5gAiQQQp8dFiPgAsQwzvqyH1rykFFlPQIn43UxF5Y3ZExTMIs2SPvb2S+rXzJD9uLCKZ1AUJXRKkIPK7aYrE8jf/qPlWYYsTGYO0zPtvnVIzVd6kgzkfmgIFCvD5558beUS8CVy5coUmTZowaNAgJk+ezL59+/TkjA4eHh7MmTOHixcvEhUVRZkyZRg+fDghIYlX6zBv3jwUCoW+pCIxkitvMkT//v2ZPXu2WR+M8PBwky22ly5dSokSJfD29sbb25tRo0YRFxeX5LywsDAaNWpEgQIF2LhxY6rImcjISGbOnImnpyeLFi1i4sSJXLx4kW7duuHg4ICLiwvffPMNq5bOwiOPEwoFJu9FB6UC7JRahvp+ycD+vuTLl49u3brx4sULfv75Z0JCQtixYwe+vr4UK1aMw4cPJzEItgShoaH07duX7l3aUskzB6U8suGWwx4XZztcnO1wy54FzwJZqeSZnXxuTm+NnAHQoLRoe1vYsmVLkm3fvn08ffrU+sEU4rpuyZbZzR8B1q1bR926dZk3bx4dO3Zk3Lhx7Nu3L+ULbXh3sBvJKPbFmJwpjSzA9yB1PycRMxfDhbaOqCmDGAvejN/XC6nNByFuxiV6zXQwCNbBMDFqiKXI2nQ0sjb9n5nrw5DKrtbx56ZmJnyNeLtUQtbGS5HnjuTImZQQiXSG9UESqt8hC5s5COmkM/59H8kZyDxxYsWKFVSuXFnfYbFq1aqsXLkybYNaEScyO2wx4j+EUETtUgFpB6cztupAwsSZHWG57wATESlhNaAVohgxxHOkpOlDTJuEaZFYkdbyJt1rJX6c+JcEcqYoMtl2JMHEzBDTkbKgBogixVpyJg5YjQSJJsjn5oeU+vqROiXOAxJMxXIjrP0MxDNmCvLZPUCUSV/w3pEzYIsTGRUn0jLvZyqCxpwPDQhJMXfu3FR5mKSEp0+f4uPjw6effkrTpk25dOkSLVu2TPZh3MPDg7lz5/Lvv/8SGRlJmTJlGDFiBM+eieX5vXv3GD9+PEuXLjXrV2Oug1NiNGnShOfPnydpCamDKYImKCiI/fv3c+fOHX777TfOnz/PhQsXqFOnDrdu3dKf9+jRI+rVq0etWrVYunSp1e3Mg4OD+fHHH/Ukyf/+9z8OHjxI8+bNk7xvX19flixZwgdZtZQv5sKtq2dQq+KwU4pKRa2KIybmNZEvQlm5aCpN6pTi3p2beHl5cfr0aQICApg+fToNGzbE0TGhHOnOnTtoNBqKF7de29mvXz86duzIJ598gkKhIEc2IWTKFXGhXBEXPAtm4wMX+7dKzAgUaFFatL0tLF26lG+//ZY1a9awZs0aevXqxZQpU6hduzarVq2yery3VTNqzhwyuf2WGj/q0Lp1a6ZOncrChQtp1qwZy5cvp0WL9GpVYEOGIgDpwNEEeeLXwQ3JgF5CZOq6v1VF/PmnEOImMVGzEiFqaiG17DosIamZiyFBk0bzR1MKmjuIykSNqNj/QdbGPcColfQjxIZgAMJPWYsw5BmkItKwZDdSymQqEZwStMivYRqS6KyJPAe0Qvwodca/pXinVekWInPEiRUrVvD7778zffp0Hj16RFBQEH5+fsycOTNVscEQmdVbwFrYYsR/ALFI+Y8nwgx/ixgAD0NUJ6aQAyFq7iLqmCNIzGiNmMcTf30w4gFjysIkCJlk04OgMRUoFiBlP9mQuPYrsAWJSboYpgVGIirQDoj/mqXeOiAB53egBFJ3G4coce4gKiRrSJMYEkzFyiOeNX0Q4qsTYs4chqiChiFE2jswh6QetjiRUXEiLfN+piJozCloAOrUqYOTk1O6ZhxiYmLw8/OjXLlyODk5ce3aNQYNGpSk/Co5FCpUiLlz53LhwgUiIiIoXbo0I0aMwNvbm8GDB5t9r2A5QaNUKvH19TXy/jCEoQeNDkuWLKFAgQL069eP7Nmz4+7uzs6dO+nSpQs1a9Zk1apVBAYGUqdOHbp06cLUqVOtIiFu3LhBnz59KFOmDKGhoRw7doxNmzZRo0YNs9eUKlWKSpUqsXHjRpwd7CjgloXvvm3J3YBjbF+/iAkj+vBtp0bs2ryAti0aEPTwIVWqVKFBgwbJtiU7fPiwnmCxBhs3buTChQvpJrNLT2jJXKZeIL5MV69eZfPmzWzevJmAgAAUCgUnT55kypQpVo/3tibUVq1asXLlSrRaLSdOnMDV1ZX8+fPTuHFj9uzZQ3h4OOHh4ezZs4fGjRtbZfyowxdffIGnpyeDBg0iOjqalStX6g2HbXhHEYIoYj5E5B462CNsxk2gP6YXzCALvubIgnYHUMXgmBrjlqudSdrdQo2xB00aF9+mEqOLkDX38PjbzYOUHlVF+KPTyNtsjJTjf2Xla95DFC11EO7pdPw4+a0cJwJZU/dBOprrnncWAJdJMP7Nam6A9xSZJU7Mnz+fP//8k08//RRXV1c++OADPvvsMzZv3mx23WIpMuvC21rYYsR7DC0iBfRCYkMNhFyfD1hqM5QDkSbeAX5GlDdVkIl4MVLeU8XMtToiP70IGkOp5UtE/alB2Pk8wAhE8WMP1ENKd3shnaX6IJ0LLfVMf4K870KIMXExJF5eArqTsteNDneRgNAaUcY0RAJDPoTNv4IEpIVIKbKFvsHvA2xxIuPiRFrm/bfuQXP9+nWzXirJKWgUCgX9+vVjzpw5NGrUKE33oNVq2bRpE8OHD6dChQocO3aMUqUs6V1hHoUKFWLevHl6cubQoUNUrVqVZ8+emWxHrHvYNdcaODG8vb0pUaIEwcHBuLu7Gx1L7EGjUqlYsGABUVFRDBw4UL9foVAwcOBA6tWrR7t27Xj8+DG//vorgwYNsvh9njhxgqlTp3Lo0CF8fHy4fv26VUZ3Pj4+jBo1ihs3bvDXX39x/vx5li6aTbt27egxZzpeXl6MHj1a7zWUPXt2Xr58meyYOoLGGgQHBzNgwAC2bdtmsaHw20ZmqhkFePDggdHfXt68eXnw4AFubm7Y21vnTJ+eNaOdOnXi4MGDPHv2DA8PDyZMmKAvn+vbty/NmjXjr7/+wtPTk6xZs7Js2TIA3NzcGDNmDNWqiTvf2LFj9YbBlho/6jBy5EgqVaqUohG3De8AYhBlzC+IYaAh2iKS6JKJL0oGCqTGpjmiqBkHnE90znpkQTsaySAC3EL6PoOwGyVIE8Ix7soag/hU5oi/NcPb9UX4oi8Rtf4WrCtD+hexDghA1txTMc9jmYIWWZ/rvBtDEcV8OyTJmjln7IxBZogTERERFC1aNMl+nbFuapGZvQWshS1GvKc4h5AnBxH22B9RW6YWroip/ACEWJgUvz8AiRuVTFyTngRNYiZ/NWI+lgUhn3SoHn8/PkisBAkcc7BMjXIdKYdaiSiP2iFKGfM5XmO8Rror6bxkrsXvL4q42DdFSpmsUfG8x7DFiYxBWub9DCFoklPQnDp1yuy1Xbp0YeTIkdy9e9fkL9kSnDp1isGDBxMZGcmSJUv47LPPUjWOOTg6OnL58mW2bdvGjh0Yd5lJAAAgAElEQVQ7KF26NH369GHIkCFG5kCBgYEUKFAAFxfLZg83Nzfat2/PkiVLGD16tNGx8PBw8uRJMMbasWMHSqWSHj16mCSHIiIiiIiIoE6dOsycOZPq1avz8ccfm31tjUbDzp07mTp1Kvfv32fIkCGsXLmSbNksq+p//Pgxu3fvxt/fnz179hAVFcWDBw+YMWMGixYtonbt2nz77bcAVKhQgYsXL+oJF0sJmgEDBlh0LyAEXd++ffH29k5W8ZPRyAwTqiHq169PixYt6NChAyBtSOvXr09UVFQSk2pLkF4T6rp165I9rlAozLLzPXr0oEePHkn2W2r8aHi+De84tAgTMQxpa2SIykjNer00jK9Auna8QCTchtAgcu6ViEfNaIzLmyogEvM0IHFidAuSRByKaSnt8/jbaohwSitIvvxfi6jKpyJeykMR5Y2lX/MXSKmSP6KaL4U85/yBJFVtMI3MECeSS3KkNQGSWRfe1sIWI94zPELm6RWIYmMeoiJJr6cqV4SE0CBKkj+RONQGmZAN7TH/Rdj3tHqnxCBJAR1Bo0WSFYr4eyiQ6HwFUnoFImlcDXyC+NOYw3HES2Yboo7pjhBcliQ9bpNAyBxA2vQ5InG5DxIwSvOelyulDrY4kTFIy7z/VgmaZ8+eoVKpzCouklPQAGTLlo1u3bqxYMECJk+ebNVrP3jwgJEjR7J//35++eUXunXr9kYyGQMHDqR79+40b96c5s2bM3LkSCZNmkSpUqXo27cvgwcPJleuXBaXNxmiX79+tGzZkuHDhxt5xYSHhxspgGbNmsXz588ZMmRIkjF27tyJt7c3a9eupVGjRmzdupU2bdrQv39/Ro0aZfSZxMTEsGbNGqZNm4aTkxPDhg2jffv2KfrUqFQqjh8/ru+4dO/ePRo2bEiTJk347bffWL16NQEBAdSrV49nz56xcOFCfH19AahUqRIXLlywmKAJDg7m6dOnFiuRANasWcPNmzeNWmhmPrzd8iVLMHfuXLZs2cKRI1J03K1bN7744gsUCgUHDhywbrB3RJZuw38EZ5Ds4OFE+wsgGcxvSJ+C4KdIzY8OdREJuU5Ro0YYiRUIKaNDOmRGE1sLzEISo6bKlv5CVOz+CDnyfwhRMxlRkBtCBWxC+KtiiNeMJUsSLfJcoVtvv4h/jY6IfYOlCvn/NjJHnLh69arJ9YxWq+X27cRspxWwxQkbMhuiEWXLFGTy+wEhatLbWPY0Mqn2QUp3fkMm7RlIOVVbhKj5iPQzCNZVXuiY/GOISa+CpObEoYjX2lkkuVAP6VTYCZE+ziahjEiDBBE/pDQqJ/KZ9QeMCwKM8QoxRtMFicD4/SUQo7QmQH3eXwf4dIMtTryLeKsEjU49Y84rJDkPGh18fHyoVasW48ePt6jbUGRkJH5+fsydOxdfX19u3LhhsWrFWmzbto1z586xfPly/b4iRYqwcOFCI6LGx8eH6OhoPvrIOtfHihUrUqRIEbZv3067du3QarWoo6KJfhrKBzlyAKLMOX36NF988YVRC2EQYmLIkCHs2LFDrxxp06YN1apV45tvvmHv3r2sXr0aV1dXFi5cyMyZMylfvjyzZ8/ms88+S9bjJSgoiF27duHv78++ffsoVqwYTZs2Zfbs2dSsWdOI1OnZsyeenp6EhITQpEkTvL299UbHFStWNFJRpUTQHD58mNq1a1tMtgUFBTF48GB27dplZDSc2aAFNNrM1ZpCoVBQtWpVXF1dadiwIdHR0URGRpI9e+qKed+Fzhs2vOd4iBg0Jvanc0YWpENJ38XfIGRhC5J13YR0ldgOjAcuxB9TG/wMqetckQiGyvUrwGPEeybxImANUka0BynfB6nQqoIkO3cjzydahEtagCjJ1wMp2bSHA3uRtfYxoCyiRF+FcfmVDZYhs8SJq1evvrGxbXHChkwBDTI5jkRMedsjJI31vSlSRizQE5mAdfZ+HwBjke5QMxHC5k+EMb+OlAilFTqCRhcoZsX/2xZjhctD4HNE0bKZBNb+EOKfozM7Xh5/b9OQEqQi8ffeA/OlR4FIgNiFqGReI+qc+gih0xTrSoxtsMWJdxQZQtCYQ758+QgPD+f169dmyZeSJUtStWpVNmzYQLdu3cyOpVarWbFiBWPGjOHTTz/lwoULSQiL9MTz58/p168fa9asMSnVKlq0KIsWLWLUqFFMmjSJ5cuX07p1a8LCwvS+F5agn48PO36ZTsFFOwg9eBIUCjqqVdjtGseB8Us54+GCvVqbpAxq9uzZ+Pn5sW/fPry8vIyOFSxYkL179zJmzBjKlCmDnZ0drVu3ZufOnUnajOsQFxfH0aNH8ff3Z9euXTx8+JBGjRrRsmVL5syZQ758+UxeB5ArVy7atGnD0qVLGTFiBPXr1+evv/6iS5cuVKxYkUWLFunPzZ49e7K1idb4z2i1Wnr16oWvry+VK1e26JqMRGZgvA2xePFiFi1aRFhYGLdu3SIoKIi+ffumyrg7M9eMpgZarZY1a9Zw+/Ztxo4dy/3793ny5AnVq1fP6FuzwRSikGzeVBJ8XnToishAPNL5NXcgLIYOsxDDRZAFbitE9j0ekZYY4hfkoWA0UmNvAbRqNU//OsiTHfuIefKMz7VanrnnxrlJXWa1bkisvT3eia6ZjXBGe0nakjs/sm7+GXkmcQK8kQRn0kJagQYRB+nW21FIN6iuiFej5Xb8NphDZogTyZn4pwXvU5ywxYh3GIcRheUZRB64HnE+f1OYgphwbSepMucDRDkzCGHSpyGE/h6ke5KVShp1TCxPtuwmZM8RYgNDUSjtcFydh3x2n5Nncx0UKMUTR4cbyCQejkzq9Q2OZQEmIK31OiPlTiAqn7Xx95f4qTMaIWJ0hmO6JrMlgd4IIVMPm/FYGmGLExmDtMz7b5VzSomgUSqVeHh4GLW+NYX+/fsze/ZstFqtyeMHDhygatWq/PHHH/z555+sXr36jZIzAMOGDaNly5bUq5e8SYGOqMmdOzdKpZKSJUsyZswYi1ydH678k9wj5tPiUgjP9h1Dq1ajVamw0wJaLdE371Hs4CWWaQoRN28D6tcxaLVaxo8fz+zZszl8+HAScgbg8uXL9OzZkwULFtCqVSty586No6Mjnp6eRuc9ePCARYsW0a5dO/LkycPQoUNxdHRkwYIFBAcHs379erp165YsOaND//79mT9/Pmq1mrZt27J161ZAPGiuXr2qN3m1REFjKUHzxx9/8OTJkyTkVWZFZnBdN8TcuXM5evQoOeLVWiVLluTp06epHi+zuq6nBr6+vhw/flzvh5M9e3b69euXwXdlQxJokKxeSeAnjMmZTxBZ+QrSn5x5gZgp6tAckYIbQoH4C5wjqaJHjbTh1i1a75p/qdjQcAInzWNvwdqc7zaUh8s2E+L/DyV3HSJkxRb+7T2aBgVq8fPY31AEPwNkgTMeWeP7k5ScAenm1A/YiChpsiOJ1lyJzgtFmnh0Q9poT0ISwesQssYPUdzYyJn0QXrGiefPn9O+fXvKlClD2bJlOX78+Bu++5TxvsQJW4x4B3EbUcrURboNrUK6771JciYAIeQ7Ip5l5vABMmnrFDZXECKkPcbd/8zgdVAwV4dP4e/8NbnkO46g1VsJOXmYp8qDPNizkXPeg9jHJ9wqtoS44vFr8HPIe3+FmCLXTzToQ6Tk60tEsqkrX3JDPsMsSLC5hpBLjeOPtQCWIgFjDhJsbiBqmybYyJl0gC1OZEycSMu8/9YVNF9//XWy5+h8aEqWNK9ha9KkCQMGDODUqVNGJq83btxg2LBh/Pvvv/j5+dG+fXurWy+nBgcOHMDf358rV65YdH54eDiRkZGsW7eOu3fvMmnSJEqWLImvry/ff/99kpbZWq2Wq8OncH/hOtTRr3FK5kvkjBI0Wh4s20T48fNsqejOwdMnOXz4sFEHHq1Wy6FDh/Dz8+Ps2bMMGDCAmzdv4ubmxsuXLxk4cCCVK1fm+++/5/bt2/j7+xMcHMznn39Ou3btWLBggVXdmxKjSpUqFChQgP/7v/+jRYsWfPfdd7x69YqsWbNSpEgRrl27RoUKFciRI4dZgiYiIoIbN25YZMJ07949RowYwf79+63uOJQxUKDNZKtOR0dHoxb0KpUq9d+vd2RRbSlOnjzJuXPnqFRJWizkzJmT2NjYDL4rG4xwEMmCJu6gVBxhDdrx5swFhyMKGBBmY34yr6XEWCVjD8TF/6xCWq4uQ2TiozCqDXoZcJMTjbqiiohE8zrG5PDql1FkA1xm/MHB+WupvmsZY6uU5zliHJx4djyFCI3uI16Os5GFwxjETqc5oro/hSRBYxD1ew+EU3oXZtt3F+kbJwYNGkSTJk3YtGkTsbGxREdHp9vYqcJ7FCdsMeIdwguEJJmFTHY/IZNf1jf8umrgW6T8Z6aF19xCCIy7yOT8O1J21B4piaqQ9JKwo2c53ao36texaM38Dapjo1ErorkRPIu7H62k5q+rydansLDyexEXdx0uIUqetQgB8xVC1FREEiIDkHrW2gg5czf+ujJI56emSIIkZdcKG1IFW5zIKKRl3s9UChqwzIfGzs4OX19ffWeWsLAwvv/+e2rVqkWtWrW4evUqHTp0eCvkTHR0NL169WLevHl6ZUFKuHjxIhUqVECpVFK8eHGWLFnCqVOnCAoKomTJkowbN85IUXNj/EzuL1yPOvq1xfeleRVD+L9XKf+/w+zftVtPzqjVajZt2kSNGjXo1asXrVu35u7du4wePRo3Nzfu3r3LmjVrCAsL4+HDhwwcOJB///2XJUuW8OTJE9asWcPXX3+dJnJGB13b9Dx58lCxYkV9qUzFihW5cEEMGJJT0Bw7doyqVasakQYmPwuNhp49ezJ48GAqVDARrTIhtIAGhUXb20K9evWYNGkSr169Yu/evXTo0IGWLZNL8SSPzMp4pwb29vao1Wr9nBMSEoJS+R4Wxb6LuInU0H+KMTmTg4Re0F/w5siZg0g9jw5TgJQEnYYdnBohC27DqUsFLEIUNX2AexB5/TbHPvmK2JAws+SMIbQxsaieR/BPg29wPR/AChLIFA2wE0mQjkPW0CeQxGgWIASxP3iFfKS6hOkmxDPyV0SVbiNn3izSM068ePGCQ4cO0bNnTwAcHBys7s63b98+Xr1KXDOYNrwvccIWI94BqJBuTJ5IC+guiCfKGN48OQPijn4cIWcsXWJfBMojdaYTgDvI/e5Gyp2+BAwaUoYdO8epZj1QRUSaJWcMoYmN4fWjEI52bU+0e5CY/JZCJp+DiFHwh8jk3w+Jt6sRl/cZiG9PLGKGvyv+upnx93k1/pxG2MiZNwhbnMi4OJGWef+tRQeVSsXdu3eTVcZAyp2cdOjRowc7duxg4sSJlClThtevXxMQEMCwYcMsMg9OL4wbN47q1atb9aBqqoNT8eLFWbp0KadOneLBgweULFmS8ePHc3frbu78vhx1tPV/zHYaLQXUdjwYP5tXr14xf/58SpcuzfTp0xk5ciRXr16la9euHD58mMGDB1O2bFmqV6/OsWPH+Oqrr7h//z6BgYFERUUxZsyYNJWzmEKHDh24ePEi165do23btvz555+A5QSNpeVNCxYs4OXLlwwdOjT9bv5NQyumXpZsluC3337Dy8uL8uXL06lTJ16/tpzs02Hy5MnkyZOHChUqsHDhQpo1a8Yvv/xi9TgQHzCUlm3vAgYOHEjbtm15+vQpo0ePpk6dOowaNSqjb+u/jXBEMVMO6Xqhgx3CONxEsnxv0iv8FdJ6VYdPEEIlJRgSNBURdc8FZBFs2LAuDlgEKs9oTlT7BtXLKDBT+msOWaKiqdu4O6rQcGKRCq8qwAYkgeyPqNNPIGRNDSTheRXoCzxA1PXnkGealKkhG9INVsSJkJAQqlatqt8Mvd4A7ty5Q548efD29qZSpUp8++23REVFWXU7K1eu5KOPPqJmzZoMHTqUHTt2WFS+nczbe2/ihC1GZHL4I0RDP2SOPYu4oCduLf2mcBdRRDZFiCFLoGuFZ/g44YYofu4i3jG74o9/Ba8PBHO6ZS+rkr3yOhrieMkJ1TdocsVKnWt1JOlxBjEmu4Iw+r8iClAvJL4+Qfxy/BEV0j3EWf6FdbdgQxpgixMZFifSMu+/tRKnO3fukD9//hTJk8KFC+vb+JqDVqvl6NGjaDQali9fzv79+61qs5xeOH36NKtWreLSJQsKPg1w8eJFsya1xYsX548//uDWrVtMnDiRzZNXUkad+jyk5nUM95ZvptWmRXz4cQ2WL19O/vz52bVrF61bt+bQoUNUqFCBJk2asGrVKipXrmzE7uXOnZt//vmHX375hUqVKrF48eI0qSYM4ejoqFcfDR48mEmTJqFWq6lYsSJ+fn5A8gTNoUOHGDt2bLKvcevWLcaOHcuRI0dSbA+e2ZBe/jJBQUHMmjWLgIAAnJ2d+fLLL1m/fj3du3e3ahylUkmbNm1o06YNefLkSfmCFPAuZD0tRZcuXahSpQr79u1Dq9WydetWypYtm9G39d9EHLIAHA+EJTrWDFHNlHtL9zIeIYJAiKAlWJYWMTQJ1i2+lYjSpy1SizQBfWY0SL0DVXQUKKwjZ0CEQ+rXMaxeupGpw3rTDPGmdECSsL8ia/AqyLPDNhK6O+lQCNiHtOH+BFiJqNdtePOwNE7kyZOHM2fOmD2uUqk4d+4cs2fPpkaNGgwaNIjJkyfz888/W3wvK1asAODRo0ds2rSJfv368ejRI1QqlcVjJMb7EidsMSKT4jJCJOxGlDNbEcP2t/l3p0WIewUSuyx97SfAM0wbA7shxMl3SMenmXB30yo0WVJbVqchNiScJ8X+pkBwM2l1PQaJa/v+n73zDo+i6uLwu5seIEhJgRQghCZgQhMQRAIKSAlBWuADEUIToijNAvIJ0gQJIhFQQZEaLBBQP0ApotJCS4IiCNKkBCGhJiFlM98fN7vZTd0ySTbLvM8zD7tT7twJyT13zj3ndxBOoUxECu+zCEH7boCfXhPdcrYXEQ6e9xH5seXAwVreUexE2WDJuF9qb6ynT58uNr0JRATN+vXrCz0eFxfHpEmTSExMZM6cOSxYsICGDUt/KpiRkUF4eDiLFi0y+WU1Pj6e4cPz1s8wpG7dukRNn8nPX3VH0liWp6yRJGY/24cjHk6MGDGCe/fu0a1bN4YOHcqaNWuKrSJlb2/Pu+++y7PPPsuQIUPYuXMnCxcuLLBalamMGTOGwMBA5syZg7e3N/v379dF0EiSVKiD5uHDh5w4cYK2bdsW2nZ2djbDhw/n7bffLpPfEcuQN2c0KyuLtLQ0HBwcSE1NpWZN45eFJEli5syZREVFkZ2dDYg0w1deeaVYB1mhlJOwdGNJTk7Gw8ODQYNylV8zMzPLid6RjSAh8nImI0p76tMYEUrdpRT7cxSRl6/lXQzz9gsjG0ORx7yTbzVCX+AF4FuQ3pU4f2YlGpX5IcPZaQ+pumQ1kZNHsl+t5oWc23RFzJ9bUfxkwQ4xJ++MKNbxKkJOwYb+zK0Q+eyEj48PPj4+Ol2/fv36MX/+fJPaWLduHb/++isnT56kevXqREREGC3iXyA2ZCcUG2Fl/IsICfwUke66GBFZWRbq5WsQCu1RGDo0ikMbaRlYxDnVgNmQPT6DS/U2kp1u/vuEJj2Vv6VPqflsd6Elo30nfwJ4HeHBf4qif4bP5vQ7POeanQidGs8irlGwEMVOlBWWjPul5rc0Rn8GCteguX79OiNHjqRr16707duX+Ph4IiIi8PX15bvvviuJLhfJggUL8PX1ZfDgwSZdp9Fo+OOPP4yK+Lnw8VrIeSG2BAdNNnbb9uHp4cGmTZu4du0aq1evZuDAgSaV+G7fvj1xcXHcvHmTVq1amRw5VBC+vr4EBwezbt06QkNDiYmJwdPTE0dHR/75559CHTRHjhyhUaNGVKxYsdC2lyxZgiRJTJgwweJ+ljYS8qmue3t7M3nyZPz8/KhRowaVK1emSxfj31QXL17M/v37OXLkCMnJySQnJ3P48GH279/P4sWLzX9GK80ZNYfmzZvj7u5O/fr1qVevHu7u7tSuXZvmzZtz7Nixsu6e7ZOAcL70wtA5445YkYyjdJ0zmYgJqHb4boYI7zaGC4h61CDy8gvLClYD/eHOZ3GkO980t6c6NKkP2f3jrzRE+LkOIxZF22LaSk4bhEzBPoQfKW8Qk4J8yGknvLy88PX15cwZ8Qe0e/duHn/ctFCz1157jbi4OEaNGsVHH33E1KlTi1xEMQZbsROKjbASHiIiNwIQEY0RiCjH1ygb58wNhKOiHYaV/oxBG2lphLRi4q+7kBxMj7DMy4P0C9w/9JfIdV2JqNwUj/iZdsS4n2F1RKTSMoSOzROIFCiFEkGxE2VnJywZ963OQePr68uVK1d0K/VpaWnMmTOHpk2bUrVqVc6cOcO4ceN06SpaodnS5NSpUyxZsoQVK1aYLET8999/4+npaZSgcNK+w0iZ5od86VM5W8XUCa/RrFkzi4TpHnvsMaKjo5k8eTKdOnUiKiqq0HLnxhIREUFUVBShoaFs2bIFSZJ0UTSFOWiK0585c+YMc+bMYfXq1djZ2VnUv7JCklRGbcXljN6+fZutW7dy4cIFrl27RkpKCuvWrTO6H2vXrmXjxo3UqVNHt8/f359169axZs0a85/PSgdUc3juuef43//+x61bt0hKSmL79u307NmTZcuWMW7cuLLunu2SiNB4aQbs0tvviKicdA4ROl7a2Y0LyV3dtEOUEDV2oVxff6YxxfY9+cBRsjWZRZ9kBM4PUgjfF8sQjNenLAw3hE5kH0TK088WtqdQOMbaCWNYunQp//nPf3jiiSeIi4szWSPl1q1bfP755zx8+JBp06bx5JNPMnToUHMeS4et2AnFRpQxEvAVoprQmwhnwu8IwdpqZdctXgFSETbC1Kl5AuCDSGcqhlu796N5YJpWSIE4Q/L7x4QeWjjgbWY7KoRD6igieqY7wklmujyighEodqJs7IQl477VOWhcXFyoXLky169fZ8OGDTRs2JC4uDgOHz7MggUL8qlF9+vXjz/++IM///yzpLpugEajYeTIkcycORNf3+JKceQnPj6ewMCi4hFzybpbsPaKOagdHci8fU+WtlQqFS+99BIHDhzgyy+/JCQkhJs3zV/B7dixIwBJSUmoVCoSEhIMHDQPHjzI5wQqykGTlZXFsGHDmDlzJnXr1jW7X2VNtpGbNmdUu40ePdqgnV27dlGnTh3c3d1xcHDghRde4MCBA0b3IzMzk+rVq+fb7+7uTmameS+GEtY7oJrDoUOH6Nq1q+57ly5dOHjwIG3atCE9XZFNlZ00hDhKPcQqnn6g4QBE+PV8hKegtDmN0IfRMgXhQDKWvPoz2YiySScR5U3XIMqCTwQGQ8bHd5AsyN3WJ/1GkiztaBmC0LN5G5H+ZLkbSSEvxtoJYwgKCuLo0aMkJCQQExNDlSpVTOrLvXv3uHz5MpcuXeLixYvcvXvXogUhW7ITio0oQ2KB9ojyz24IZ/42oPhXkpJlC0Jwd4aZfUmg6PQmPeQa27OzMsi8L8+7BCAWIWIRjqoliBDMU/I1ryBQ7ETZ2AlLxv1SW1c01kEDUK1aNbp06YKrqyvr1q0rMlLC0dGR0aNH8/HHH5dKJM2yZcuws7Nj7NixZl2ft4JTWloaiYmJuu369eu6zx3//ZfKMvVbkkBlL28kSb169di/fz8zZswgKCiI1atX89xzz5ncjkqlYvz48Sxbtow+ffoQExNDUFAQmzZtwt7eHgcHB9LS0nB1FXUONRoNBw8eLDR644MPPqBChQq8/LKp8aLWg4TK6ApNxeHn58ehQ4dITU3FxcWF3bt307JlS6OvL6qMeXElzgtFVT4qbxhLjRo1eP/99wkLCwNg06ZNeHp6otFolFKqciIB0YgV0LyZsE8idASeKu1O6ZGNWFXUpvnXR0y+C+IBIgIo76Yf3PYtwiGjKfyWarW9WI2UYfKhdpA/2rAuIt1pJqL09tqcfQqWI6edkIP27dvrtoiICHx8fCxr0IbshGIjyoDLCO/wekSUxkrgJURUY1lzB1ExKhDhxDeVdEQpvW6Iykh57ch1w++qq/I8tEqllv1dAmdEycCuiP+flghbPhpFxEwGFDtRdlgy7peog+ZqYhpH426TfDsFJ7encHbJvwqvz8WLF3njjTe4ePEiI0aM4KOPPjLKcI0ZM4amTZsyd+5co1KHzOXixYvMnDmT/fv3F9kvjUbDzZs3DRwv2u3rr7+mcuXKREdHk5iYSHp6Ol5eXvm25s2bU+nXv+HcFVn6np2RgUMVudw9uTg6OjJ//nyee+45hg0bxqBBg5gzZ47JL+5Dhw5l+vTpfPrpp8yePZtNmzbx1ltvAbmVnLQOmvj4eGrWrFmgOPPvv//OokWLOHr0aLmf9Mgl6tW6dWv69etH8+bNsbe3p1mzZvmibIoiPj6+wL8rSZLMKtetu96GDO+GDRuYOXMmoaGhALRr144NGzag0Wj46quvyrh3NsJBRNTIoTz7fRHRMmGUfTWIpYB+cFooQpy4oEmzMdHmRixUOkhVUKscycYyMXnUapy8LK/OVhAOwGyEoyYE4V+zLKBZQYucYvKWkpAg8vNSU1N19tpSrOjxLEKxEaXIA4QmygcIp/7biEGnUll2Kg9TEELF35M//TUbSCKfk8VguwBkIdJpFxbQfnVEuT0voD44n/IwjM40E7W9Q4m8SwDQAxEV9BIwFlEifCVlm4JmIyh2omywZNwvEQfN6bP3Wbrqb/786z5qNWRmZeNb/2X6jzzMk82r8kp4Xbxr5FYAunfvHnPnzuWzzz5jwoQJeHp64uvra/QLtre3N507d2bt2rWMHz++JB6J7OxswsPDGTZsGImJicTFxRXogLl+/TpJSUlUqVKFGjVqGDhdatWqRWZmJm+//TYtWmwrkxkAACAASURBVLTQCbYWpmNz7r4dZ9+LIjvN8qTMm64OrF6/jpCQEFlKJOelc+fOxMXFER4eTtu2bdmwYYPREVMgnDBDhgzh+PHjXLt2DTs7O/7991/u3Lmjc9B4egqZ98LSmzIzM3nxxReZN28etWrVku3Zygq5ymwDzJw5k5kzZxZ/YgFoNEUs31uAtQ6o5lC9enWWLl1a4LGAgIBS7o2NcQkxuY7Os78C8BZCYFEeG18wEmK1s4BVSYPtKvkVcRfI1Ac1QhjGC6hB7sTbCzzpyJm3IrHUP5Pu7MSXIc/SCRHpUhK1ZZ4BfkXIBu0EPgbZokQfVeS0E5Zy8OBBwsPDefDgAZcvXyY+Pp5PPvmEZcuWmd2mrdgJxUaUAhrgS0Q+ZSIwGJgLWMt0UBs1+T3C8fAsIkryYwxtyQ0Kjpp0JXf8rw5cRKQGBWJgE/Ag3wDutbcLV16IQfMg1aJHkNI0eEzvACcQCxAd8t/LImogBIM/RNj9JxBhl51kvMcjiGInygZLxn3ZHTSHjiUzfd4fPEw3zGZTqZ3IyJTYH5vEiZN3iJofRB1fZ1atWsV///tfnn/+eU6ePEnNmjVZsmQJZ8+eNem+ERERjB07lnHjxpkk3JuRkcGNGzfypRfl3a5cuYJGo+HSpUvExsYaOF/q1atn4IjRan3k5e7du7zzzjsMHjzYKOeT34h+nJ1V8H+sKdhVdKXC0Of5eudOJk6cSGBgIKGhoYSGhuLv729x+1qqV69OTEwMK1asoH379syfP58RI0YY/f8xbtw4OnbsSI8ePfjuu+9o2rQpCQkJ+YSCf/31V503Up+5c+fi5eVFeHi4bM9UZkiQbbngvtWizRkt7/Tq1avI3+9t27aVYm/KL4n/PuT+gyxqeDpTsUKOWbqHiIyJRIRza1EBIxAlPmtYcNM0xES4iLBw3Wah86NQKmM4sdYKWYJYNdxN7mS8kKjyCtTG7etG3DkUZ1FXqvjVoE2LJmxAVJttjZh/dwUKr5VnOlUR+pKrELIQnyIqRSmYgZXZiddee42dO3cSEhICQGBgIL/88ovZ7dmCnVBshDxIksSV62k8TM/Gt4YLzs55BsQ9iAjLeMSAsgWhZ1LSZCIiYQpz3BcVNbkL2ItIv9LagCAMbYK+U15/IJ6C0CWLxKg3uWod2+BQuZLFDppqTdvgXMtTOJiigCpAT3KNRQWLmheoEf+XwcAghCPrDURZQaUqvekodqLUkWPcl9VBczUxjenz8ztn9JEkSEnVMG7qURLPTKZa1Yr873//o3nz5rpzatWqxe7du026d4cOHbCzs2PPnj0EBweTnJxcYHRL3n3379/Hw8MjX4pR48aN6dy5M15eXtjb29OrVy927Nhh0E9TOXnyJE2aNDE6MsixWhU8ewZzffNOi/66VGo1fRe+S39HRx4+fMju3buJiYnh/fffx9PTkz59+hAaGkpQUJDJVany3Uul4uWXX6ZDhw4MGjSIHTt28OmnnxolItWwYUOaNm1KtWrVdDo0eSs5SZLEr7/+yqJFiwyuPX78OB9//DEnTpyw+BmsAW1ZPFvG2gZUc5g8eTIAmzdvJjExkSFDhgCwceNGXcSXQuGc+use7y/9i3+upeFgryIzM5tO7TyY6FoP13ftxMRXn06ICWlhwoga4BbFT5QTgbsl8EBa3IE6FD7R9kJMyl3yXPcJuQ6aJzFOADIN6rqMIk6agkZl3uTbroIr9aaMxgeRgpSFyNSKAd5BaMaE5hyTI/5SBYxEOGhezGn3LaxDGqI8YY12Im/xBEurKJZ3O6HYCMs5dCyZyBVnSbqdgZ1aRXa2RO9uNXj5JX/sz6uFs2IbIlImGiEUb8nvjX7UZHGRk7dyzs/LY+SO+U+SO+7vRaTurEI4NqpjXmpuPEZV+dOiUqnwfz2c029Fkp1pXlS+XUVX/BeHiwpYKQjR+hjgO0SUizPwHMJY9MJyY9EMOIaIkp2PWLDYgCiRrmA0ip0ofeQY92V10GyKuUJWpnE60Cmp6fQdNIv/vvl8vhdqPz8/Ll26lP+alJRCI1y0zpaePXui0WioVKlSgdouQUFBeHl56SJgqlatWqzDZMCAAYSHh1vknIH8AsHGcLdLa9K/2Y6TmX9cdq4u1H1zLOocTRhnZ2d69OhBjx49WLFiBYcOHSImJob+/fuTmZmpi6x5+umndaXMzaFx48bExsbyxhtvEBQUxNq1a+nQoUOx10VERDB37lxOnz5Nr1698jlozp49i5OTk0EKU3p6OsOGDWPRokV4e5tb88/6sKacUdkpJ5U3iuOZZ54BYNKkSRw9elS3v1evXiYJMT+KnD3/gFffjtc59DNyolR27/mXi/dS+OTf5thpx70AYBKiROoZhJhJQQ6YfzG+FIGpVCdfehFeCCfLG4A2yO8/GAr9mkLeCk7FcRoYAB4nO1LRzp970hkklWm1klQO9jj7eFFjQHfdPntE5HoHYBFCFiAGeB4RZR+as1kaf9kQkfI0DeiM0EP2s7DNRw1rshO+vr4cOHAAlUpFZmYmS5YsoVGjRuY3aAN2QrERlhF7PJlpc/8gPcNwYI/Zfp3rPz1k7uYmYgyehyjT7FxEYw8xLtKlsKhJJ3LHfX+EIH1BNsGzkH7EIjzRYxFRoJaQgBiQjeUW+H7flwsZX/JQfR1JMs1Qqp2dqNyiCdWeaS12VCDXEGQhBvIYch02aqAd0AfojfnGogIizLIrIje2GSJy50UUAWETUOxE6SLHuC+bgyY9I5v/7U4ky0i5CrXamT/OOesEdfWjW86fP8+ZM2cYMGCAQfRLVlaWztGin2LUqlUrvLy8cHNzo0+fPsTGxlKvXj1ZnmvLli3Ex8fz5ZdfWtyWKSW2s7OzmT9/PkuXLmXVG6+i+ijaZC0ataszHr06UXfyyAKP29nZ0a5dO9q1a8eCBQv4448/iImJYcqUKVy8eJGePXsSGhqqq6hlKs7OzixZsoQuXbowcOBARo4cyYwZMwpM/9LSs2dPXn31VVq1asWdO3eIi4ujXr163Lsn1DJ/+eWXfI6emTNnEhAQoPNQ2gYqNNY24siMnKrrO3bsYMKECWg0GkaOHMmbb75pcPzSpUuMGDGCmzdvUrVqVdatW4ePjw979+7l9ddf1513+vRpoqOjCQ0N5aWXXmLfvn1UrixUMlavXk1QUFCB909JSeH8+fO6lMELFy6QkmKMCuyjy7Iv/i4w2jJTLXGpYhqHPJJpd6uaCKs4B5REUTb9nP4C9F2KyunXEUauc6Y6InfeXBL0PhfnoFkPjAFSQI09rTWf82vFF3iY/S9ShnE5WSoHexyrV6HNrjXYOTsVfA4ikCcQ+C9C9mAroljVXcRCaSgiMt+cEcsJoeW5EzEHfw/oZ0Y7jybWZSdWrFjBhAkTuHr1Kt7e3nTp0oWPP/7YojattTqHqSg2wnQkSSLyk3P5nDMg3jkOp97m7NAH1JtYUTjmC3Lc6ztgCoqaVCEiPbRjfUMKj3qsjPlOgQzEoFkDIWBsCTdyNiNLbPMbEAb2NyvSduY6fl36All37iEZqS+odnbCtY4vrWJWFByhbo9IRwpG2L84hKNmCyJVaSLCnmkdOuYYi76ISKShCBHhHcByRKSSQjEodqKssGTcl81Bc+Pfhyanlly7kYqLawWqVa1iEOXi6emJJEk8//zz+Pv76/a7ubkVe48XX3yRL774grlz51ryOADcuXOHiIgINm7ciItL3lh000lISDDKiXDz5k2GDBlCamoqR44cwcfHh2uBQcSHv0n2w3SRJ1YMdhVcqNG/O02XzzLq/0WlUtGkSROaNGnC9OnTuXz5Mtu2bWPp0qW8+OKLdOrUidDQUHr27En16kVX48pLjx49OHHiBMOGDaNDhw6sX7++UO0bOzs7Xn75ZXbs2MHx48c5ffo0gYGBugiavALBhw8f5vPPPyc+Pt4mUpu0SFiXx1tu5MwZ1Wg0jB8/np9++gkfHx9atWpFSEgIjz/+uO6cyZMn8+KLLzJs2DD27NnDW2+9xdq1awkODiYuTmh3JCcnExAQQJcuXXTXLVy4kH79in9dXLx4MR07dsTf3x9Jkrh06RKffPKJPA9og2RlZXM84U6hx9PsNezwSaTdv9VMj4ixwzCnv7DJdt6cfnPYCmzS+/4RwkljDhKGDprCJt9pwATgM719anCYXZmnx2zh8LPhPEg4h4bUIifBdhUr4FqrJq1//BInD+PLZNTOuf0ERHT/94gS2meALoj599OYPrnoini/Go6Yey9BHjkDW8ba7ET16tVZv369bO1Zo7aAuSg2wnRu3Ezn5q30Qo9n2mWz9+C/1GtawEBekdwxvyki9aYw53uJ1rPNYT7wOyIVy9Jis1o7UZwTPxtR4WkaYuA+BC7NavL0sBgOPfci6f/cQpNexMuiSoWdqwuVWzalVcxy7CsaMSKrEFEuzRCG4W+EnYxBeN9nIVLRtM6a9hj/8/dFpDnNR6wWHESkPD1l5PWPKIqdKDssGfdlG5ayNKZnuKnVau7cuU8F1/wlmb/77jtatWpFkyZNTGpz3LhxdOjQgRkzZuDsXFSsY/FMmTKF3r17G5WaUxzZ2dn8/vvvNG3atMjzfv31VwYPHsyQIUN47733dGlGNfs/T4W6fuwd8yaq+L9wdHISzho9VHZq1E5OuNT2JuCtl6k5sIfZTgs/Pz8iIiKIiIggOTmZH374gZiYGCZMmEDz5s0JDQ2ld+/e1K5d26j2vLy82L59O0uWLKF169YsWbKEwYMHF3hueHg48+fPR6PR4OvrS2ZmpoGD5o033gAgLS2NYcOGsXTpUpvM5TbCD1eukWtAjY2NJSAgQOf0CwsLY+vWrQYOmlOnThEZGQlAcHBwgSLT33zzDc8//7xZ0WLdunXj7NmznD59GhB6Sk5OBUckKEBWllRg2r4+qfZ5Vvceo/AIF30HTDVKp9z2XYSarpaeiGgac7lEbiSOI1C/gHPOIPQV9B05NYGNQAdwwI12zTdxK+4g51UrSVYfQ+WqJjtDpD2pHRyQsrN5rGVT/KeMwqPr06gsyP2ujljMfIlcSYIvEYE9TyHm310wvsiWB8LhE5Vz/ReAZYnFto812IlZs2YVekylUvHOO++Y3ba1TrxNRbERppOeno1aXfgvQLYK0tpkQwT5U4zkVDa3lFPAbITgbS8Z2jPGQXMLGAb8D+iPEPXNcQy5+HjxTNz/+LfOXv5O+Yx72adBrUbKyAS1CrWjA1JmFtU7tcV/8kiqPt3K/AXQuuRG0WjLim8BViC88FXJDcM0xljYkZsTOxixGvBfRCn10nC0lVMUO1E2WDLuy/brXK2KI5lZpi11OjupC3TOgBAKvnTpkskOmgYNGhAUFMTXX3/N0KFDTbpWn927d7Nz505+//13s9vQ5/z581SrVo3HHis4Hi87O5v333+fJUuW8MUXX/D88/mTS92aPc67mf8w5/N3qXsxietfbSfzzj2krCzs3SpSpU0QdSYMp3Kzxwu4g/lUrVqVoUOHMnToUNLS0ti1axdbtmxh9uzZ+Pj4EBoaSp8+fWjatGmRg7hareb111+nY8eOOgHhqKgo3NwMlxPc3d3p3bs3Bw8e5IlmrfCr24THPGpz9uINXCtW1uUqTp8+naCgIPr37y/r81oL1ibqJSsm5IzevHnTIGdz9OjRjB49Wvf96tWrBoJjPj4+HD582KCNwMBANm/ezIQJE9iyZQv3798nKSmJatVyIweio6OZOHGiwXXTpk1j1qxZdO7cmfnz5+cbWBcsWMDUqVMBocqu/7v49ttvyxLJZ4s4OanxdHfm+o2C0zad7dS06lVFRKQUldNflkwFruV8roQIt7bkT1Zff+Zx8qdUbQRGI0q1aumCEGf0yPmeCKq1atyldrhL7Uj79Dq37A6QefsukiTh8Jgb1Z5pTYW68iu9FCZJ8BbQIGe/VhOzKFSIyrHPIBw/gxFzeyuNYC5zrMFOVKiQf2U9JSWFVatWkZSUZP7E2wq1BUxFsRHmU9PLmaL8Aq4udgSNqmzdERQaRGqTG8IhIQcJCMd8YYPpfsRiwb+IEt4vk882qX+wx+v6c3h99RwPnjhP8m/HyLxzF5XaDoeqlXHv0h7nGh75mrYID4T2zgiEHduJMBJbEZ59F0QopdZYFBXY2QaRSjUe4aD5CaH9Zi0l1a0MxU6ULnKM+7I5aCq7OfBEo8ocKyJs3eDG9ip6POtV6HE/Pz8uX75sVl/Gjx/PnDlzzHbQpKamMnr0aJYvX57PeWAuRQkE37x5kxdffJH79+9z9OhRfHx8Cjxv165dZGVl8fzggajVaupPj5Clb6bg4uJCr1696NWrF1lZWRw4cICYmBh69+6NSqXSiQy3a9euUFXuZs2acezYMSZOnEizZs3YsGEDrVu31h3PliTGv/YWPQbep0ZNXyRJwt7enqQUNZ+s/5E/L6dwO/Ec0dHRJCQkFHiPco+VlcUrCYwdUN3d3Q1Etszhgw8+ICIigtWrV9OhQwe8vb0Nfj+vX7/OyZMn6dq1q27fvHnz8PLyIiMjg9GjR/P+++8zY8YMg3ajo6N1g/C8efMMBuEdO3Yok+9CUKlUjBhUi0XLzxaoQ2PnpKL7a17WtQqqz16EcKGWhUDBw7bxFKY/k4YQv9S/nxoRKv4Whp6Lj8gVt2wALsNr4Kvua2HHTEdfkkAiV5KgC0LGIRShG1m7iDaeQLxnTEZoYa7GsqrqNomV2IlJkybpPt+/f1+30BQWFmZwzBysbeJtKoqNMB8HBzUDenuzccsV0vPYCZUKKrja0a61uTmlpUQUcAjhPJCjBB4IZ35BrxPZCEGvtxGOioMUHIIoAQsQwr0vQEU7fyo2sFTy3UQqInRl+iJKlf+CiKzRCg3bIaJjijIWbogFiq6IaNZAhJ0cULJdL3codqLUkWPcl3VR6j99fXF2Mq5JO7WKfr0Kn9FqI2jMoUePHty4cYMjR46Ydf2MGTNo06YNPXr0MOv6gijMQfPbb7/RvHlzAgMD2bt3b6HOGRAeucmTJxtdpruksbe3p0OHDkRGRnL+/HliYmJ47LHHeO211/Dy8mLEiBF89913pKWl5bu2QoUKfPLJJyxYsICQkBDmzp2LRqMhU5PNqYsPULvWoLZ/fZycXXB2ccXewRG1nT0ODo48SNOgcfRkY8wvVKlStQyevOSRgOxslVFbeURCiHoZsxWHt7c3//zzj+77lStX8lXzqlmzJps3b+bEiRPMmTMHwCCa7auvvqJPnz4GAtY1atRApVLh5OTE8OHDiY2Nzf8cenGjUp4Y0rzfFQzp1smTAb19cHRQ4WAvfo9dXOyoVNGeD997AreKhYuJlympiGoSWjrk+W4uBenP/AW0xdA5UwPYgwjz1v/7uI+I4tEyGasIO9FKEswEjiMi7TWIIhwtcvbHU3ClWhfEAvB4RIX170uhv+UJa7ITycnJTJ8+nSeeeIKsrCyOHz/O+++/j4eH+avwctqJskKxEZYxPKw2we3ccXRUo11TcXWxo3pVR5bODcLezornQBcQzpLuiFBAOchEpEzlfZ1IAkIQFQX7IAbbwvJDfwMOIyojWlbdWB4cEClLUcA/wBHgTeAmYnGiDuJZZiHsZN4/myGIVYCGwEByI3QUAMVOlIWdkGPcl7W7TzavStdgz2KdNJKUQfhgH3xqFi68a0kEjZ2dHePGjTNLFTo2NpZ169bx4YeWlOLIT94KTtqUpn79+vHJJ58wf/78IqsbnThxgj///LNQ3ZayRqVS8cQTTzBjxgyOHz/OkSNHCAwMJDIyEi8vL/r27cvatWtJTk42uK5v374cPXqUH3/8kee79yDh3B1S07OL9fY6u7ji4lads1dTbXaSk43KqK1ckhOSaMxWHK1ateLs2bNcuHCBjIwMoqOjCQkJMTjn1q1bZGeLFbh58+YxYoRhjcuNGzcyaNAgg33Xr18HxGAaExNTYLqlfkpf3vQ+WxKtLglUKhWjh9Zh4ydPMmpoHf7T15ep4+uzdU1bGtWXJ3KxRPgvQvgQRNrVZ8hjSfOW2I5GeDD09z+HmIg+U8D1KwFtAKsnYtJqhWglCX4BtiN0H6cDjfX2560tEoLQhlyCSH/K7/J/dLEGOzFlyhRatWpFpUqVOHnyJO+++y5VqlSxvGEZ7URZodgIy7CzUzH99Yas/qgFIwbVZkg/X2ZMasg3q9oU+Q5R5kgIMS41Qm9Frv/qMwgnjb6I/EGEF/wnYCnwFSJUsTAWkCsgZm2ogJYIzZ7fEYsUCxDe+ncRz61vRLTGwh+RVzsdEW7ZHLAs8NqmUOxE6doJOcZ92f1Jk8fV062KOjkaNp+teYiDA1RxOsC2r9/SvTAVhCURNAAjRowgJiaGW7duGX1NRkYG4eHhLF68GHd3uWIRBfoRNLdu3aJnz55s27aNI0eO0L1792KvX7hwIRMmTMDRsWDNHmujdu3aTJgwgb179/L3338TEhLC5s2bqVOnDs8++yxRUVG6qAdfX192797NuNffJfVhptH3kCS4n5pF4u3CVf7LM5Jk3FZekWtAtbe3Jyoqiq5du9KoUSMGDBhA48aNmTFjBtu2bQPg559/pkGDBtSvX58bN24wbdo03fUXL17kn3/+4ZlnDN96//Of/9C0aVOaNm3KrVu3mD59er57x8fH4+bmRqVKlUhISMDNzU33/eTJk5b9gB4RPN2dGfyCLy+/5M9zz3jg6GDFy+FHgUi97zMpWMzXVFIQpcS1rEMISmpXAdWIChjbydWb0ScTWKz3fQLWp9lTAFpJgu+AWKAdwt/VKGf/NnKdMTURkgV+iMIf+upwif8+ZMPmf1i++jy7fvmXjExTS3+VX6zBTixatIhr164xe/ZsatasaTAOWpomLufEe8eOHTRo0ICAgADmz5+f73h6ejoDBw4kICCA1q1bc/HiRd2xefPmERAQQIMGDdi5c6du/+LFi2ncuDFNmjRh0KBBPHxoqKul2Ah58PN2ZdjAWowd5k/71tWxs+bIGRCaKj8hSmr7FnOuKegLBGtTmjog8kr3IwSTi/rRnEKEIkZgvHp7WVIPmIJ4tmuIaNJGiNDKZxAadeEII5KFsJN7EYajLcK5k2MOLv2TypebLrHiy/P8FnsLjaYcT6BNRLETpeugkWPcl13zWrsqOrC3D9/9dJ19+2+R9lBDpYr2qLMucPfmL6z8bBmdOnVixowZzJ49u8B2LImgAVHGq0+fPqxatUpX9ac45s+fT61atQgLs6QUR37u379PYmIiAQEB7N+/n0GDBhEWFsacOXOKjJrRcvHiRXbu3Mny5cuLPdcaqV69OsOGDWPYsGGkpqby448/EhMTw7vvvkvt2rVFRajQPnjXaWTyAJEtQWJSBl5VnGxqNUpCZVVl8UoCOR+ve/fu+Ryd+orx/fr1K7Rcdu3atbl69Wq+/Xv27Cn2vhpN3rV+BZslA+E10L77N0es4snBH+SGbdsjJvdavBACwR2LuH4TIjQchFrvWJn6VYoUJkkwFRFdEwr0QMzVgxEZA6MlCfs1F/hq6xUkCTKzJFxc7IhccZbImU/QsF6lMnmW0sJa7ERRi22WItfjaTQaxo8fz08//YSPjw+tWrUiJCTEoNrfqlWrqFKlCufOCY27N954g02bNnHq1Cmio6P5448/uHbtGs8++yx//fUXiYmJfPTRR5w6dQoXFxcGDBhAdHQ0L730ksF9FR4xEoHXEZ5kucfieERKkDtCm+V7xKC5ElHpsDg+QESjjJe5X6WBFyKdeBRwD9iB0Kv5Bvgc4XDqhjAW+xDG4w3Q/Cgxt+Np9p64hUaTjUYj0uQqVrBn6dxAvGtYcSSWDCh2wnh27NjBhAkT0Gg0jBw5kjfffNPgeGRkJCtXrsTe3h53d3c+//xzatXKr0wtx7hfYkXJKrs5MKSvH0P65laLSE6uRd26r3Pv3ny2bt1KmzZtCAgIMDBmWry9vblx4waZmZlGOTEKIiIighdeeIHJkycXKlir5dSpUyxdupTjx4/L/qL/+++/8/jjj7No0SIWLVrEqlWr6Nmzp9HXL168mPDwcCpXLipmsXzg6uqqExLOysrit99+IyYmhmWfrWfQ8FdxdjZ9oNRkS9xLzaJyBSvVrDAHKxH1KkmswF4oKBjPAkC78GEHrEI+C6qfxpSl9/lZRDSNZxHXSgiRYi2jARmihssSrSRBZ8TjHUPMwxchCnuEInxS03cncuO7q5CZO1impWlIAyZMj+ebVW2oVNGGa6/KbCdq165NpUqVsLOzw97e3mJxdjmQy07ExsYSEBCAv78QQw0LC2Pr1q0GDpqtW7fy7rvvAsKpHxERgSRJbN26lbCwMJycnKhTpw4BAQHExsbi5+dHVlYWaWlpODg4kJqaSs2aNeXpsEL5RZuHuRL58xQSEIK5TwLXESlN4zEuheoawp6MpvhyetaOG0IMeABi8eRncgWGNyNs9DPAIFiVcJF9B26RYZfrIEhN05D2UEPEW3F8vaqNdWsZWYpiJ4zCGCd+s2bNOHr0KK6urixfvpypU6eyadMmy29eAKUaS161alX69u3LypUrcXd354cffuCNN95g7969+c61t7fHy8urwJVtY2nRogU1atTg+++LlhbUaDSEh4fz3nvvGZTrlYv9+/dz48YNYmJiOHLkiEnOmaSkJNauXcuECRNk71dZY29vT8eOHfnwww8Z/9pbZjlnQAw8yfeyij+xnGENIYklhYT1hSQqKBTKn4jQaS1vAEEytf0QUX1JHxUifWoHRTtnAH4kN+zdDiGqaEPoSxLEIeQc0oFRkkTixktQQBUwAI1GYvvuxFLrZ1kht53Yu3cvcXFx1jHpxng7cfPmTVq2bKnbPv30U4O2rl69ajC/8/HxyTe/1D/H3t6eypUrk5SUVOi13t7eTJ48GT8/P2rUqEHlypXp0qVLyf1AFKyfzYiIjneBQAoY1AAAIABJREFUBjK3LQEHEOmwdjmfi0tp0mcJQrNFrshPa8ERUSJwGXAFUTVrCnANMjZl843PVR7a5bcTkgQpqRoOxCaVanfLAsVOFP8+oe/Ed3R01Dnx9QkODsbVVeQGtmnThitXrpTQU5VBjYfx48ezYsUKsrKyaNiwIRs3biQsLIzTp0/nO9dSHRoQUTTFiQVHRUXh6OjI6NGjLbpXQRw4cICZM2dSr1499u3bh5+fX/EX6bF8+XJ69+6dryqNraGx0L2bqbEtzQEJ0Egqo7Zyicr6VNcVFApEg8hx1ytfzTsytX0OkSevL6jiBuwCZmBchQ396JlBCJEWG6YeokDVjw+zsU/KKPS8h+nZxMbdLrV+lQWKncjd3N3dOXr0qG4riflcXm7fvs3WrVu5cOEC165dIyUlhXXr1pX4fRWslNuIaJYgRIUkOUlGpO/cQ+R8Hkd4ro3lHsK73Q8hqGurqIHWwDzgT7i6J61IPbbUNA1xf9wtpc6VDYqdyN2KcuQb48TXZ9WqVTz//PMl9lil/vrTrFkzfH19+e677wDo1KkT8+fPp0ePHty8edPgXEt1aAD69+9PfHw8Z86cKfD4hQsXeO+99/jss89kLV+dnZ3NwoUL6dOnD76+vrzzzjsmp2qlpaURFRXF5MmTZeuXtaKyUD3chuRndNhyBA0oETQK5YRliCoZWlYijwDv1wgdm7g8+7cgakobw3FEeSMtU2ToVznB3l5V7ADo6mwNNWRLFmPtRHERJiA0BLt06UKLFi0KPF4WyGUnvL29dYUJAK5cuZJv4Uv/nKysLO7evUu1atUKvXbXrl3UqVMHd3d3HBwceOGFFzhw4IA8D65Q/piMKA29CpGnKReHEFWatGP9IozTm9HnU4ST5hGyEQCODdRo7Aq3E2o1uDjb/kqgYifEJpcjf926dRw9epQpU0ruD6pMkrMjIiKIioqiT58+AAwfPpxz584RGhrK7t27cXYWs185ImicnJwYOXIky5YtY8mSJQbHJElizJgxTJkyhfr15SjFIUhKSuKll17i5s2bHDp0iMDAQF0FJ1NYs2YNrVq1onHjxrL1zVpxclRhQgGnfDhbc+UXMynPzhdjUJwvClbPJeAtve/jEcKPlvAQMZEvKLDT3sT29aNnuiIqezwiODqoCWryGMcS7hR43MVZTbdOxeWHlX+MtRPaiWlR/Pbbb3h7e/Pvv//y3HPP0bBhQzp06CBDL81HLjvRqlUrzp49y4ULF/D29iY6OpoNGzYYnBMSEsKXX35J27Zt+eabb+jUqRMqlYqQkBAGDx7MxIkTuXbtGmfPnuXJJ59ErVZz6NAhUlNTcXFxYffu3bRsaUpYg4LNsBshVPsmwvEuBxLwIULs1gehbfMhpqfXZuRc1wnTom5sgJqezlSv5sjV6w8LPO7goKZTe3mr9lojip0oHmOc+AC7du1izpw57Nu3DycnJ8tvXAhl8lbbt29fTp06xZ9//qnbp9V/eemll3RKz3JE0ACMHTuWtWvXcv/+fYP9X375JUlJSUyaJF8s4sGDB2nevDkNGjTgl19+AaBy5cpUrVrVpHY0Gg2LFi0qUe+cNeFZxQm1mX9gKhW4P1Y+yo8biyRBtqQyaiuPKBo0ClaPBIxBlMAGUSp1noVt/o2oJa3vnNFfCW2EyKc3hgvAV3rfp1rWtfLI+BH+ODvln8Y4Oqio7VuB1s1Ns7vlDbnthHYy6uHhQZ8+fYiNjS3J7heLnHbC3t6eqKgounbtSqNGjRgwYACNGzdmxowZbNu2DYDw8HCSkpIICAggMjJSV4q7cePGDBgwgMcff5xu3brx8ccfY2dnR+vWrenXrx/NmzenadOmZGdnl0pqlYKVkYKoLFQPkZoqB7cRaugTgZ6IaMk7CE0yDxPb2ghc5ZGLngER7fH6mHo4Oea3E05Oato0r0rd2hXLoGelh2InjLMT+k78jIwMoqOjCQkJMTjnxIkTjBkzhm3btuHhYeofommUiYPG0dGRUaNGGWjDqNVqVq9ezeXLl5kxQ4xwckTQAPj6+hIcHGyQG5yYmMjUqVNZtWoV9vaWBxJJksSiRYsIDQ1l6dKlfPDBBzg6OpKQkGBW9MzWrVupWrUqTz/9tMV9Kw9UrmCP2kwPTQUnO5wdbS+U3aZTnIwcTBUHjUKZsRbYqff9E8CSqs3fIFZWj+vtC8awDGugCe0txrDkd7AFfSun1K9biSVzAqlbuwJOjmoquNrh6KimcwcPPpobiJ0tV+bIQS47kZKSolvESklJ4ccff6RJkyYl3PtikNlOdO/enb/++ou///6badOmATBr1izdJNzZ2Zmvv/6ac+fOERsbq6v4BDBt2jT+/vtvzpw5Y6A7MHPmTE6fPs3vv//O2rVrS3RFVcFKeQfhMF+JKGFtKbGIlKbtiMiXzYjKfAmYHiWprfLXFBFl+QjSpkVVZr/1ODU8nXF2UuPqYoezk5o+3Wsyc2qjsu5eqaDYieLthDFO/ClTpvDgwQP69+9PUFBQPgeOnJRZ/ckxY8bQtGlT5s6di5ubGyCMo3757VatWskSQQNCnPjVV19l7NixqFQqXnnlFUaOHElQkOWlOJKTk3nppZe4ceMGhw8fpnbt2rpj8fHxBAaaMusWzp4FCxYwZcoU2Ut+WysqlQqf6k5c/vehSeXg1Crw8ZBDEML6sDHd43woAsAKVssNDKshDQHM1YJLR6Q0RentUyFWWt8BBuvtN3byfQvxMqBlKsZX8rAxGjdw48ulLbl+4yH3H2TiXcOFCq42XFo7D3LZiRs3bujSzrOyshg8eDDdunWTp3ELUOyEglVzGFEd6WXA0iwPKaetqUBN4DdEOW2ALOAPRNUmU9iec90aHlkbAdC2ZTW++qwq/1xLIz09G9+aLjg/AhplWhQ7YRzdu3ene/fuBvtmzZql+7xr1y55bmQEZTaL8fb2pnPnzqxdu5bx48fr9ru7u/P999/TsWNHPv/8cy5fvowkSRY7KoKDg8nOzmbfvn0kJyeTkJDA2rVrLX0MDh06RFhYGH379uWbb77B0dEwPj0hIYF+/fqZ1OZvv/1GUlISoaGhFvevPOFRxYnU9Gxu3c0wykmjVoGfhzNuNjgZlwDJhsNHtCGJCgpWySuIEHMAd0S0ijn8DQwEjunt8wDWA8/mfE/QO2asg2YZkJbzuTbQ18z+2RA1PJ2p4WmbzvrCkNNO+Pv7Ex8fL0tbcqHYCQWrJgNR4a8mMN/Ctm4DI4AYoDfwBSJqRstZhLPftPVeWIBIzw2zsH82gEqlws/btay7UeoodqJ8UqZrE1qxYClPXFWjRo3YuHEjI0aMwN7enqQky2vUq1QqIiIiiIyM5JVXXmHVqlU6MWJzkCSJyMhIQkJCWLJkCYsWLcrnnAHMSnFasGABkyZNws7u0fHuaqnl6UzNak6oVBSqSaPOOVanhgseVWw0nNjIcMRym+KEkuKkYKXEICosaVkKVDejnW8RqUf6zpmOiKpNWudMGvCX3nFjJt9pOX3SMokyXGpRKFMUO6HYCYWyYx4iOmUF4GZBO7EIW/E9EImo5Fclzznad2JTXidigX2IaFA5q0oplC8UO1Eu7USZOmg6dOiAvb09e/bsyXdMW3774cOHxMXlrUNqHkOHDmXnzp08++yztG9vfimO27dvExoaSnR0NIcPH6Z3794FnpeSksKVK1dMqhB16tQpjhw5wrBhw8zuX3lGpVJRs7ozzQIq4V3dCYccDQHt35Wzg5pani40q+dGNTfbEgbOS7Zk3FYuUTRoFKyRO8A4ve8hwAAT20gHJgD9EGVNITelaRdQQ+/cU+TqyLgjBCCLYzUixQmgKjDcxP4p2BSKnVDshEIZ8AcwB5Gi2sPMNrQpTe0RduA34HUKTkVKQDjiG5rQ/kKgMkLAWOGRRrET5c9OlKmDRqVSMX78eAOxYH2GDx9OrVq1iIiI4OHDgkukmUJsbCyOjo54eppfevPw4cM0b94cf39/fvvtN+rUqVPoub///juNGjUySYT4gw8+YPz48bi4yKE0Vn6xt1NTo5ozzeq50bKBG83qudGqgRtP1K2E+2OO2Jlb8qmcIEIS5fF4nzlzhqCgIN3m5ubGhx9+WOLPUBy2OKAqlHOmANdzPrshUolM+R08j5hsf6S3zx0hNjwTyBsUqR8p/IQR99IAi/S+RwAVTOifgk0hp52wVhQ7oWB1aBCpTZURIr7mcAeRmvoa0A04AbQu4vwERJU/Y4PGzyHEhV/GMnF7hXKPYifKp50oc/m1IUOGsG/fvkLFgLt27Yqrq6tB+W1zSElJYdSoUXzwwQesWbOG9PR0k66XJIkPP/yQXr16ERkZyeLFiwtMadLH1PSma9euERMTw7hx44o/+RFCrVJhb6d6ZASTtcg1oDZo0IC4uDji4uI4duwYrq6uOpGvssQWB1SFcsweDIV3FwLeJly/GRGmflRv3zOIlKbnCrlGX3/GmPSmLQhdGwBnTBeMVLA5lIm3YicUSpml5IoDu5tx/RGErfgO4XDfioiGLIp4TEtvikRE3LxqRv8UbA7FTpQ/O1HmDpqKFSsydOhQVqxYUeDxWrVq0b59e4Py2+bwzjvv0K5dO8aMGUOTJk349ttvjb729u3b9OnTh/Xr13P48GGjX25NddAsWbKEIUOGUK1aNaOvUbBRJKG6bsxmCrt376Zu3brUqlWrZPptJBJCdd2YTUGhxEnFMAy8IzDSyGszEClNfYG7OftUwHRESlPNIq41RSBYQgg+ahmOeS8HCrZDCdkJa0GxEwpWx3lgGiKtaZCJ10oI5047RFWmX4CJFB85mQxcwXgHzU2EyPBQDFNqFR5NFDtRLu2EVXR33LhxrFq1qsA0plq1anH16lW2bt3Kxo0bWb16tcntHz58mA0bNrB4sSjFoRUnNobY2FiaN29O7dq1i01pyospJbbv3bvHypUrmThxotHtK9guEpCdbdx28+ZNWrZsqds+/fTTQtuNjo5m0CBTZxUlgI3mjCqUU2YgJt4gIlM+wzjreIGCU5p2AO9RtHivRP4Up6L4BbHySk7fFFPxyGOKnSiXKHZCwZqQgDGIVNXlmJb+ehfoj4ho6YqIrGxr5LUnc/41toJTFPAQmGxC/xRsFsVOlE87YRW1H+rXr09QUBBff/01Q4cONTjm5+fH5cuXDcpv16pVi+DgYKPazsjIIDw8nA8//JDq1UUpjp49ezJhwgSOHz9O8+bNC7xOkiQ++ugj5syZw4oVK3jhhRdMeiZJkkhISKBp06ZGnf/pp5/StWtXateubdJ9FGwXY8MN3d3dOXr0aLHnZWRksG3bNubNm2dhz+ShvA2WCjbKEQzLaM8CAoy4LgYRxXJHb18HYCNFR81ouYZYGQUx4W9UzPkL9T6/YGQfFWye8hyWbgyKnVCwGlYjoiKXI0pXG8sxhNj8JcQ4PgnTnDumVHBKQThoQjBNUFjBplHsRPnDKiJooPCollq1anHp0iUgt/x2WFgYp0+fNqrdefPm4e/vz8CBA3X77O3tGTt2bKHixLdv3+aFF15g7dq1HDp0yGTnDMA///yDq6sr7u7Fx6BnZGTw4YcfMmXKFJPvo2C7yJ0zun37dpo3b26RSLac2KLHW6GckYEQe9SuHLVAVNEo7prXgT4YOmemAbsxzjkDhulNDRGRO4XxB/CD3vepRt5DweZRtAUUO6FQClxHRC12AEYbeY2EcJY8BWQioiAnY5pzBoStqA54GXHuFwjHv2IjFPRQ7ET5sxNW46Dp3r07N27c4MiRIwb7PTw8uHfvHmlpaUBu+e0ePXpw8+bNItv8448/iIqKYtmyZfkEZkeOHMnmzZtJSkoy2H/kyBFatGiBr68v+/fvx9/f36znMSW9aePGjTRq1IhmzZqZdS8F20MysiSeKWXxNm7caB3pTYh5iy0OqArljPfJDR+3B1ZRdFzpReBpDCt3VEekNM0u5tq8mJLe9IHe52eAVibcR8FmKQk7YU0odkLBaogA0jA+/fUuImrmFYRI/AmEo8YcEjCuyl8WQnT4KYTOjYICip0or3bCahw0dnZ2jBs3Ll9Ui1qtxtfX16DK0/DhwwkLCyM0NLTQ8tsajYbw8HBmz56Nj49PvuPu7u6EhITw+eefAyIlaenSpfTo0YOFCxfy0Ucf4eRkbD27/BgrECxJEgsXLlSiZxTyIUmSUZsxpKSk8NNPP5kVDVZSyDmg7tixgwYNGhAQEMD8+fPzHb906RKdO3fmiSeeoGPHjly5ckV3zM7OTleCPCQkRLf/woULtG7dmoCAAAYOHEhGRobFz6xgRZxC6MRoeZOic/y3As2AWL19TyO0BLqacX9jBYKvAuv1visrowp6yGknrBFbnHgrlDO+RVTpmwnUN+L844gqTVsQwu7bAHNrf2iA3zFOf+ZbxCKC8jqhkAfFTpQ/O2E1DhqA8PBwtm7dyq1btwz2a3Vo9Hnvvffw9fVl+PDhBZbfXrp0Kc7OzowaNSrfMS3jx49n+fLlJCUl0a9fP1avXs3Bgwfp27evxc9irINm+/bt2Nvb89xzhdVhVXhU0WiM24yhQoUKJCUlUbly5ZLttLGo5FNd12g0jB8/nu3bt3Pq1Ck2btzIqVOnDM6ZPHkyL774IgkJCcyYMYO33npLd8zFxUVXhnzbtm26/W+88Qavv/46586do0qVKqxatUq2x1coYzSI1KbMnO8NEVWXCiIDEdoeimFK01uI0tymlOLWx9gS20vI7Wdj4Hkz76dgk8hpJ6wOGe2EgoJZ3AbGIxwuk4o5VwKWIcR/M4B9CGeJJb+f5xCRO8ZW+auP0J9RUNBDsRPlz05YVXerVatGaGhovhchfR0aLWq1mi+++IJLly7x3//+1+DY+fPnmT17Np999hlqdeGP+OSTT+Li4kLjxo2pWbMmBw4coG7durI8S3x8vFEOGm30TN4ULIVHG2PzRcuxw1s2j3dsbCwBAQH4+/vj6OhIWFgYW7duNTjn1KlTdOrUCYDg4OB8x/P1TZLYs2cP/fr1A2DYsGHExMSY96AK1sfHwKGczypgJVBQwOQlhOaAvohwNWA7MBfzZfbTAX0ZtcJMxV1ghd73KZiuX6Bgsyh2onyujCqUIyYBtyg+/fUuMBDhzHkWkdIkR5qR1pFf3OvEXkTkzmSs7M1OoaxR7ET5tBNW92ccERHBsmXL0Oi58vz8/PI5aECsfG/dupUNGzboym9LksSYMWOYOnUq9erVK/Q+2pSmy5cv4+XlxdKlSy1KadInNTWVS5cu0bBh0RLqR44c4fz58wwYMECW+yrYFkrOqNiKKyN+9epVfH1zSyr4+Phw9epVg3MCAwPZvHkzAFu2bOH+/fs6/amHDx/SsmVL2rRpo3PCJCUl8dhjj2Fvb19omwrllIuI6BctERQ8kf4OkdJ0WG9fe0RKUzcL+3AKEcUDUJXChYU/Be7nfPYGrENCSsGKUOxE+Zt4K5QTfkKI7k4Fgoo47wRCYH4zQtfsO4Q2mRwkIKr8PV7MeQsAT2BoMecpPJIodqL82QmrKLOtT4sWLahZsybff/89vXv3BkQEzc8//1zg+XnLb1+8eJHk5GQmTpxY6D3u3r1LeHg458+f59ChQwQHB3P27NkiHTqmcOrUKRo0aICDg0OR5y1cuJDXX3+92PMUHk3KszfbGIwdLI0tI14UH3zwAREREaxevZoOHTrg7e2NnZ0dIPRpvL29OX/+PJ06daJp06bWkwqmIC8SogJHas53P0QkjD6ZCAfOojz730Ro1shhNfOmNxX0t5COoRjxa4CjDPdWsCkUO6GgUAKkIGxFfWBGIedIiAjH1wB34GeEE19O4oEGFF3lLwHYCcwp5jyFRxbFTpQ/rC6CBoQ2jL5YcGERNFq05bcHDBjApEmT+Pzzz3Ur33k5duyYrtTwgQMHaNy4MeHh4Sxbtky2/huT3nTu3Dn27t3LyJEjZbuvgm0hZUtGbeUSI73dxgy63t7e/PPPP7rvV65cwdvbUBikZs2abN68mRMnTjBnzhwAHnvsMd31AP7+/nTs2JETJ05QrVo17ty5Q1ZWVqFtKpRD1iBWRbV8AlTU+34ZkdKk75yphihxPQ/5ljSMEQjeAFzL+eyG8aVdFR4pFDthm5NzhTJmOiLaciUFOz3uAWHAOKATIrJSbucM5FZwKoqFQAXg5RK4v4JNoNiJ8mcnrNJB079/f+Lj4zl9WiTp16pVK59IcF46deqEn58fIF7G8iJJEh9//DHdunVj3rx5fPzxxzg7i1F37NixrFmzhpSUFFn6n5CQUGyJ7cjISEaPHk3FihWLPE/h0USSQJNt3FZekUvUq1WrVpw9e5YLFy6QkZFBdHS0QTUmgFu3bunExOfNm8eIESMAuH37Nunp6bpz9u/fz+OPP45KpSI4OJhvvvkGgC+//FIX0adQTkkEXtf7/iKGqUrfI8LYD+ntewoRvt5d5r4UV2I7G8PS2mMRThoFBT1Kwk5oNBqaNWtGz549S67jJmCL4o8KVs4hhDj7OESlvrzEIVKavkU47n9AvpQmfe4idNCKep24DEQDo4AqJdAHhXKP3HbC2mwE2KadsMruOjk5MWrUKF1Ui1b/QVOExPS3335LSkoKo0aNyld+++7duwwcOJCVK1dy4MCBfJovtWrV4umnn2b9+vV5mzWL4io43bx5k+joaF555RVZ7qdgm2RnS0Zt5RE5c0bt7e2Jioqia9euNGrUiAEDBtC4cWNmzJihq8r0888/06BBA+rXr8+NGzeYNm0aAH/++SctW7YkMDCQ4OBg3nzzTR5/XCR7v//++0RGRhIQEEBSUhLh4eEl9eNQKA1eQVTkAPAAInM+ZyLEd3vpHQehO/Az4Iu8SBTvoPkfQqcGwAGYIHMfFGwGue3EkiVLaNSoUQn22HhsVVtAwYpJR1T480E4X/SREFGXbRBpsnsRqa8l9SZljEDwhzn9er2IcxQeeeS0E9ZkI8B27YTVadBoGTNmDIGBgcyZM4dKlSpRtWpVEhMTC0wzuH37Nq+88gpff/01bdu2ZfDgwQwfPpz169cTFxfHgAED6NKlC2vWrNFFzeQlIiKCiRMnMmrUKIsqKkmSVKyDJioqiv79++Pl5WX2fRRsGwklZ9QUunfvTvfuhmEOs2bN0n3u16+friKTPk899RQnT54ssE1/f39iY2Pl66RC2bEF+Ebv+1JE6tJlRJj6Qb1jVRGpUD1KqC83EFVBQEzsGxdwzkK9z0MoXERY4ZFGbjtx5coVfvjhB6ZNm0ZkZGTxF5QC5W1SrVDOmYdwjv+AYdTiPWAMIlqlK7AWoTtTkhTnoLmNEJIPQ+ipKSgUgJx2whptBNimnbDKCBoAX19fgoODWbduHVBwqW0tkyZNom/fvrRr105XfvvixYv06NGDrl27Mnv2bJYtW1aocwagc+fOZGRk8Ouvv1rU76tXr+Lg4ICnp2eBx1NSUli+fDmTJk2y6D4KNo6tl8Wz0ZxRBSvkNiJUXUtvoD9iAt4MQ+dMW0T4ekk5Z8AweqY+4JLn+GHgF73vk0uwLwrlGxPsRHHV8ABee+01FixYgFptJVNDxU4olCYnEaLx/8EwrTUeaAl8lXP8f5S8cwaEg6YKooJfQaxAiBlPKYW+KJRfZLQTVmcjwGbthNVG0ICIaomIiGDs2LH4+flx+fJlnnrqKYNzfvrpJ/bs2WOwCp6ZmYmHhwc7duxg1qxZhIWFFXsvlUqlEyfu0KGD2X0uLnrmiy++oH379tSvX9/seyg8Ckhkl1vvi3GUt8FSoZwyBaE/A1AZoS3wJqIsad7z5iBSikqS4gSC9aNnelJ8eVWFRxjj7URx1fC+//57PDw8aNGiRaFVM8sCxU4olAoaYCTCRmir50nAZ8CriMjKvQgR+dIigcKr/D1E2LIuFK1Ro6Agk52wVhsBtmknrMgFlp+OHTsCsG/fvgIjaB48eMDo0aNZsWIFlSpVAuDEiRO0aNGCGjVqcPjwYSIjI43+RRo2bBg//fQTV69eNbvPRTlosrKyiIyMZMoUxd2tUDxStnFbecUWPd4KVsZuYJXe97eBwRg6Z6oA3+XsK2nnDOQvsa3PWWCz3vepJd8dhfKNXHZi//79bNu2jdq1axMWFsaePXsYMmRIyT9AMSh2QqFU+AiIRaS/VgfuIyJpxgDPICIrS9M5k42I6ClsvXcdIl1WsREKRiCHnbBWGwG2aSes2kGjjWr58KN1JD98kl+O+TDj/VN8tfUK9x9k8c477/D000/TrVs3JElixYoVdOnShVmzZrFixQqCgoLYuHEjAwcO1FWEKgo3NzcGDRpUYOivsRRVYvvbb7/F29ubtm3bmt2+wqOBJIFGIxm1lUckbFN1XeH/7d15eFNl+v/xd5amC2vZBFIQSqCU0lL4URCdcaCCKDh1gUEEBSkMCEVRQB0Gp14iXEXA+aqA+h11oIxCZwQENyqOqHwVsYK0VStYVttSEJCdrsn5/RETWrqdQprknN6v68plm5yEJ7U9n+TO89yPH7lI5a2pY4DngB0VrrsB5wtvb25GUNsMmr/j/OMAGEjDbNsqdMOTOZGSkkJ+fj6HDx8mLS2N+Ph49xJzX5GcEF5xEJiPMwfuxXmO7g/8G1gIbMHZWN7bY7pI7bv89cO5xbcQtfBUTvhjRoB+c8KvlzgdP1HMnv0D+KXYxqkjFhTFwLYvTrDjm1O8vPoAp46G8N9353Pu3DmmTp3Kjz/+yJdffllp+VB8fDyLFy/mjjvu4KuvvqJt29oXjs6YMYOhQ4cyf/58LBZLvcecnZ3NX/7ylyrXK4rC0qVLSU5OrvdjisZJ0fMSJw1Ws4XGJON8kQvOpMu+4va5OPsJeGPWjEsp8GOF7yu++P4FWFXh+yeofmq7EBVITghxDRScW1SbgZeB13EuaQoFtuGcPeMLrryqbvnSe8A+YB2SEUIVyQlEJtfiAAAdm0lEQVTt8dsCzbFfipn82G7OXyjHaAqs1Ay1uMQ5D6t1h6HMW7SXL9KnMGTIH9i5cyfBwVd2XIRJkyaRm5vLXXfdxSeffFJrs+CoqCgiIyPZuHGjqt41FRUXF3Pw4EF69uxZ5bZPP/2Uixcv+tW+8cJ/KYBGd9BWTY8nVOFb+/afJ/fgBZocNXPD8lYEY3LeUF7hoFAgFee22t62F+e23gAtqbyF9wqcW7wC2HA2MxaiFg2VE4MHD3YvMfc1yQnhSYqikJ1zlp/ziwhtGcDA71oRsM3o7OcyD3gLGIZzCZG3Z81UlIVzjUN1PciWAF2AqhtTClFFQ+SEP2UE6DMn/LJAoygKTyz4jvMXynHUsiau3G5k7/6LjBi1hBdSEmp9zIULF3Lfffe5t9+urQP1zJkz+fvf/17vAk1OTg42m43AwMAqty1dupS5c+f6V+dr4b8UUHReoVF7QtXheVd42C8nS3hiwXfkHy0CwFhkwHGLQlJON+7+ucIe1QNxTlu/3ifDrLq8yfXLfRFYWeG2ueCqLQlRI8kJN8kJUZeDRy7y5LPfc+ZsGYqiYDQYMFyA+Tf15OaVbWA/8CzOQo2vz7/ZQHcg5Irrv8S5TPcl/PQdnPA7khNuWsoJv6wW/Jh7nqPHimstzrgYTUHsPdSa8vLaDzYajaxevZrDhw/z9NNP13psQkICR44cITMzsz7DJjs7mz59qs5HzM7OJisry2+aKQlt0PM22wr6bOolvK+szMH0J/Zw6MhFikscFJc4uGS0U2x2sKLXAT5rf8J54Gyc21f7qjgDlbfYrri86Z/Ar7993RaY4LURCY2TnJCcEHU7e66MpL9kcuyXYoqK7c6cKLZz0WznmeY/8r39rLOp/FP4vjgDl3dwutJSnDtKJXp3OELbJCe0lxN+WaBZ/24BJaXqt6dxOBR27v61zuOCg4PZvHkza9euJTU1tcbjzGYzDz30ECtXrqzxmOrUtIPTsmXLeOSRR6qdWSNETRwORdVFq/R4QhXet33nSc6dL8deTWSUmB28GnUQ5R0Fngfq31bMs6prEFyOszmwyyNA1ZW6QlRLckJyQtTtva2FlJY6qn0TWmJ08Po9h2Gwt0dVg3M4+6dd+XZiL/AukAQ08faghJZJTmgvJ/xygtzhvEv1quSVljnILyxSdWy7du14//33GTx4MNdff32Na+j+/Oc/ExERwZIlSwgNDVX12FlZWdx6662VrsvLy+ODDz7gpZdeUvUYQoBzmZ9Dozs0qWJQ31HdHz7MEv7ri52nKCq213j7saYlnIsvp4VXuwHXoLotttcDh3/7OgSY4c0BCS2TnLhMckLUZvtXJ2v+4NcAmblnvTug2nz/23+vLNA8DwQCM707HKFtkhOXaSkn/HIGTX0ZAEM9VpZFRkbWuf12u3btGDlyJKtWrar29ispikJWVlaVJU4vvPACkyZNomXLlqrHJwSAQ1FUXbRKjxVv4X2GulLMAAZ/+D36BTj229cGIArn3NwlFY6ZgnP6uhAqSU6oz4n09HQiIiKw2WwsXry4yu0lJSXce++92Gw2Bg4cyOHDh923paSkYLPZiIiI4KOPPnJff+bMGUaPHk3Pnj2JjIzkq6++utanLBpAXRngDxHh5irkVyzQFAJrgAfxbfNioUmSE9p7P+GXBRpb1ybUp5duQICRzmH1mxMeHx9PSkoKd9xxBydOnKj2mJkzZ7Jy5UocKprhHDt2DIPBQPv27d3XnT59mlWrVjFr1qx6jU0IcBb91Fy0SK9rRoX3/eHGtgQH1fy5SFiHYJo387PZMzacU9S3AXt+u84EPObtQQmtk5xQlxN2u52kpCS2bNlCTk4O69atIycnp9Ixb7zxBqGhoezfv5/HHnuMJ598EnBuAJGWlsYPP/xAeno6M2bMwG53ztqbNWsWt912G3v37iUrK4vIyEhP/xiEB8T/vh2BgdW/sTAYoH+supnyXpENtAA6V7huOc4dAOf4ZERC4yQntPd+wi8LNH9KCCMgQP3QysqKiIqof3+XxMRExowZw1133UVxcXGV2wcOHEhoaCjp6el1Ppar/4yhQpn+1Vdf5Y9//COdOnWq5Z5CVKUosmZUiydU4X03DWhNm9YWzOaqvyyBFiNJieE+GFU1qlveVHH2zBicW6cKoZLkhPqcyMjIwGazER4ejsViYezYsWzevLnSMZs3b2bixIkAjB49mk8++QRFUdi8eTNjx44lMDCQrl27YrPZyMjI4OzZs2zfvp3JkycDYLFYZLa0nxo5tD3BQaZqZ9JYLEamjO/i9THVKIvKu/ydB14G7sFZ3BeiHiQntPl+wi8LNN3Dm9K1UwgmU90/TYvFQIhxN5E9I3j99dfdn2qotXDhQsLCwpg0aVKVmTIGg4GkpCRVzYKzsrIqNQguLi5m+fLlzJ07t17jEcJFz13XUXky1doJVXif2WTg5edi6d2zORaLkSYhJkKCTTRtYuLJh3swqH9rXw/R6coGwZnA1grXPe7d4Qh9kJxQlxMFBQWVPiwLCwujoKCgxmPMZjMtWrTg1KlTNd730KFDtG3blkmTJtG3b1+mTJnCxYsXPfPchUc1bWLmf5f1pVuXJgQFGgkJduZEq5YBLH6qNxG2Zr4eopMD+I7Ky5teB84iGSGumuSE9t5P+GWBBuC55GhaNg/AXEtHn6BAIwP7teK9t5/knXfeYc2aNfTp04cPP/xQ9VSturbfHjt2LBkZGezfv7/Wx7lyi+0333yT2NhYoqOjVY1DiIoURcFud6i6aJXDqO4iRF1CW1hYkRLLmuX9mfdIBIv+GsX7b97IrYOv8/XQLrtyi+1lFb4fCvT17nCE9klOXL6cOHGC/v37uy//+Mc/Gnxs5eXlfPvtt0yfPp09e/bQpEmTanvbCP9gbR/M6pf687/L+jJvVgRLknuzKXUQcf60vOkIzhkzrrcTZcD/ADcDA301KKFlkhPafD/ht8NtHWph9Uv/j7i+rQgIMBAQcLn0FRxkIjjIyJg7w1j4lyiMRgMDBgzg888/Z9GiRcyePZuhQ4fy7bffqvq3XNtvv/XWW1W23w4ODiYxMZFXXnml1seouMW2w+Fg2bJlPP64lLvF1VMciqqLGv7WyFCva0aFb4V1DGbwTW2Jiw3FbPajeCsDKra7aA2kVfj+Ce8OR+iHJ3PC39QnJ9q2bcuuXbvcl6lTp1Z6LKvVSl5envv7/Px8rFZrjceUl5dz9uxZWrduXeN9w8LCCAsLY+BA5zvn0aNHq37dKXynW5emDLmpLbG9W2I0+tmLDFch3zWD5t9AHpIR4ppITmjv/YRfbrPtEtrSwtKnozlxqoStnx6n4FgRAQFGenZvRvxNbQkMrDy9xmAwcOeddzJixAhef/11Ro4cydChQ1m0aBGdO3eu4V9xateuHR988EG1228/9NBDxMXF8eyzzxISElLlviUlJeTm5tKrVy8A3nvvPZo1a1bjFt5CqOHJk6WrkeH69espLS3l0qVLHnvsq6W1k6UQV+0noPS3r5vh3FrbtRo3FucMGiGuglZfVKvlqZyIi4sjNzeXQ4cOYbVaSUtLY+3atZWOSUhIIDU1lUGDBrF+/Xri4+MxGAwkJCQwbtw4Zs+ezdGjR8nNzWXAgAGYTCY6derEvn37iIiI4JNPPnG/DhTiqmTj7D3TG9y7/EUBt/tyUELrJCe0x68LNC5tWwcyfnTtBZaKAgICmD59Ovfffz9Llixxrw2eN29erQ3cKm6/vX37diIiIgDo2rUrN954I2vXrmXKlClV7rd3717Cw8MJCgoCYOnSpTz++OOVGgYLUS8KeOp86mpkuHr1asDZyNBisXjmwa+WBqvZQly1isubeuHsKeDyOH62x6vQDA/mhF/yYE6YzWZWrFjB8OHDsdvtJCYmEhUVRXJyMv379ychIYHJkyfzwAMPYLPZaNWqFWlpzmluUVFRjBkzhl69emE2m1m5ciUmk/MDwuXLlzN+/HhKS0sJDw9n1apVnhmwaJyyubzL30c4+9Gswo/XOwi/JzmhSbr+k2/WrBnPPvss2dnZnDx5koiICF588UVKS0trvI9r++2RI0dW2n575syZrFixotreNhWXN+3YsYPCwkLuuecezz8h0WgoqJ+SWNfae39tZOjJKYnp6elERERgs9mq7QFw5MgRbrnlFmJiYhg8eDD5+fkAZGZmMmjQIKKiooiJieHf//63+z4PPvggXbt2JTY2ltjYWDIzMz3yvEUjVLFBsAK4/vw6A3/y/nCEPtQnJ7TKkzkxYsQIfvrpJw4cOMD8+fMBWLBgAQkJCQAEBQXx9ttvs3//fjIyMggPv7wD3Pz58zlw4AD79u3j9tsvT2eIjY1l165dZGdns2nTJkJD/aifidAe1w5O4Jw90xEY57vhCO2TnNDmEiddF2hcrFYrb7zxBv/9739JT0+nV69erF+/vsZGwtVtvz106FAuXbrEjh07qhxfcQenpUuXMmfOHMxmTUxOEn5LQVHUXepae++vjQw9dUK12+0kJSWxZcsWcnJyWLduHTk5OZWOmTt3LhMmTCA7O5vk5GTmzZsHQEhICGvWrOGHH34gPT2dRx99lDNnzrjvt3TpUjIzM8nMzCQ2Ntajz180IhULND9W+Ho2EODlsQgdUZ8TWqXHF95CVOsCcABngWY3sA14FPDxhGehdZITWsyJRlGgcYmOjmbLli288sorLFy4kJtuuqnaggs4t9+2Wq0kJiaiKApGo5GkpCRWrFhR5VjXDk779u1jx44dPPjggw38TITuKXis67o/NjJU8FzX9YyMDGw2G+Hh4VgsFsaOHcvmzZsrHZOTk0N8fDwAQ4YMcd/eo0cPunfvDkDHjh1p165dpZlzQnhExSVO53/7bygw2QdjEfrhwZzwR57MCSH83g84f+ljgKVAc2BqrfcQom6SE6pzoq7Z+Nu3b6dfv36YzWbWr1/v+SdTQaOMtWHDhrF7926mTZvGvffey6hRo8jNza10jNFoJDU1lUOHDrm33544cSLp6ekUFhZWOta1xOn5559nxowZ1TYSFqI+PDklsX379u5GhoB/NDJUWe1WU/EuKCigU6dO7u/DwsIoKCiodEyfPn3YuHEjAO+88w7nz5/n1KlTlY7JyMigtLSUbt26ua+bP38+MTExPPbYY5SUlFzDExaN1kngaDXXzwCaenksQlc8mRPFxcUMGDCAPn36EBUV5X7d41MezAkh/J6rkB8KvA1MA1r4bjhCHyQn1OWEmtn4nTt3ZvXq1Ywb1/DrDhtlgQbAZDIxceJEfvrpJ/r378+gQYN4+OGHK3167tp++8033yQ1NZWWLVty77338tprr7mPOX78OKWlpZhMJtavX09SUpIvno7QG8Wza0ZdjQxjYmLIzMzkr3/9awM/gbqpPaHW1WNHjWXLlvH555/Tt29fPv/8c6xWq7vJI0BhYSEPPPAAq1atwmh0nhZTUlLYu3cv33zzDb/++ivPPfecx567aES+q+a6QOBhbw9E6I4HcyIwMJBt27aRlZVFZmYm6enp7Ny50wtPonZSoBGNRjbOXf7eBkzALN8OR+iE5ISqnFAzG79Lly7ExMS43yc0pEbfKCU4OJh58+YxZcoUFixYQGRkJHPmzOHRRx8lODi4yvbbSUlJjPzjOK7rMp7cg5c4ceIkUXFPsGTZPxk3bhxt2rTx9VMSuqDg8OB6UFcjQ3+i9kW1q8dOTaxWK3l5ee7v8/PzsVqtlY7p2LGjewbNhQsX2LBhg3tHt3PnzjFy5EgWLVrEDTfc4L5Phw4dAGcgTZo0iWXLlqkbsBAVZVdz3UTgOm8PROiP53LCYDDQtKlzSldZWRllZWV+sROlFF9Eo5EN9AT+CYwHrLUfLoQ6khMurg98XaZOneru21ndbPyvv/7ao+Osj0Y7g+ZKbdu2Zfny5ezYsYNdu3YRERFBamoqDofDvf322Psm8I+1F+nc+39YuyGPnbt/5cDPRhTLQL7+MY4LhvGcPCXLIIRn6LnruoLnPhmNi4sjNzeXQ4cOUVpaSlpamntXDpeTJ0/icDjX16akpJCYmAhAaWkpd999NxMmTGD06NGV7uNayqgoCps2baJ3797X/sRF45N1xfcGYI4vBiL0yJM5YbfbiY2NpV27dgwbNszdt8xXPJkTQvg1hcvF/CJgrg/HInRHcsJ5qWtTFX8iBZor9OjRgw0bNrBu3TpeffVV+vXrx8cff8ygG/9AXPzr7M4+D4YA7I6KrwiMGE2BHMorY9Ks3VKkEddMAem6rvKFt9lsZsWKFQwfPpzIyEjGjBlDVFQUycnJvPvuuwB89tlnRERE0KNHD44fP+7eYvU///kP27dvZ/Xq1VW20x4/fjzR0dFER0dz8uRJnnrqqQb7WQh9+jn/Ehv3F7Cp81EKg507AnIX0MOnwxI6UZ+cULNU1GQykZmZSX5+PhkZGXz//ffef1JXkAKN0LufDpxnw78KeL95Ib/mlsJIIMrXoxJ6ITmhLifUzMb3pka/xKkmrh2eNmzYwPTp0wnr8TCGwFhqW3Zmtyucu1DOU4tzeHVpX+8NVuiPAvZybXZUV8Xg2Z03RowYwYgRIypdt2DBAvfXo0ePrjJDBuD+++/n/vvvr/Yxt23b5rkBikalqNhO8nM57M4+g6Ep0AteMhzg5mOtmT+nJwHy2YjwhHrkRF1LRStq2bIlQ4YMIT093bczBz2cE0L4k9NnS3lywfccPHIRRzkYo+DvhlzujrMyUwn3i6UjQgckJ1SpOBvfarWSlpbG2rVrr/2Br5JEXy0MBgOjR4/mm13ZKAExlJXXfR+7XSH34AUO511s+AEKHVNX7ZYZNEL4n6eX5LA76zSlpQ5KTA5KzA5KTQ7+r+Mpnt+dW/cDCKGK53LixIkTnDlzBoCioiI+/vhjevbs2dBPoE6SE0KPFEXh0aey+engBYpLHJTaHRSbHZSaFDZ/d5R/vZ1X94MIoYrkhJqcUDMb/5tvviEsLIy3336badOmERXVcFPdZAaNCl9mnCXAYsZerK4CWV7uYNOWozw6tXsDj0zolaKA4tDvDBrXmlEh9KagsIhdmWcoLav6YqfE4GDrZ78wbWJXQltYfDA6oSeezInCwkImTpyI3W7H4XAwZswY7rjjDo889tWSnBB6lfn9WY4eK6K8vGpOFJc4eGvDz9x3dxgBAfI5urg2khPq1TUbPy4ujvz8fM/8Y3WQAo0K+UeLKFZZnAGwO+Dwz5cacESiMXBotAGwWvLCW+jRnu/PYDQBZdXfHhBg4Ie95/jdQNnxT1w7T+VETEwMe/bs8chjeZLkhNCj3dmnKarlfYWiwOG8S3QPb+rFUQm9kpzQHinQqGA01v//vNGkw98W4VVaXr5UJ5mWLnTKZDRgoPZfbpPkg/AQyQkhtMdkNGA0Qk0TGxRFckJ4juSE9kiBRoWu1zchOMhYa7W7ooAAAxHdmjXwqISeKYqCQ89NgpHmj0Kf4vqGYrfX/LdbXq7Qp1cLL45I6JXkhBDadGNca95cn0dJafV/v0GBRrp0CvHyqIQeSU5okw6fkufdPKgN1KObugG46/YODTcg0Sg4FIeqixa51oxK80ehN21aBXL7Le0JCqwar0GBRu67J4yQEPlsRHiG5ITkhNCeCFsz+kS1wGKpPiceejD8qmbvC1EdyQnt5YQUaFQItBj50x+t1b7gvpIlwEBc31Zc1zbICyMTuqWA4lBUXbRKjydUIQBmP9SdhNs6YLEYaRJiokmIiaBAI+NGdWLyuC6+Hp7QC8kJyQmhWSnzoxh8YxssAQZ3TjQJMTFzSjdGDm3v6+EJvZCc0GROyMd4KiWO68Le/efJ/uEsxSXVVxktAQY6tA8meY7vtxwT2qag7ZNlnTR4shRCLZPJwCNTbEwe14Xv957DaDQQ3bM5QUEmXw9N6IjkhBDaFRhoInlOJA9P6cbe3PMEBpqI7tlcdm4SHiU5oU1SoFHJbDKwJDmaN946xPr3jgJQVGwHnNMRHQoM/X1bHp3WnZBgeREurp2um3qhzxOqEBU1CTEzsF8rXw9D6JjkhBDaFtrCwqD+rX09DKFjkhPaIwWaejCbDEybEM6DY7vw+Y4THPz5Ig67grVDMLf8vh1Nm8iPU3iIAo6a2vvrhB5PqEII4TWSE0IIIWojOaFJUlG4CoEWI7cOvs7XwxA6pqDgsNt9PYwGoxj02XVdCCG8RXJCCCFEbSQntEkKNEL4o9+aeumZHiveQgjhNZITQgghaiM5oUlSoBHCT8kJVQghRG0kJ4QQQtRGckJ7pEAjhF9ScCiyZlQIIURNJCeEEELURnJCi6RAI4QfUnQ+JVFBnydUIYTwFskJIYQQtZGc0CYp0AjhpxQ9d1036POEKoQQ3iQ5IYQQojaSE9ojBRoh/JGi767roM+u60II4TWSE0IIIWojOaFJUqARwg8pgEPHUxJBnxVvIYTwFskJIYQQtZGc0CYp0AjhjxR9T0nU65pRIYTwGskJIYQQtZGc0CQdTgoSQg8UFIe6ixpdunQhOjqa2NhY+vfv38BjV+G3NaNqLmqkp6cTERGBzWZj8eLFVW4/cuQIt9xyCzExMQwePJj8/Hz3bampqXTv3p3u3buTmprqvn737t1ER0djs9l45JFHUBR9fwIhhNAaz+VEXl4eQ4YMoVevXkRFRfHiiy96Yfx18HBOCCFE4yM5ocWckAKNEH5KURyqLmp9+umnZGZmsmvXrgYctXqeOqHa7XaSkpLYsmULOTk5rFu3jpycnErHzJ07lwkTJpCdnU1ycjLz5s0D4Ndff+WZZ57h66+/JiMjg2eeeYbTp08DMH36dF577TVyc3PJzc0lPT3d4z8DIYS4Fp7KCbPZzPPPP09OTg47d+5k5cqVVc6jvqDHF95CCOFNkhPaywkp0AjhhxRFwVFuV3XRKodR3aUuGRkZ2Gw2wsPDsVgsjB07ls2bN1c6Jicnh/j4eACGDBnivv2jjz5i2LBhtGrVitDQUIYNG0Z6ejqFhYWcO3eOG264AYPBwIQJE9i0aZPHfwZCCHG1PJkTHTp0oF+/fgA0a9aMyMhICgoKGvop1MlTOSGEEI2R5IQ2c0J60Ajhh353Q1dOFs5RdWxRUVGlZUtTp05l6tSplY4xGAzceuutGAwGpk2bVuV2bxvepg0nVS61quv5FRQU0KlTJ/f3YWFhfP3115Ueo0+fPmzcuJFZs2bxzjvvcP78eU6dOlXtfQsKCigoKCAsLKzK9UII4S88nRMuhw8fZs+ePQwcONAj47xa9cmJNm3aNPBohBBCeyQnLtNSTkiBRgg/5OnlNF988QVWq5VffvmFYcOG0bNnT26++WaP/hv14e3lQsuWLWPmzJmsXr2am2++GavVislk8uoYhBDCkxriPHrhwgVGjRrFCy+8QPPmzT3++PUhy0qFEOLaSE5ok8Ym/AghrobVagWgXbt23H333WRkZPh4RJ5jtVrJy8tzf5+fn+9+vi4dO3Zk48aN7Nmzh0WLFgHQsmXLGu9rtVorNRKu7jGFEEJPysrKGDVqFOPHj+eee+7x9XCEEEL4GckJ75ACjRA6d/HiRc6fP+/+euvWrfTu3dvHo/KcuLg4cnNzOXToEKWlpaSlpZGQkFDpmJMnT+L4bZvBlJQUEhMTARg+fDhbt27l9OnTnD59mq1btzJ8+HA6dOhA8+bN2blzJ4qisGbNGu68806vPzchhPAGRVGYPHkykZGRzJ4929fDEUII4WckJ7xHCjRC6Nzx48f53e9+R58+fRgwYAAjR47ktttu8/WwPMZsNrNixQqGDx9OZGQkY8aMISoqiuTkZN59910APvvsMyIiIujRowfHjx9n/vz5ALRq1Yq//e1vxMXFERcXR3JyMq1atQLg5ZdfZsqUKdhsNrp168btt9/us+cohBAN6csvv+Rf//oX27ZtIzY2ltjYWD788ENfD0sIIYSfkJzwHoOiKHVvfC6EEEIIIYQQQgghGozMoBFCCCGEEEIIIYTwMSnQCCGEEEIIIYQQQviYFGiEEEIIIYQQQgghfEwKNEIIIYQQQgghhBA+JgUaIYQQQgghhBBCCB+TAo0QQgghhBBCCCGEj0mBRgghhBBCCCGEEMLH/j9DP77tuT+0IQAAAABJRU5ErkJggg==",
            "text/plain": [
              "<Figure size 1440x288 with 9 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "fig = plt.figure(figsize=(20,4), constrained_layout=True)\n",
        "\n",
        "ax1 = plt.subplot2grid((1, 3), (0, 0))\n",
        "ax2 = plt.subplot2grid((1, 3), (0, 1))\n",
        "ax3 = plt.subplot2grid((1, 3), (0, 2))\n",
        "\n",
        "node_cmap = plt.cm.coolwarm\n",
        "edge_cmap = plt.cm.cool\n",
        "\n",
        "####### ORIGINAL\n",
        "degree = np.array(list(dict(G.degree).values()))\n",
        "pos = nx.kamada_kawai_layout(G)\n",
        "#pos = nx.spring_layout(G)\n",
        "nx.draw(G, node_color = degree, cmap=node_cmap, pos=pos, node_size=degree*30, ax=ax1)\n",
        "\n",
        "plt.colorbar(get_cbar(edge_cmap, [1]*G.number_of_nodes()), label='Edge weight', orientation=\"vertical\", ax=ax1)\n",
        "plt.colorbar(get_cbar(node_cmap, degree), label='Degree', orientation=\"vertical\", ax=ax1)\n",
        "ax1.set_title('Original')\n",
        "\n",
        "######### REWIRED\n",
        "H = nx.from_numpy_matrix(np.asmatrix(rewired_adj))\n",
        "#totally connected lead to totally disconnected\n",
        "if len(torch.nonzero(rewired_adj)) > 0:\n",
        "  edges, weights = zip(*nx.get_edge_attributes(H,'weight').items())\n",
        "  new_degree = np.array(list(dict(H.degree).values()))\n",
        "else:\n",
        "  edges = []\n",
        "  weights= np.array([0])\n",
        "  new_degree = np.array([1]*H.number_of_nodes())\n",
        "\n",
        "#posH = nx.kamada_kawai_layout(H)\n",
        "\n",
        "#draw\n",
        "nx.draw(H, node_color = new_degree, cmap=node_cmap, pos=pos, node_size=new_degree*30, \n",
        "        edgelist=edges, edge_color=weights, width=np.array(weights)*200, edge_cmap=edge_cmap, ax=ax2)\n",
        "#colorbar\n",
        "fig.colorbar(get_cbar(edge_cmap, weights), label='Edge weight', orientation=\"vertical\", ax=ax2)\n",
        "fig.colorbar(get_cbar(node_cmap, new_degree), label='New Degree', orientation=\"vertical\", ax=ax2)\n",
        "ax2.set_title('CT-Distance')\n",
        "\n",
        "\n",
        "######### REWIRED and row-stochastic\n",
        "new_matrix = normalize(rewired_adj,axis=1, norm='l1')\n",
        "H = nx.from_numpy_matrix(np.asmatrix(new_matrix))\n",
        "#totally connected lead to totally disconnected\n",
        "\n",
        "if len(torch.nonzero(rewired_adj)) > 0:\n",
        "  edges, weights = zip(*nx.get_edge_attributes(H,'weight').items())\n",
        "  new_degree = np.array(list(dict(H.degree).values()))\n",
        "else:\n",
        "  edges = []\n",
        "  weights= np.array([0])\n",
        "  new_degree = np.array([1]*H.number_of_nodes())\n",
        "\n",
        "#posH = nx.kamada_kawai_layout(H)\n",
        "\n",
        "#draw\n",
        "nx.draw(H, node_color = new_degree, cmap=node_cmap, pos=pos, node_size=new_degree*30, \n",
        "        edgelist=edges, edge_color=weights, width=np.array(weights)*3, edge_cmap=edge_cmap, ax=ax3)\n",
        "#colorbar\n",
        "fig.colorbar(get_cbar(edge_cmap, weights), label='Edge weight', orientation=\"vertical\", ax=ax3)\n",
        "fig.colorbar(get_cbar(node_cmap, new_degree), label='New Degree', orientation=\"vertical\", ax=ax3)\n",
        "ax3.set_title('ROW Stochastic CT-Distance')\n",
        "#plt.tight_layout()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "PovEwYLkwBJv",
      "metadata": {
        "id": "PovEwYLkwBJv"
      },
      "source": [
        "#### CT Embedding"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 304,
      "id": "n--uDZMZwA1H",
      "metadata": {
        "id": "n--uDZMZwA1H"
      },
      "outputs": [],
      "source": [
        "def visualize(z, color, size=10, save=False, title='img', ax=None):\n",
        "    plt.figure(figsize=(5,5))\n",
        "    plt.scatter(z[:, 0], z[:, 1], s=size, c=color, cmap=\"seismic\")\n",
        "    plt.colorbar()\n",
        "    plt.xticks([], [])\n",
        "    plt.yticks([], [])\n",
        "    if save:\n",
        "        plt.savefig(title+'.pdf')\n",
        "     \n",
        "    plt.show()\n",
        "\n",
        "def graph_vol(G):\n",
        "    A = nx.adjacency_matrix(G)\n",
        "    D = A.sum(axis=1)\n",
        "    D = D.squeeze()\n",
        "    d = np.zeros(G.number_of_nodes())\n",
        "    for i in range(G.number_of_nodes()):\n",
        "        d[i] = D[0,i]\n",
        "    vol = d.sum()\n",
        "    return vol\n",
        "\n",
        "def commute_times_embedding(G):\n",
        "\n",
        "    A = nx.adjacency_matrix(G)\n",
        "    n = G.number_of_nodes()\n",
        "    vol = graph_vol(G)\n",
        "    D = A.sum(axis=1)\n",
        "    L = nx.normalized_laplacian_matrix(G)\n",
        "\n",
        "    # Finding eigen values and eigen vectors\n",
        "    e, evecs = np.linalg.eig(L.todense())\n",
        "    # Sort them (both e's and evecs's) ascending\n",
        "    idx =e.argsort()\n",
        "    e = e[idx]\n",
        "    evecs = evecs[:,idx]\n",
        "\n",
        "    #Compute CTE\n",
        "    Lambda = np.diag(e)\n",
        "    Lambda = fractional_matrix_power(Lambda, -0.5)\n",
        "    Lambda[0,0] = 0\n",
        "    CTE = np.sqrt(vol)*np.matmul(Lambda,np.transpose(evecs))\n",
        "    # Embedding is in the columns\n",
        "    return CTE"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "zYYSEjPgDNVD",
      "metadata": {
        "id": "zYYSEjPgDNVD"
      },
      "source": [
        "**Select a graph, calculate Spectral CT, predicted CT by CT-Layer and plot**\n",
        "\n",
        "Ideally we select connected graphs for a fair comparison with Spectral CT (CT computed spectrally has complex numbers if disconnected)\n",
        "\n",
        "You can select the graph yourself, either from the list of connected graphs `connected_graphs[IDX]` or from all the graphs in the dataset `next(islice(iter(indiv_loader),IDX,None))`"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 305,
      "id": "MnqHa-NMAZBj",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MnqHa-NMAZBj",
        "outputId": "15020db8-6e40-4f9a-b7a7-8440e5a10aca"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "(True, 1, 12)"
            ]
          },
          "execution_count": 305,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "G = to_networkx(graph_i, node_attrs=[\"x\"], to_undirected=True)\n",
        "degree = np.array(list(dict(G.degree).values()))\n",
        "nx.is_connected(G), len(list(nx.connected_components(G))), G.number_of_nodes()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 306,
      "id": "mWtWIrBpAZ-T",
      "metadata": {
        "id": "mWtWIrBpAZ-T"
      },
      "outputs": [],
      "source": [
        "#Spectral CTE\n",
        "degree = np.array(list(dict(G.degree).values()))\n",
        "spectral_CTE = commute_times_embedding(G).T"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 307,
      "id": "zUDHsu94DhHi",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zUDHsu94DhHi",
        "outputId": "0427d3cc-3c2e-4b67-8d47-4d6f7be11619"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "correct pred? False\n"
          ]
        }
      ],
      "source": [
        "#Learned CTE\n",
        "graph_i.to(device)\n",
        "pred, mc_loss, o_loss = model(graph_i.x, graph_i.edge_index, graph_i.batch)\n",
        "loss_graph = F.nll_loss(pred,graph_i.y.view(-1)) + mc_loss + o_loss\n",
        "print('correct pred?', bool(pred.max(dim=-1)[1].eq(graph_i.y.view(-1)).sum().item()))\n",
        "learned_cte = model.ct_embedding.detach().cpu().squeeze()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 308,
      "id": "phnN2K2X4wN-",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 322
        },
        "id": "phnN2K2X4wN-",
        "outputId": "0e965ed5-91f6-46a0-ae74-74eeb9df8980"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 1152x360 with 3 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "fig = plt.figure(figsize=(16,5))\n",
        "\n",
        "ax1 = plt.subplot2grid((1, 3), (0, 0))\n",
        "ax2 = plt.subplot2grid((1, 3), (0, 1), rowspan=1)\n",
        "ax3 = plt.subplot2grid((1, 3), (0, 2))\n",
        "\n",
        "#pos = nx.spring_layout(G, seed=44)\n",
        "nx.draw(G, node_color = degree, cmap=\"coolwarm\", pos=pos, node_size=degree*30, ax=ax2)\n",
        "\n",
        "learned_cte_2d = TSNE(n_components=2, learning_rate='auto',\n",
        "               init='random', perplexity=6, early_exaggeration=100, random_state=200).fit_transform(learned_cte.real)\n",
        "ax3.set_xticks([], [])\n",
        "ax3.set_yticks([], [])\n",
        "ax3.set_title('CT-Layer - Learned CTE')\n",
        "ax3.scatter(learned_cte_2d[:, 0], learned_cte_2d[:, 1], s=degree*10, c=degree, cmap=\"coolwarm\")\n",
        "\n",
        "spec_cte_2d = TSNE(n_components=2, learning_rate='auto',\n",
        "               init='random', perplexity=6, early_exaggeration=100, random_state=200).fit_transform(spectral_CTE.real)\n",
        "ax1.set_xticks([], [])\n",
        "ax1.set_yticks([], [])\n",
        "ax1.set_title('Spectral CTE')\n",
        "ax1.scatter(spec_cte_2d[:, 0], spec_cte_2d[:, 1], s=degree*10, c=degree, cmap=\"coolwarm\")\n",
        "\n",
        "plt.show()"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "provenance": [],
      "toc_visible": true
    },
    "gpuClass": "standard",
    "kernelspec": {
      "display_name": "Python 3.9.5 ('base')",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.9.5 (default, May 18 2021, 14:42:02) [MSC v.1916 64 bit (AMD64)]"
    },
    "vscode": {
      "interpreter": {
        "hash": "975064d8d3d76841da6545d74f7ce08046b8a22628138d3f8f012e5fa8e266fc"
      }
    }
  },
  "nbformat": 4,
  "nbformat_minor": 5
}