{
    "cells": [
        {
            "cell_type": "code",
            "execution_count": null,
            "metadata": {
                "nbsphinx": "hidden"
            },
            "outputs": [],
            "source": [
                "!pip install graph-pes"
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {
                "vscode": {
                    "languageId": "raw"
                }
            },
            "source": [
                "\n",
                "# Custom training loops\n",
                "\n",
                ">  FYI, you can open this notebook in [Google Colab](https://colab.research.google.com/github/jla-gardner/graph-pes/blob/main/docs/source/quickstart/custom-training-loop.ipynb) and follow along interactively 😊\n",
                "\n",
                "\n",
                "`graph-pes` provides all the components you need to train a [GraphPESModel](https://jla-gardner.github.io/graph-pes/models/root.html#graph_pes.GraphPESModel) in any way you want.\n",
                "\n",
                "Here we'll implement a custom training loop involving gradient accumulation, and train a [SchNet](https://jla-gardner.github.io/graph-pes/models/many-body/schnet.html) model on the [QM7](https://jla-gardner.github.io/load-atoms/datasets/QM7.html) dataset.\n",
                "\n",
                "\n",
                "## 1. Data\n",
                "\n",
                "[GraphPESModel](https://jla-gardner.github.io/graph-pes/models/root.html#graph_pes.GraphPESModel) models act on [AtomicGraph](https://jla-gardner.github.io/graph-pes/data/atomic_graph.html#graph_pes.AtomicGraph) representations of atomic structures. `graph-pes` provides the [from_ase](https://jla-gardner.github.io/graph-pes/data/atomic_graph.html#graph_pes.AtomicGraph.from_ase) class method to easily convert between [ase.Atoms](https://wiki.fysik.dtu.dk/ase/ase/atoms.html#ase.Atoms) objects and [AtomicGraph](https://jla-gardner.github.io/graph-pes/data/atomic_graph.html#graph_pes.AtomicGraph) objects.\n",
                "\n",
                "Here we use the wonderful [load-atoms](https://jla-gardner.github.io/load-atoms/) package to load a dataset of [ase.Atoms](https://wiki.fysik.dtu.dk/ase/ase/atoms.html#ase.Atoms) objects, before converting them to [AtomicGraph](https://jla-gardner.github.io/graph-pes/data/atomic_graph.html#graph_pes.AtomicGraph) objects."
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 1,
            "metadata": {},
            "outputs": [
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "cb46438c21944202b9c325ba232ad433",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Output()"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "text/html": [
                            "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"></pre>\n"
                        ],
                        "text/plain": []
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "text/plain": [
                            "QM7:\n",
                            "    structures: 7,165\n",
                            "    atoms: 110,650\n",
                            "    species:\n",
                            "        H: 56.00%\n",
                            "        C: 32.32%\n",
                            "        N: 6.01%\n",
                            "        O: 5.40%\n",
                            "        S: 0.27%\n",
                            "    properties:\n",
                            "        per atom: ()\n",
                            "        per structure: (energy)"
                        ]
                    },
                    "execution_count": 1,
                    "metadata": {},
                    "output_type": "execute_result"
                }
            ],
            "source": [
                "from load_atoms import load_dataset\n",
                "\n",
                "structures = load_dataset(\"QM7\")\n",
                "structures"
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "Split the dataset into training, validation and test sets. You can see that each of these [load_atoms.AtomsDataset](https://jla-gardner.github.io/load-atoms/api/dataset.html#load_atoms.load_dataset) objects is just a lightweight wrapper around a list of [ase.Atoms](https://wiki.fysik.dtu.dk/ase/ase/atoms.html#ase.Atoms) objects:"
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 2,
            "metadata": {},
            "outputs": [
                {
                    "data": {
                        "text/plain": [
                            "Atoms(symbols='CNC5H13', pbc=False)"
                        ]
                    },
                    "execution_count": 2,
                    "metadata": {},
                    "output_type": "execute_result"
                }
            ],
            "source": [
                "train, val, test = structures.random_split([0.8, 0.1, 0.1])\n",
                "train[0]"
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "Convert the structures to [AtomicGraph](https://jla-gardner.github.io/graph-pes/data/atomic_graph.html#graph_pes.AtomicGraph) objects using a cutoff of 5.0 Å:"
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 3,
            "metadata": {},
            "outputs": [
                {
                    "data": {
                        "text/plain": [
                            "AtomicGraph(\n",
                            "    atoms=20,\n",
                            "    edges=314,\n",
                            "    has_cell=False,\n",
                            "    cutoff=5.0,\n",
                            "    properties=['energy']\n",
                            ")"
                        ]
                    },
                    "execution_count": 3,
                    "metadata": {},
                    "output_type": "execute_result"
                }
            ],
            "source": [
                "from graph_pes import AtomicGraph\n",
                "\n",
                "train_graphs, val_graphs, test_graphs = (\n",
                "    [AtomicGraph.from_ase(structure, cutoff=5.0) for structure in split]\n",
                "    for split in (train, val, test)\n",
                ")\n",
                "train_graphs[0]"
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "## 2. Model\n",
                "\n",
                "Before we instantiate our model, let's create a quick helper function to visualise the performance of our model on the training and test sets:"
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 4,
            "metadata": {},
            "outputs": [],
            "source": [
                "import matplotlib.pyplot as plt\n",
                "from graph_pes import GraphPESModel\n",
                "from graph_pes.utils.analysis import parity_plot\n",
                "\n",
                "%config InlineBackend.figure_format = 'retina'\n",
                "\n",
                "\n",
                "def analyse_model(model: GraphPESModel):\n",
                "    for name, data, colour in zip(\n",
                "        [\"Train\", \"Test\"],\n",
                "        [train_graphs, test_graphs],\n",
                "        [\"royalblue\", \"crimson\"],\n",
                "    ):\n",
                "        parity_plot(\n",
                "            model,\n",
                "            data,\n",
                "            property=\"energy_per_atom\",\n",
                "            units=\"eV / atom\",\n",
                "            label=name,\n",
                "            c=colour,\n",
                "        )\n",
                "\n",
                "    plt.legend(fancybox=False, loc=\"lower right\")"
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "Now let's create a [SchNet](https://jla-gardner.github.io/graph-pes/models/many-body/schnet.html) model. To account for energy offsets, we wrap this PES model in a [AdditionModel](https://jla-gardner.github.io/graph-pes/models/addition.html#graph_pes.AdditionModel), and combine it with a [LearnableOffset](https://jla-gardner.github.io/graph-pes/models/offsets.html#graph_pes.models.LearnableOffset):"
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 5,
            "metadata": {},
            "outputs": [
                {
                    "data": {
                        "text/plain": [
                            "AdditionModel(\n",
                            "  schnet=SchNet(\n",
                            "    chemical_embedding=PerElementEmbedding(dim=16, elements=[]),\n",
                            "    interactions=UniformModuleList(\n",
                            "      (0-2): 3 x SchNetInteraction(\n",
                            "        (linear): Linear(in_features=16, out_features=16, bias=False)\n",
                            "        (cfconv): CFConv(\n",
                            "          Sequential(\n",
                            "            (0): GaussianSmearing(n_features=10, cutoff=5.0, trainable=True)\n",
                            "            (1): MLP(10 → 16 → 16, activation=ShiftedSoftplus())\n",
                            "          )\n",
                            "        )\n",
                            "        (mlp): MLP(16 → 16 → 16, activation=ShiftedSoftplus())\n",
                            "      )\n",
                            "    ),\n",
                            "    read_out=MLP(16 → 8 → 1, activation=ShiftedSoftplus()),\n",
                            "    scaler=LocalEnergiesScaler(trainable=True)\n",
                            "  ),\n",
                            "  offset=LearnableOffset(trainable=True)\n",
                            ")"
                        ]
                    },
                    "execution_count": 5,
                    "metadata": {},
                    "output_type": "execute_result"
                }
            ],
            "source": [
                "import torch\n",
                "from graph_pes.models import AdditionModel, LearnableOffset, SchNet\n",
                "\n",
                "# ensure reproducibility\n",
                "torch.manual_seed(0)\n",
                "\n",
                "model = AdditionModel(\n",
                "    schnet=SchNet(\n",
                "        cutoff=5.0,\n",
                "        channels=16,\n",
                "        expansion_features=10,\n",
                "    ),\n",
                "    offset=LearnableOffset(),\n",
                ")\n",
                "model"
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "We can see that the model has not been fitted yet, and performs poorly on the training and test sets, predicting energies that appear to be very close to zero:"
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 6,
            "metadata": {},
            "outputs": [
                {
                    "data": {
                        "image/png": "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",
                        "text/plain": [
                            "<Figure size 350x300 with 1 Axes>"
                        ]
                    },
                    "metadata": {
                        "image/png": {
                            "height": 308,
                            "width": 354
                        }
                    },
                    "output_type": "display_data"
                }
            ],
            "source": [
                "analyse_model(model)"
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "While not essential, it is strongly advised to make use of the [GraphPESModel.pre_fit_all_components](https://jla-gardner.github.io/graph-pes/models/root.html#graph_pes.GraphPESModel.pre_fit_all_components) method before training. In all cases, this \"registers\" the elements in the training data with all of the model's [PerElementParameter](https://jla-gardner.github.io/graph-pes/building-blocks/nn.html#graph_pes.utils.nn.PerElementParameter) parameters, allowing for correct parameter counting (i.e. parameters corresponding to e.g. embeddings for elements that do not appear in the training set will not be counted in `graph_pes.utils.nn.count_used_parameters`).\n",
                "\n",
                "For the [AdditionModel](https://jla-gardner.github.io/graph-pes/models/addition.html#graph_pes.AdditionModel), the [GraphPESModel.pre_fit_all_components](https://jla-gardner.github.io/graph-pes/models/root.html#graph_pes.GraphPESModel.pre_fit_all_components) method dispatches to the [GraphPESModel.pre_fit_all_components](https://jla-gardner.github.io/graph-pes/models/root.html#graph_pes.GraphPESModel.pre_fit_all_components) method of its components:\n",
                "\n",
                "- in the case of [LearnableOffset](https://jla-gardner.github.io/graph-pes/models/offsets.html#graph_pes.models.LearnableOffset), this makes estimates of the **mean** per-element offset energies.\n",
                "- in the case of [SchNet](https://jla-gardner.github.io/graph-pes/models/many-body/schnet.html) model, and indeed all [GraphPESModel](https://jla-gardner.github.io/graph-pes/models/root.html#graph_pes.GraphPESModel) this also makes estimates of the **variance** in the per-element energies.\n",
                "\n",
                "We can see from the model summary below the effect that this pre-fitting has had:"
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 7,
            "metadata": {},
            "outputs": [
                {
                    "name": "stdout",
                    "output_type": "stream",
                    "text": [
                        "[graph-pes WARNING]: \n",
                        "Estimated per-element energy offsets from the training data:\n",
                        "    PerElementParameter(trainable=True)\n",
                        "This will lead to a lack of guaranteed physicality in the model, \n",
                        "since the energy of an isolated atom (and hence the behaviour of \n",
                        "this model in the dissociation limit) is not guaranteed to be \n",
                        "correct. Use a FixedOffset energy offset model if you require\n",
                        "and know the reference energy offsets.\n",
                        "\n"
                    ]
                },
                {
                    "data": {
                        "text/plain": [
                            "AdditionModel(\n",
                            "  schnet=SchNet(\n",
                            "    chemical_embedding=PerElementEmbedding(\n",
                            "      dim=16,\n",
                            "      elements=['H', 'C', 'N', 'O', 'S']\n",
                            "    ),\n",
                            "    interactions=UniformModuleList(\n",
                            "      (0-2): 3 x SchNetInteraction(\n",
                            "        (linear): Linear(in_features=16, out_features=16, bias=False)\n",
                            "        (cfconv): CFConv(\n",
                            "          Sequential(\n",
                            "            (0): GaussianSmearing(n_features=10, cutoff=5.0, trainable=True)\n",
                            "            (1): MLP(10 → 16 → 16, activation=ShiftedSoftplus())\n",
                            "          )\n",
                            "        )\n",
                            "        (mlp): MLP(16 → 16 → 16, activation=ShiftedSoftplus())\n",
                            "      )\n",
                            "    ),\n",
                            "    read_out=MLP(16 → 8 → 1, activation=ShiftedSoftplus()),\n",
                            "    scaler=LocalEnergiesScaler({'H': 0.1, 'C': 0.592, 'N': 0.58, 'O': 0.49, 'S': 0.1}, trainable=True)\n",
                            "  ),\n",
                            "  offset=LearnableOffset({'H': -2.98, 'C': -6.68, 'N': -4.29, 'O': -4.25, 'S': -3.37}, trainable=True)\n",
                            ")"
                        ]
                    },
                    "execution_count": 7,
                    "metadata": {},
                    "output_type": "execute_result"
                }
            ],
            "source": [
                "model.pre_fit_all_components(train_graphs)\n",
                "model"
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "Analysing the model, we can see that this pre-fitting has drastically improved the model's performance for ~free! This makes the following training loop more robust."
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 8,
            "metadata": {},
            "outputs": [
                {
                    "data": {
                        "image/png": "iVBORw0KGgoAAAANSUhEUgAAAuAAAAJoCAYAAAAnPVS8AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuNSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/xnp5ZAAAACXBIWXMAAB7CAAAewgFu0HU+AADNQ0lEQVR4nOzdd1iT5/4G8PsNIxEFEUFEBFHEPbCeaqtFqbNH66p2WKtWW0eHVq3drg5PT5e12uE4ravV/uqgrlpbrai11q24UESjoCgiIARICOT9/UETGUnIfEPg/lyX10XevOOhQ+48PM/3K4iiKIKIiIiIiCQhc/UAiIiIiIhqEgZwIiIiIiIJMYATEREREUmIAZyIiIiISEIM4EREREREEmIAJyIiIiKSEAM4EREREZGEGMCJiIiIiCTEAE5EREREJCEGcCIiIiIiCTGAExERERFJiAGciIiIiEhCDOBERERERBJiACciIiIikhADOBERERGRhBjAiYiIiIgkxABORERERCQhT1cPgNxbUVERbt68CQBo2LAhPD35nxQRERGROZwBJ7vcvHkTYWFhCAsLMwRxIiIiIjKNAZyIiIiIJKVSqdCwYUMIggBBEFC7dm0UFRW5eliSYQAnIiIiIsmoVCo0b94ct27dMhx77733atQyVgZwIiIiIpLMAw88UCZ8f/rpp3j11VddOCLpMYA7wfbt2zFv3jwMHDgQrVu3RmBgILy8vFCvXj107twZr776Ki5cuODw554/fx6vvfYaOnTogHr16qFWrVqIiIhAnz598J///AdKpdLhzyQiIiKyxtKlSw2z3TUxfAOAIIqi6OpBVCdFRUXw8vKq9DwvLy+89957ePPNN+1+piiKmDVrFj7++GOz66c+//xzTJs2ze7nlZaamoqwsDAAQEpKCho3buzQ+xMREVH1c+DAARw9ehSvvPKKq4fiEjVnsY2E6tati9jYWHTt2hXNmjVDSEgIfHx8cOPGDcTHx+O7777D3bt38dZbb8Hf3x+TJ0+263kTJ07E//73PwBA586dMX78eLRv3x516tRBeno6Dh8+jI0bN0IQBEd8e0REREQWU6lU0Ol08PPzMxzr3r07unfv7sJRuRZnwJ2guLgYHh4eJt+/cuUKOnfujKysLAQFBSEtLc3s+eZ89913eO655wAAb7/9Nj744AOTQbuwsBDe3t42PccUzoATERGRKfoNlzqdDpcuXSoTwmsyrgF3gsrCdNOmTfHEE08AAG7fvo3ExESbnpObm4uZM2cCAAYOHIj58+ebneV2dPgmIiIiMqV0tZPbt2+jbdu2rh5SlcEA7iK+vr6Gr9VqtU33+OGHH5CVlQUAeOeddxwyLiIiIiJ7GSs16Oh9aO6MAdwFCgoKsHnzZgCATCZDixYtbLrP+vXrAQCBgYF48MEHDcfT09Nx6dIl5OTk2D9YIiIiIisYC981tdqJKdyEKRGtVou0tDT89ddf+Oijj5CUlAQAGD9+fJnZcEvpdDocOXIEANC+fXuIoogvv/wSn3/+Oa5cuWI4r2PHjpgyZQrGjRsHmcz6z1upqalm309LS7P6nkRERFQ9MXxbhpswnUipVKJp06Ym3+/fvz9++uknmzYkXL16FREREQCA4cOHQ6fTIS4uzuT5Q4YMwU8//WT1OnBrKqdwEyYREVHNxfBtOS5BcYHAwED83//9H7Zv327zbuDMzEzD19u3b0dcXBxatGiBLVu2ICcnB7m5udi2bRtatWoFANi8ebNDao4TERERlcfwbR3OgDuRVqs1dLwsKirC9evX8euvv+Lbb7+FQqHAa6+9hrfeesume//555+IiYkxvG7YsCESEhIQFBRU5ryMjAx07NgRN27cgJeXFy5fvmzVLLUlS1C6dOkCgDPgRERENdXNmzfRsmVLw/4zhm/zauwMuCAIdv9ZuXKl2Wd4eXmhXbt2aNeuHaKjozFw4EAsXrwYf//9NwRBwNtvv43x48fbNH6FQlHm9euvv14hfAMls+1vv/02gJIPBJs2bbLqOY0bNzb7JyQkxKbxExERUfXRsGFDnD9/HvXr12f4tkCNDeCu1KFDB3zwwQcAgBUrVuC3336z+h7lN27269fP5Ln9+/c3fK3fuElERETkSI0aNUJ6ejrDtwVqbBWU8+fP230Pe2Z/hwwZghdffBEAsGHDBrMB2pjGjRtDEAToVxDpu1EaU/q927dv2zBaIiIiontUKhV69uyJ1atXl2mwY0vFtZqoxgZw/eZEVym9XOTq1atWX1+7dm00adIESqUSAFBcXGzy3NLveXrW2H/lRERE5AClN1x27twZx44dY5dLK/Fjiotcv37d8HWdOnVsukePHj0MX1++fNnkecnJyYavQ0NDbXoWERERUflqJxqNBnv37nXxqNwPA7iL6LtYAiWNdGwxfPhww9fmaoCX3nhZunIKERERkaVMlRrUL6kly7EMoYP9/PPP6Nq1q9n14fv27cPAgQOhUqng6emJM2fOoGXLlhXO0zfBKb3UpDSdToeOHTvizJkzqF27Ng4dOlThV0Dnz59Hly5doFKpEBQUBKVSCR8fH/u+yVJSU1MNa8xZhpCIiKh6Yp1vx+KCYAf7+eef8eSTT2LgwIHo3bs32rZtC39/f2g0GiQnJ2Pr1q346aefoNPpAABz5swxGr4tIZPJ8M0336B3797Iy8vDQw89hNdffx0PP/wwgJKg/9///hcqlQoAsHjxYoeGbyIiIqr+GL4djwHcCQoLCxEXF2d2WUitWrXwwQcfYMaMGXY966GHHsLatWsxbtw4ZGdnG2p+l+bl5YVFixbhySeftOtZREREVLMwfDsHA7iDffzxx+jZsyf27duHM2fO4NatW0hPT4dMJkNAQADatm2LXr16YcyYMQ5rYjN8+HDcf//9WLx4MbZv346UlBTodDqEhYWhT58+mDZtGpo3b+6QZxEREVHNMW7cOIZvJ+AacLIL14ATERFVX4WFhWjbti0uXbrE8O1AnAEnIiIiIqO8vb1x9uxZbN++HcOGDXP1cKoNliEkIiIiIgAla753795d5pi3tzfDt4MxgBMRERGRYcNl3759sXHjRlcPp1pjACciIiKq4UpXOxFFEU8++aShjDE5HgM4ERERUQ1mrNTgRx99hDp16rhwVNUbAzgRERFRDcU6367BAE5ERERUAzF8uw4DOBEREVENw/DtWpLXAb99+zYuX76MmzdvIi8vD15eXvD390d4eDiaN28ODw8PqYdEREREVGMUFRUhKiqK4duFnB7A8/LysHnzZuzYsQN79+7F9evXTZ4rl8vRqVMn9OvXD8OGDUOHDh2cPTwiIiKiGsXT0xOdO3fG9u3bATB8u4LTWtGfOHECixcvxvr165Gfnw8AsPRRgiAAANq2bYuXXnoJo0ePho+PjzOGSXZiK3oiIiL3NGzYMDz00EMM3y7g8AB+4sQJzJ49Gzt27ABwL3Q3bNgQXbp0QefOndGgQQMEBASgXr16KCgoQGZmJrKysnDx4kUcOXIECQkJ0Gq1JQMUBNSvXx+vv/46pkyZArlc7sjhkp0YwImIiIis49AAPm7cOKxZswY6nQ4AcN9992HUqFEYPnw4wsPDLb5PYWEh9u3bhx9++AFxcXHIycmBIAho0qQJVq9ejYceeshRQyY7MYATERFVbSqVCq1atcL48ePx3nvvuXo4BAdXQVm1ahU8PT0xYcIEJCYm4ujRo5g+fbpV4RsAvL290adPH6xYsQK3bt3C6tWr0bJlSyiVSvzxxx+OHDIRERFRtaWvdnL9+nW8//77mDNnjquHRHDwJswXX3wRb7zxhmFG1BHkcjmeeeYZjBo1CuvXr0dxcbHD7k1ERERUXRkrNVi3bl0Xjoj0nLYJk2oGLkEhIiKqeljnu2qTvA44ERERETmPFOFbmabF1v25SErRIl+tg49ChqgwLwyK8UVEiJfDnlNdMYATERERVRPODt+JSg2WxmXjVJKmwntnkjWIi1ehY5Qck4b5o1UEK9eZwlb0RERERNWAs8P3obMFmLEw3Wj4Lu1UkgYzFqbj0NkChzy3OpJ8BvzUqVPYv38/Ll++jNzc3Eo3VQqCgG+//Vai0RERERG5p7Vr1zp15nvesgxotJZtHVQXinh3eQYWTGvAmXAjJAvg58+fx3PPPYdDhw5ZfI0oigzgRERERBaYOHEiUlNT8cEHH+CTTz5x6JrvpXHZFodvPXWhiGVx2VgwPdhh46guJAngly9fxkMPPYTs7GxDZ0xfX1/4+/tDJuMqGCIiIiJHeO+99/Dcc8+hSZMmDrunMk1b6bITU04maXA1TYsm3JhZhiQBfM6cOcjKyoJMJsPMmTPxwgsvICIiQopHExEREVVLKpUK77//Pj766KMyxx0ZvgFg6/5cu67fsj8XU54IcNBoqgdJAviuXbsgCAKmTZtW4T8SIiIiIrJO6Q2X586dw9atW532rKQUrV3XX7Lz+upIkvUfOTk5AIDhw4dL8TgiIiKiaqt8tZNt27bhyJEjTntevlpn3/Ua+66vjiQJ4PpOiZ6eLDtOREREZCtjpQY/++wz3H///U57po/CvrjoI+d+v/Ik+SfSv39/AMDhw4eleBwRERFRtWMqfM+YMcOpz40Ks28DZXM7r6+OJAngr776Knx9ffHJJ58gMzNTikcSERERVRuuCt8AMCjG167rB9t5fXUkSQBv0qQJNm3ahMzMTHTr1g2///67FI8lIiIicnuuDN8AEBHihY5RtjXTiY6SswShEYKoL8wtgeTkZHTr1g0ZGRmoV68emjdvDh8fH7PXCIKA3bt3SzRCslZqaqphjX9KSgoaN27s4hERERFVLy1atEBSUpLhtZThWy9RWdJeXl1oeWxUeAvshGmCZAH8r7/+wujRo6FUKmHJIwVBMHTCrKxdPbkOAzgREZFzbdy4EY8//jhEUXRJ+NY7dLYA7y7PsCiEK7wFzJ0QiK5ta0kwMvcjSQA/d+4cunbtivz8fIiiCIVCgaioKIs7Ye7Zs8fZQyQbMYATERE538aNG3H16lWXhW+9RKUGy+KycdJMZ8zoKDkmDvPnzLcZkgTwJ598EuvXr4dcLsdnn32G8ePHQ6FQOPuxJAEGcCIiIsdSqVRQKBRVunyzMk2LrftzcSlFi3yNDj5yGZqHeWFwjC/XfFtAkn+zBw4cgCAIePvtt/Hiiy9K8UgiIiIit6PfcOnn54dz585V2RAeEeLF9vJ2kKQKSlZWFgDgkUcekeJxRERERG6ndLWTpKQkREdHu3pI5CSSBHD9sgRupiQiIiKqyFipwfHjx7twRORMkgTwQYMGAQD27dsnxeOIiIiI3Iar63yT9CTZhHnr1i1ER0ejqKgIR44cQUREhLMfSRLhJkwiIiLbMXzXTJLMgAcHB2Pnzp3w8/ND165dsXz5cmRnZ0vxaCIiIqIqieG75pJkBrxZs2YAgPz8fKSnp0MQBAiCgMDAQIs6YSYnJzt7iGQjzoATERFZj+G7ZpOkto1SqSzzWhRFiKKI9PT0Sq8VBMFJoyIiIiJyjevXr5dZDWBt+NbX4U5K0SJfrYOPQoaoMC/c10qB44nqCscHxfgigvW5qwxJAvjYsWOleAwRERGRW2jZsiWOHTuGrl274r333rM4fCcqNVgal41TRjpRnknWIC5eZfJ4xyg5JrFDZZUgyRIUqr64BIWIiMh2Op0OMpllW/IOnS3Au8szoC60PbopvAXMnRCIrm1r2XwPsp8kmzCJiIiIajKVSoUePXrg5s2bZY5bGr4TlRrMW2Zf+AYAdaGId5dnIFFZcQadpMMATkRERORE+g2X+/fvR8uWLSuEcEssjcuGRuuYRQvqQhHL4rIdci+yjSRrwI25desWzpw5g8zMTABAQEAA2rVrh+DgYFcNiYiIiMihylc7ycnJwfbt2/Hcc89ZfI8DCflG13zb42SSBlfTtGjCjZkuIWkAF0URy5Ytw5dffolz584ZPadNmzaYMmUKJkyYwAooRERE5LZMlRq0NHyb23DpCFv252LKEwFOuTeZJ9kSlKysLPTo0QMvvvgizp07ZyhFWP7PuXPn8MILL6BHjx5s1kNERERuyd4634fOFmDGwnSnhW8AuJSiddq9yTxJZsBFUcSQIUNw4MABAED9+vXxxBNPoGvXrmjYsCEA4ObNmzh8+DB++uknZGRk4K+//sKQIUOwd+9eKYZIRERE5BD2hm/9hktHrfk2JV+jc+r9yTRJAvjatWvx559/QhAEPP300/j666/h6+tb4bwxY8bgv//9L1566SWsWbMGf/75J9atW4eRI0dKMUwiIiIiuziiw6UjN1ya4yNnLQ5XkeSf/Nq1awEAPXv2xJo1a4yGb706depg1apV6NmzJ0RRxPfffy/FEImIiIjsNnz4cLvCtzJN69RlJ6U1D+MGTFeRJIAfP34cgiDg5ZdftviaKVOmAABOnDjhrGEREREROdT69esRFBQEwPrwDQCrt2c7YVTGDY4xPSFKziXJEhR9qcGmTZtafI3+XP21RERERFWdn58fLl26hC1btuCZZ56x6tpDZwuw90SBk0ZWVnSUnCUIXUiSGfC6desCAG7cuGHxNWlpaQBK/kMmIiIiqopUKhWOHz9e5pifn5/V4Vu/8VJ0/tJvKLwFTBzm7/wHkUmSzIC3a9cOe/fuxYoVKzBw4ECLrlmxYoXhWiIiIqKqRr/h8s6dO4iPj0f37t0tuk6ZpsXW/blIStEiO7cYufk65OTpIIpASE4qeibvRniWEooiNdSeClyrF4G9kb2R5tfY7jErvAXMnRCIVhFyu+9FtpMkgI8YMQLx8fGIi4vDvHnzMHfuXLNNdt5//31s3LgRgiDg8ccfl2KIRERERBYrX+2kd+/eyMnJgbe3t8lr9hzLw+KfspCdW7H8X0RmMkYkrEPL2+crvBd15yJ6X/oNF4JaY0OHkVAGRNo05ugoOSYO82f4rgIEUXT+Lzu0Wi06dOiACxcuQBAEtG3bFs8++yy6du2KBg0aQBAE3Lp1C4cOHcKqVatw5swZiKKI1q1b49SpU/D0lLRhJ1khNTUVYWFhAICUlBQ0bmz/p3MiIqKqzNpSg4lKDT74LgM3MoorvBeSk4rHEv4PHdJOQIbKI5nGQ44lD07FmZBoo+9PfdIfKbeKcClFi3yNDj5yGZqHeWFwjC/XfFchkgRwAFAqlejduzeuXLlSaYt5URTRrFkz/PHHHwgPD5dieGQjBnAiIqpJrA3fh84WYNaS2ygul73NzXhXRuMhx6ex71SYCY+OkmPB9GCr70fSk6wCe0REBBISEvDqq6+ibt26JlvR161bFzNnzsTJkycZvomIiKjKsCZ8H0jIx4i3UvHWVxXDd7u0k5gZP9+m8A0A8mINhiesK3PMyxPcWOlGJF3bUbt2bXzyySeYP38+jh07hjNnzhjKDAYEBKBdu3bo3Lmz2fVTRERERFKzNHz/elCFL37MhEZr/D4RmcmYfPALyIsL7RpPq9vn0TDnOm76hdp1H3INyZagUPXEJShERFTd5efno1mzZpWG7//OPYqA33YYKpjoBBm8izTw1dyFokgNmShCgAjzC3Ett7t5P/zYaazhNZeguA9JZsBXr14NABg6dKjFdb1VKhU2bdoEABgzZozTxkZERERkjkKhQOPGjQ0BvHz4Xr/oMGov+R8et3FJia3Cs6+WeX0ySYOraVputnQDksyAy2QyCIKA06dPo02bNhZdk5ycjKioKMhkMhQVFTl5hGQrzoATEVFNoNPp8NBDD2HEiBFlwvd3b+7EAys/gbxYI/mYUuqG471+H5Y5Niy2DqY8ESD5WMg6Vb6+H1fIEBERkavJZDL89ddfZY6tX3QYD6z82O713LZSe9WqcOxSionF51SlSFYFxVrF/2wZZg1wIiIikpJKpULTpk2xbNkys+fVXvI/l4VvALjm36TCsXxNxSY/VPVU2XR74cIFACXVUYiIiIikULrayaRJkwAAEydOrHDeL2vO2FxG0FHiI/tUOOYjr7Jzq1SKUwL4vn37jB4/cuQIMjIyzF6r0WiQnJyMTz/9FIIgIDo62gkjJCIiIirLWKlBlUpV4bz5M/9E/58+lnJoFSQGtTZagrB5GDdgugOnBPDY2NgK3S5FUcT48eMtvocoihAEwfDpk4iIiMhZLKnzfSAhHz/N3YVJBxe5ZNOlnsZDjo0dRhp9b3CMr8SjIVs4bQmKsc2T1myobNy4Md5++20MHTrUgaMiIiIiKquy8P3rQRU+X5uJ0NvJmOmAJjr20HjIseTBqRXa0AMldcBZgtA9OCWA79mzx/C1KIro1asXBEHAt99+i6ZNm5q8ThAEKBQKhISEGErbERERETlLZeF79S/ZWLktByE5qXjxr89dGr4Tg1pjY4eRRsO3wltgK3o34pQA3rNnT6PHu3TpYnEdcCIiIiJnsmTmO371CcxMWOeyDZc6CEgIicbGDiNNtp1XeAuYOyEQrSLkEo+ObCVJFZQrV64AAEJDjf+HQ0RERCS1xYsXGw3fV/5MwuWFG1H7xEm8lXMDMrimJ4m5GW+96Cg5Jg7zZ/h2M5IE8CZNKtapJCIiInKlt956CxcvXsTKlSvx2WefYWDL3tjXeRJCr51DhAvGo/aQI6VeBK75N0F8ZJ8yM97163qgUaAn8jU6+MhlaB7mhcExvlzz7aYkaUVP1Rdb0RMRkbs7e/YsNIcyoHj7fZdVNymUeeGTh2ebnO2ObOyF5W+HSDwqchbJG/Hk5uZi165dOHXqFDIyMlBQUGC2Oop+8yYRERGRPVQqFZYuXYpXX321zHFPZRFkb78HbxdtsNTKPPFNt2lml5qwwU71IlkA1+l0eP/99/HZZ58hLy/Pomv0tcAZwImIiMgepTdcXrp0Cd98843hvVtzvkaoi8K30j8CP3QebzZ8A2ywU91IFsCfffZZ/PDDDxBFER4eHqhfvz7S09MhCAIaN26MrKwsQ7cpQRAQGBgIHx8fqYZHRERE1VT5aidLlizBtGnT0LJlS1z5Mwmh185JPqYcb1988vBsk5VNymODnepFkt9n7Ny5E99//z2AkiCenp6OXbt2Gd6/evUqcnJycP78eUydOhUymQz16tXDjh07DBVUiIiIiKxlqtRgy5YtAQCXF250ybiOhD9ocfhmg53qR5IAvmLFCgBA27Zt8d1336FevXoVWtUDQMuWLbFw4UJs2rQJycnJGDBgAO7evSvFEImIiKiasaS9vOziJVcMDfGRfSw6jw12qidJAvjff/8NQRDw0ksvWXT+oEGDMHbsWFy9ehWLFi1y8uiks2PHDgiCYPgzb948u+6n1Wrx66+/Yvr06ejWrRsCAwPh5eUFf39/3HfffXjttddw+fJlxwyeiIjIjRgL3x+98TaGXgWOdJ2Iv9s8g11tn0eDdKXkY0sMam3R7Dcb7FRfkqwBT09PBwC0aNHCcMzDw8PwtUajgVxe9j+uESNG4LvvvkNcXBxmz54txTCdKi8vDy+88ILD7nf79m20bt0ad+7cqfDe3bt3ceLECZw4cQKLFi3Cxx9/jFdeecVhzyYiIqrKjIXvN1t0w/Bv90EEEPDPsSAXjE3jIcfGDiMrPY8Ndqo3ScsQBgQEGL729b23mSA9Pd1QS1qvQYMGAAClUinJ2Jxt9uzZuHr1Kho0aGD4QGIPjUZjCN/R0dEYMmQIunbtiuDgYNy9exc7duzA4sWLoVarMW3aNNSqVQsTJ060+7lERERVQeGFK0j9Og53/k5EcW4+dIVa+BTkwKtYi363/8Qt3b163m/5ROC5TNeX8dN4yLHkwakVKp4IAhAc4IEgf0822KkhJAngwcHBuHbtGjIzM8sc8/b2hlarRUJCQoUAfvXqVQCAWq2WYohOdezYMSxatAhyuRzz58/HhAkT7L6nIAjo27cv3nvvPTzwwAMV3n/44YcxfPhwPPzwwygoKMDrr7+OkSNHlvngQ0RE5E6UaVrs+/44Gv/wHZpcL6lcEmDkvOdqheL9vJIlmG/5ROA5H9c3iTPXVl4UgexcHeY85w+FXIYt+3ORlKJFvloHH4UMUWFeGBTjiwiG8mpDkgDevn17XLt2DefOncPDDz9c8mBPT3Tq1AmHDx/GihUrMHDgwDLX6Otzunsb++LiYkyYMAHFxcWYM2cOmjdv7pD7hoaG4rfffjN7TteuXfHiiy/is88+w927d/H777/jsccec8jziYiIpJKo1GBpXDaK9x3C5IOLKu1WObZWIwCAVtRVifBdKPPCzpaPmq31rS4U8dridOQVVGxOeCZZg7h4FTpGyTGJy1KqBUl+HxMbGwtRFMuUHgSAZ555BqIoIi4uDmPHjsX27dvx008/YeDAgdi1axcEQcCQIUOkGKLTfP755zhx4gRatGiBN954Q/Ln6z/wAEBycrLkzyciIrLHobMFmLEwHXcPncPkg18YDd9qnQ46na7MsbG1GuH5KhC+AcBbp8Xkg4sQkWn+57Cx8F3aqSQNZixMx6GzBY4cHrmAJAF82LBhAIBff/21zIaISZMm4b777oMoivj+++8xePBgjBw5Er/++isAIDw83CWh1VGUSiXmzp0LoGRGv/xGUyloNPf+oiq98ZWIiKiqS1RqMG9ZBtSFIkYkrIPcSLdKla4IsVlH8GTO6QohvCqRF2swPGGd3fdRF4p4d3kGEpXmfwtAVZskAbxp06a4fPkyzpw5Az8/P8NxT09P/P777xg1ahQ8PT0hiiJEseTT38CBA7F//37Uq1dPiiE6xQsvvID8/HyMGjUKvXr1cskY9u7da/i6devWLhkDERGRLZbGZUOjFRGSk4qWt89XeF+lK0KfrGPIELU4UZSLZ3LOuGCUlmt1+zwa5ly3+z7qQhHL4rLtHxC5jGRVUCIiIower1evHtasWYOvv/4aSUlJKCoqQvPmzctUTHFHa9euxa+//gp/f38sWLDAJWNIS0szNEEKCgoqsxzFUqmpqZU+g4iIyNGUaVqcSiqZ5e2ZvLvC+6XDt15v76qfHWKTd+HHTmPtvs/JJA2upmlZLcVNSVqG0BxfX1/cd999rh6GQ2RmZmL69OkAgA8//NBQUlFKoihi0qRJyM3NBVBSBlGhUFh9n/LVaYiIiKSwdX+u4evwLGWZ94yF76pS7aQy4dlXHXavLftzMeWJqv+hgypyfVHMamjmzJlIT09H165dXVZ7+z//+Q+2bt0KoGQjpqVdSImIiKqCpJR74VpRdK8ksTuHbwBQaB23gfJSqX9G5F6qzAy41ARBsPseK1aswLPPPlvmWHx8PFasWAEPDw8sWbIEMpn0n3F++OEHQ/fQpk2bYu3atTaPIyUlxez7aWlp6NKli033JiIiMiUju8jwtdqz5De47h6+AUDtVcth98rXVN1Np2RejQ3gzqDRaDBp0iQAwNSpUxEdHS35GLZv345x48ZBFEU0bNgQv//+Oxo2bGjz/Ro3dp+/1IiIqHpIVGpwK7PY8PpavQiE3D7n9uEbAK75O66/iY+cCxncVY0N4OfPV9xNba2QkJAyrzdt2oSLFy/Cy8sLbdq0wY8//ljhmnPnzhm+PnPmjOGcrl27omnTpnaNJz4+HiNGjIBWq0W9evWwc+dOREaaLvpPRERUFSjTtNi6PxenkzVIzyxGbp4OpSti743sjcDzG5Hp5uEbAOIj+zjsXs3DuAHTXdXYAN6qVSuH31Nfc1ur1VrUbn7jxo3YuHEjgJLlLPYE8MOHD2PQoEFQq9WoU6cOduzYgQ4dOth8PyIiImfTd7jUVzsxJiIzGU8fX4mmXr5Y7dce43LOYKZPE7cM34lBrXHTL9Rh9xsc4+uwe5G0amwAr04SEhLwyCOPQKVSQaFQYOvWrejataurh0VERGTSobMFeHd5SZMdU3pf3IEnTn1vqBjxgHddnAh4EAoX7K+yl8ZDjo0dRjrsftFRcpYgdGPu919wFfbss88amgmZ+rNnzx7D+XPnzjUcL7+Z01IXL15Ev379kJWVBS8vL2zcuBGxsbGO+YaIiIicoHSHS1O6n4vDb/veQr6uqMxxdw3fSx6cCmWAY5aFKrwFTBzm75B7kWu433/FNYRSqYQgCBAEwWSgvnbtGvr06YNbt27Bw8MDa9euxYABA6QdKBERkZX0HS5NaXTrNJbsew37tFnom3UMqnIh3J0kBrXGp7Hv4ExItMlzvDxL/lhC4S1g7oRAtIqQO2aA5BIOX4KSmZnp9l0s3cGdO3fQp08fQ5nAV199Fa1atcKZM6bb8NarVw+hoY5be0ZERGSt0h0ujSkqVOGXzSOQ9c+Gy9uiFnsKMzFIIX1TO2uJALLldZHh2xDX/JsgPrJPpWu+o6PkhtnsZXHZOGnmn43+XIZv9+fwAB4cHIxu3brh0UcfxaBBg5yy2ZGA06dPIykpyfD6448/xscff2z2mrFjx2LlypVOHhkREZFppTtclldUqMKRHx6Cuvhe4523fCLcInzrl5mYm+kGgLq1ZQhv6IXmYV4YHONbZh33gunBhoowl1K0yNfo4COXGT2X3JvDA3hxcTH279+PP//8E2+++SaaNWtmCOM9e/aEh4eHox9JREREbiLJRPfGokIV/l7bE1p1luGYu5QaTAxqjY0dRlq0xnvhjGCzQToixIvt5WsAQRRF04uwbHDw4EFs3boVW7duxdmzZ0se8k/XST8/P/Tr1w+DBg3CgAEDuFSlGkhNTUVYWBiAkq6ZbNxDRETmPD8/DZevlw3hhvBdkGE4VtXDtw4C9jTva9EyE73oKDkWTA928sjIHTg8gJemVCqxdetWbNu2DXv37kVhYWHJQwUBMpkMDz74oGF2vHXr1s4aBjkRAzgREVlj6me3cCb53jpndwzfAFDgWQtTh/3P4vMV3gIWTGvA9dsEwMkBvDSVSoWdO3di69at2LFjB27fvl0ygH9mx5s2bYpBgwZxqYqbYQAnIiJrLP4pE3HxKsPrY5uGIefWccPrCUFd8Ibo7YqhWSWlbjje6/ehRefqK5d0bVvLyaMidyFZAC9NFEX8/fffJpeq+Pr6on///lyq4gYYwImIyBrKNC3Gv59meK3OvY7D/9cHxdp8RHabhdd13uh96TcXjtAyi7q/itON7qv0PFYuIWNcEsDLu3r1qiGMG1uq8sADD2Dw4MF46qmnDGGPqgYGcCIistb0z2+VKUWozr2OzJR9aNRmJEJyUvHezjdcOLrK5Xn5YNrQ5ZWeN39yIB7s4CPBiMjdVIlGPE2aNMHLL7+MnTt3IiMjA+vXr8fYsWMRGBiI4uJiHDhwAG+++SZWrFjh6qESERGRHVQqFQb8KwsKb8FwTOEbikZtStq0p/k1xoWgqrsvTATwf9GjKz0vOkrO8E0mObwMob3q1KmD4cOHY/jw4YalKlu2bMH27dsNS1SIiIjIvSjTtFi8LgVfzuqKosIczGv3Au5X50JRpIbaU4E7tQMBAPXzMuCnvgsdqsgsYSkigM1thuNgRA+z57FVPFWmSixBsZRWq4WXF4vQVyVcgkJEVHMdSMjHiq13cfNOEbRFInQ6QBSB4Lup6Jm8G+FZSvgXZMJPkwO1Vo3+WUeQ8U+Hy9qQ4UTAA5DJqlrMNi7Pywf/Fz3aovDNDZdUmSo3A24OwzcREZHr/XpQha82ZCGvoOwcXkRmMkYkrEPL2+fLHFfpitA/65ghfAPAVJ9wtwjfIoDfox7BeguXnXDDJVnCrQI4ERERudbqX7KxcltOhePt0k5i8sFFkBdryhxX6YrQp1z4doc638C9JSfb2z5m8pywYE/8q7WCreLJKgzgREREZJFfD6qMhu+IzGRMPvgF5MWFZY67c/gu8FRgXaexlS45qe/nwdbxZLWq/7sfIiIiqhK+2pBl9PiIhHXVKnyn+DXG1GHfVhq+AeBkkgbzV2RAmaat9FwiPc6AExERUaUOJORXWPMNACE5qUbXfLtr+AaAsJxUNMy5jpt+oRadv/tIPnYfyUfHKDkmcQ04WYAz4ERERFSpFVvvGj3eM3l3hWPekEEu3IsY7hS+9WKTd1l9zakkDWYsTMehswVOGBFVJwzgREREVKmbd4qMHg/PUlY45i2T4Xf/zmgqq+WW4RsAwrOv2nSdulDEu8szkKjUVH4y1VgM4ERERFSpoiLjbUMURWqjx71lMvwe0NktwzcAKLS2z2KrC0Usi8t23GCo2nF4AJ8+fTpOnjzp6NsSERGRC3l6Gu9GrfZUQKUrQt/Mo/hDc0fiUTmP2su+RjonkzS4yo2ZZILDA/gXX3yBzp07o0OHDvj000+Rlpbm6EcQERGRxBrWN1634aJvCPpkHcMVnRqTcs9XmxB+zb+J3ffYsj/XASOh6sgpS1BEUcTZs2fxxhtvIDw8HP3798fatWtRUMBNCURERO5o3KC6FY4VFarwxpmvDdVORADK4urxsz4+so/d97iUwhlwMs7hAXznzp145pln4OPjA1EUUVxcjF27dmH06NFo2LAhxo8fjz179jj6sURERORE3Tv4oHate8tQigpV+HttT6jV92qDv+UTgfFuuua7tMSg1haXIDQnX6NzwGioOnJ4AO/bty9Wr16NW7duYfXq1ejbty8EQYAoisjNzcWqVavQp08fNGnSBO+88w4SExMdPQQiIiJygpdG1ANwL3xrCzIM771Wp7nbbrgsrVDmhY0dRjrkXj5y1rog4wRRFI1va3agtLQ0/PDDD/j++++RkJBw7+FCySfpzp07Y+zYsXjqqadQv359Zw+HHCg1NRVhYWEAgJSUFDRu7P5/+RIRERAXn4NV23OgytdBFAFBAHwUAsQiFXYu71EmfPdr8Tg+z86At04L41s13YNW5oWvu03DmZBoh9xvWGwdtqknoyQJ4KWdPn0aq1evxrp163Djxo2SQfwTxL28vPDII49gzJgxGDRoELy8vKQcGtmAAZyIqHpZsTUbP+zMgc7I6gljM99v+jTF8z72L9dwNaV/BH7oPB7KgEiH3XPF7BA0CWGWoYokD+B6oihi9+7dWLNmDTZt2oS8vLySAf0TxuvVq4cnn3wSo0ePxgMPPOCKIZIFGMCJiNyTMk2LrftzcTpZgxvpWhRoSjZRmpO253UkJv6f4bW7NtkBSr7XQpkXjjfugl9aD3HImu/SoqPkWDA92KH3pOrDZQG8tPz8fGzatAlr1qzBH3/8geLiYsN7MpkMRUXGu2+R6zGAExG5l0SlBkvjsnEqyfJOjf+6dhBPnVyNupocjL17Bge02W4XvsV//tzxCcS6TmNxutF9TnuWwlvAgmkN0CpC7rRnkHurEgG8tLNnz+Kpp57CuXPnIIoiBEEoE8ipamEAJyJyH4fOFuDd5RlQF1r2oz8iMxnP//0lgvPSyxw/VngXnb0rliWs6jQecix5cKpVa7wV3gKe6ueLH3/Lteifm8JbwNwJgeja1r5GPlS9VYntuVqtFps2bcJjjz2Gzp0749y5c64eEhERUbWSqNRg3jLLw3e7tJN4adds/JlxusJ77hi+AUBerMHkg4sQkZls0fnRUXIsmNYAYwb4Y8G0BoiOMj+jrT+f4ZsqY7ytlUQOHDiANWvWYP369cjOzgZQsjYcAHx9fTFixAiMHTvWhSMkIiKqHpbGZUOjtXzme8yfn+CRzMPIELW4JRbiRZ8wJ49QGvJiDYYnrMOC2Flo1tgLMgEo1gEeMkAnlpQObB7mhcExvmU2ULaKKFnTrV87fylFi3yNzuT5ROZIHsCTkpKwZs0a/PDDD1AqlQDuhW4PDw/06dMHY8aMwbBhw6BQKKQeHhERUbWjTNNateZ7wMlV+PedQ4YOl5/nX8UweRBCPKrHz+VWt88jOOc6xj/aEQ928LHq2ogQL5YWJLtJEsAzMjLw448/Ys2aNTh69CiAe6EbANq3b48xY8Zg1KhRaNiwoRRDIiIiqjG27s+1+Nz6GRfxwsW1hvANAG/6RFSb8K0Xm7wLX20MtzqAEzmC0wK4RqPB5s2b8f3332Pnzp2GSib64B0cHIynn34aY8aMQceOHZ01DCIiohovKUVb+UkoqfO98+ehyC0Vvt2t2omlwrOv4sbtYhw6W8A12yQ5hwfw+Ph4fP/999i4cSNycnIA3AvdCoUCgwcPxpgxY9C/f394eHg4+vFERERUTr7aSFedcooKVTjy/UNQa/MMx6pr+AYAhbYAADB3WQa6tVcg464O+WodfBQyRIV5YVCMLyK4ppucxOEBvFevXhAEwRC6BUHAQw89hDFjxuCJJ56An5+fox9JREREZvgozBc9KypU4dj33aHWZBuOVefwDQBqr5JZ70KtiPjjBWXeO5OsQVy8Ch2j5Jg0zJ/1vMnhnLIERRRFREZGYvTo0Rg9ejSaNm3qjMcQERGRBaLCvHAm2fgmTJ1Oh2NreyK/BoVvALjm36TSc04laTBjYTrrepPDOTyAT5w4EWPGjEG3bt0cfWsiIiKywaAYX8TFq4y+J5PJ8O9ajbCxIANAzQjfAHA2uD2eOrEK4VlKKIrUUHsqcK1eBPZG9kaa373vX10o4t3lGexsSQ5V5TphknthJ0wiIvcw/fNbRksRhuSk4r2db+Bj1RXUl3nViPCd7+UDH22+yfcvBLXGhg4joQyINByLjiqpA07kCC5pxJOcnIyDBw/i5s2byM/Px4svvojAwEBXDIWIiKhaU6ZpsTQuC+ev3AvfOl0RZLKSCNAzeTcA4PU6NWO5qAiYDd8A0PL2ecyMn1+mbf3JJA2upmnZbIccQtJW9MePH0ePHj3QokULjB07Fm+88QbeffddpKenlznvq6++QoMGDRAVFQWt1rLSSURERHRPolKDsfOuY/z7aTh0Ro3CkmrAKCpU4eDqB3Bm52REZCaj+5W9rh2ohEQAgoXnGmtbv8WKeupE5kgWwLdt24bu3bvjwIEDEEXR8MeYMWPGoKCgAJcvX8a2bdukGiIREVG1sGV/Ll76+BZS0ovLHC8qVOHvtT1RWHAbty/vwM2fn4Si2PIOme4s38vH4vCtp29br3fJwnrqRJWRJICnpaVh5MiR0Gg0aNOmDXbs2IHcXNOfIn19fTF48GAAwI4dO6QYIhERkdvTz3ovXJeF8lNc+vCt/WezJQBEe9SWdoA2SqsTYtf1azo9W+myE1Na3T6PhjnXAQD5msrrqRNZQpIA/vnnnyMvLw9NmjTB/v370b9/f9Subf5/+tjYWIiiiGPHjkkxRCIiIrd26GwBXllQcdYbMB6+3anaydfdp+NCUGubrk0Mao1GuTfsen5s8i4AgI9c0pW7VI1J8l/Sr7/+CkEQ8Oqrr8Lf39+ia1q1agUAuHLlihNHRkRE5P4SlRrMWXob2qKK77l7+AaA0Lsp2NBhJDQe1pUB1HjIsbHDSIRnKe16fnj2VQBA8zBuwCTHkKQKytWrJf/hdunSxeJr9B0zVSrjdUuJiIiqM2WaFlv35yIpRYvs3GLka0TkFxRDXVhSOrBn8m6EZykRmHcbvpq7+FK8tzxCJ8iQK6+Lm961MEa5GVrdvXXe7ha+AeCJU9/jjUe/xJIHp2LywUWQW7BuXeMhx5IHp0IZEAlFkdqu5+vb1g+O8bXrPkR6kgTwoqKSj+Q6neVrp+7evQsAqFOnjlPGREREVBUlKjVYGpdttGZ3RGYyRiSsQ8vb583eQybq4JV/G+OuH0OWeG/joDuGbwAIKMhCw5zrOBMSjU9j38HwhHVoZeafQWJQa2wsVcdb7amw6/lqr1qIjpKzBCE5jCQBvGHDhlAqlbh8+TIeeOABi645fPgwACA8PNyZQyMiIqoyDp0twLvLM6AurFglrF3aSYtnfwHgRFEu7lSD8K0Xm7wLP3YaC2VAJD6LnXXvtwDZV6HQFkDtVQvX/JsgPrIPbvqFlrn2Wr0IRN25aPOzr9ePwMRh/nZ+B0T3SBLAY2JicOXKFaxfvx5PP/10pecXFhZi6dKlEAQBsbGxzh8gERGRiyUqNZi3LAMabcXwHZGZjMkHv4C8uNDi+8V418OCOi3wquoi3nDz8A3cW4etl+bXGD92GmvRtXsje6P3pd9sfnbHtx9nG3pyKEk2YT777LMAgC1btuD33383e25hYSHGjBmD5ORkCIKACRMmSDBCIiIi11oal200fAPAiIR1VoVvvUGKBjgW8IDbh2/g3jpsW6T5Nba5ioquc0f8a0BLm59NZIwkATw2NhZPPvkkRFHEoEGD8MYbbxiWmACAUqnEX3/9hU8++QRt27bF+vXrIQgCJk+ejLZt20oxRCIiIpdRpmmNrvkGSjZcVrbmGwBUuiJMz0lEUbn9Vr4ySX7Z7XRqr1p2XW9LFRXBR4GwD1+y67lExgiiqXaUDqbRaDB8+HD88ssvEATTvaj0w3nsscfwf//3f/Dw8JBieGSj1NRUhIWFAQBSUlLQuLH7z7IQEUlt8U+ZiIs3XvXrqROrKl0+odIVoU/WMWSIWkR61ML2up3gKateNat3N+9n8ZITU6xZRy/4KBD87fuo3ceyvWtE1pDs/065XI5t27Zh6dKlaNasWZl29KX/NG7cGF9//TU2bNjA8E1ERDVCkpkW55XVsC4dvgEgubgA+7RZjhxelRAf2cfue+irqCRWshxF0b0TGv28iOGbnEby30tNmDABEyZMwLlz53D06FGkp6ejuLgY9evXR6dOnXDfffeZnSEnIiKqbvLVpsv0mqthXT58AyXVTnrJ6zt0fNVJ+SoqTe5eRV1BAw9fH9Tv2hKNX3oM3i0iXD1MquZctjCsTZs2aNOmjaseT0REVGX4KEz/QtpUDWtT4bs6bLg0Rl+G0FGMVlHRAh23yzHJW8OqJ+RU1WuBGBERkRuKMtPi/Fq9iArHalr4BiqWIXSWU0kazFiYjkNnba+6QlQZBnAiIiIXG2SmxfneyN5lXtfE8A3YV4bQWupCEe8uz0Ci0rKmR0TWcmgAnzp1KtLS0hx5yzI2bNiAH3/80Wn3JyIicoWIEC90jDK+5KF8DeuRd0/XuPAN2F+G0OrnFYpYFpct6TOp5nBoAP/yyy/RrFkzvPzyy7h8+bJD7qnVarFu3Tq0b98eTz75JC5etL2VLBERUVU1aZg/PP75qdzhxjHM+e0tLN40DkvWP4MWt89DXzP4I98oeKGkWEFNCd8AcM2/ieTPPJmkwdU00xVqiGzl0AD+zDPPoLCwEN988w2ioqLQrVs3fP3117h586ZV99Fqtfjjjz/w/PPPIzg4GM888wzOnj2Lpk2bonfv3pXfgIiIyI18suYOXvz4Frpc3ocvfp6AKQcWIOzuNSiKC+EBEQIAfX2wNp51sNG/I+bWblZjwjfgmDKEttiyP9clz6XqzeGNeA4fPoxZs2Zh165dJQ/4p6RgWFgY7r//fnTq1AkNGjRAvXr1UK9ePRQUFCAzMxNZWVm4ePEijhw5goSEBBQWlrTcFUURQUFBmD17NiZPngxPz+rR0au6YCMeIiL7zFh4EycvFmLg2U0Ycm4jyhfiVemKUAgdAmTeLhlfVZAY1Bqfxc5yybPbR8rxxavBLnk2VV8OT7NdunTBb7/9hiNHjmDhwoXYtGkTNBoNrl27hpSUFGzatMns9aU/D3Tu3BkTJ07E008/jdq1azt6qERERC71yZo7OHmxEA8q95kM332yjqEQOvxWrzMCa2AI13jIsbHDSJc9P19jukY7ka2cNp18//3344cffkBOTg42b96MPXv2YP/+/UhOTjZ5jY+PDx544AHExMRgyJAhiI6OdtbwiIiIXG7HwTwAwFMn15gM3/oNlwOzTuBQ/a4Sj9C1NB5yLHlwKpQBkWWOK7wFBNf3wNW0IqePwUfOgnHkeE5fz+Hn54fRo0dj9OjRAIDbt28jNTUVt2/fRmZmJhQKBYKCghAUFIRmzZpxiQkREdUI324uaRff4cYx+Gjzy7xnrNTgxFqhko7P1RKDWmNjh5EVwjcAvDY6AGeSNbiapnL6OJqbqdFOZCvJ064+bBMREdVk63eXbO4bemZDmeM1tc63XkrdcCx74GXc9DP9geNMsgaDYnwRF+/8AD7YTI12Ilvx9ypEREQuUPjP6onAvHTDsZoevoGSet/mwjcAXErRmq2d7ijRUXI0CeEMODke13sQERE5mDJNi637c/H3mQLILl/FI+c3o1X6OdTS5gMQUOjpjfe86kAUBCiK1AAYvvUs6XiZr9FBmaZFPV8ZBAFwbD23f8bhLWDiMH/H35gIDOBEREQOk6jUYGlcNk4laRCRmYznjn2HJtnKCufJizXw1dyrL83wfY8lHS9vZhRh/PvO67yt8BYwd0IgWkU4d4adai4GcCIiIgc4dLYA7y7PgLpQRLu0k3jxr4Xw0lnWRbEIOhTh3jRuTQ3fgGUdL/PUtk9516klQFVg+vroKDkmDvNn+CanYgAnIiKyU6JSg3nLMqDRiojITLYqfAOAv8wbu+rdh39nncDztUJrbPgGHN/xsnYtAY0CPdEuUo7BMb5oEuJlWCJ0KUWLfI0OPnIZmod5Gd4ncjYGcCIiIjstjcuGRlsyqzoiYZ1V4VvPX+aNgzWsznd5iUGtzW7ADMlJRc/k3QjPUkJRpIbaU4Fr9SKwN7I30vyMf2ipJZdB7i3DqSQNklK0iArzwqAYX0x5IsBZ3wZRpRjAiYiI7KBM0+JUkgZASUBseft8pdeodEV48m4C/lMnCh29WOYOMN/xMiIzGSMS1hn9Zxt15yJ6X/oNF4JaY4ORuuEZ2cXIyC42vD6TrEFcvAodo+SYxKUm5CIsQ0hERGSHrfvvbabsmby70vP1Gy4vFOfjybsJOKXNrfSa6s5Ux0sAaJd2EjPj51f6wabl7fOYGT8f7dJOWvTMU0kazFiYjkNnK6+6QuRoDOBERER2SEq5t9wkPEtp9tzy1U6KICKhqGYH8MSg1vg09h2cCYmu8F5EZjImH/wC8mKNRfeSF2sw+eAiRGQmW3S+ulDEu8szkKi07P5EjsIlKERERHbIV+sMX+trehtjqtTg6FqNnDq+qqZYkCGlbjiSA1sgPrKP2TXfIxLWQV5caNX95cUaDE9Yh89iZ1l0vrpQxLK4bCyYHmzVc4jsIckM+Keffor09PTKTyQiInIzPop7P0rVngqj57DO9z2X60dhft/5+LHT2Eo3XFqynt6YVrfPo2HOdYvPP5mkwdU06zfOEtlKkgD++uuvIywsDMOGDcPWrVuh0+kqv4iIiMgNRIXdK1t3rV5EhfcZvsuypM43YNl6enNik3dZdf6W/TV7KRBJS7I14FqtFlu2bMHQoUMRGhqKN954A4mJiVI9noiIyCkGxdyrYrI3sneZ9xi+K7K0zndl6+krvT77qlXnX0rhDDhJR5IAfvr0aUybNg2BgYEQRRG3bt3Cp59+irZt26Jbt2749ttvoVKppBgKERGRw3l63Pv6rtzP8PW8vGSG71Iqq/Ndmrn19BZdr7Wuukm+hr+dJ+lIEsDbtm2LBQsW4Pr169i0aRMGDRoEDw8PiKKIQ4cOYeLEiQgJCcG4ceOwb98+KYZERERkl0SlBtM/v4Xx76eh8e1kzPr9Hby38w3U1eQYzvm4dhRaevgAYPg2V+fbmEJv4+vpLaX2qmXV+T5yFoYj6UhaBcXT0xNDhw7F0KFDkZ6ejtWrV2PlypU4d+4c8vLysHr1aqxevRqRkZEYN24cxo4di0aNatbucCIiqlr0bcuTUrRIy9AiO1eH4lKTpe3STppsPS+TybC1bjT2a7PRU15zOy+aq/NtirJuBCJvX7T5mZauNddrHsYW9CQdQRRF0dWDOHz4ML777jv83//9H+7evQsAEAQBMpkMffv2xXPPPYfBgwfDy4v/c1Q1qampCAsLAwCkpKSgceOaO7tDRNVLolKDpXHZhi6XxkRkJuP1Pe8bwrdKV4TD2rvoJa8v1TCrvKxa9fB1t+lWhW+gpArKezvfsPm5s/t/bPFyFwBYMTsETUKYM0gaVeL3LV26dMGSJUuQlpaG1atXo2HDhhBFEcXFxdi5cyeeeOIJhIaG4s0338TNmzddPVwiIqrmDp0twIyF6WbDN1BSp7p0+O6TdQwTc89jY8EtKYbpFhb0eMvq8A0AaX6NcSGotU3PtGatOQBER8kZvklSVSKAA8DVq1fx0UcfYc6cObh16xYEQQAAiKIIURSRkZGBTz75BJGRkfj8889dPFoiIqquEpUazFuWAXWh+V8Ql65TXb7ayVt5ScjRFTl9rFWdtUG4vA0dRkLrJbfqGmvXmiu8BUwc5m/lyIjs49IArlar8f3336N3796IjIzEe++9B6VSCVEUERUVhY8++gg3btzAb7/9hieffBIeHh4oKCjAzJkz8f3337ty6DbZsWMHBEEw/Jk3b55D7pueno733nsP3bp1Q0BAALy8vODv74/OnTvj9ddfh1KpdMhziIhqgqVx2dBoK1+dqa9TbazU4Bs+EfCT1exm04UyL6uCsDHKgEjEDZwBwceyDZmCjwLq/8zGzYbNLTpf4S1g7oRAtIqwLuQT2cslfzscPHgQK1aswE8//YTc3JLC96IoolatWhgxYgSef/55xMTEGM5v2LAh+vTpg+TkZIwYMQKnTp3C559/jmeeecYVw7dJXl4eXnjhBYff97fffsNTTz2FrKysMsfv3r2L48eP4/jx4/jyyy+xdOlSjB492uHPJyKqTpRp2kqXneiFZylZ59sErcwT33SbZtPSk/JutuqMRi8uwp13v4H6wAmT5ym6d0L9uS+gWafWWKDUYFlcNk6a+XcZHSXHxGH+DN/kEpIFcP367pUrV+LixZJdzfr9n506dcLzzz+PUaNGwc/Pz+Q9IiMj8dFHH+GRRx4x3MNdzJ49G1evXkWDBg2Qnp7ukHtevnwZQ4cORUFBSa3TIUOGYPTo0QgPD8eNGzfw888/Y9WqVSgoKMCzzz6LZs2aoXv37g55NhFRdbTVim6IusKcahm+xTJfCyj08MbtOsE40CQGD1w7gIhspdnrlf4R+KHzeIeEb6CkOomiU2uE/rwIhReuIGflZmhOJ0Gnyoesjg/k7aPgN24ovFtEGK5pFSHHgunBhgo2l1K0yNfo4COXoXmYFwbH+HLNN7mUJAF8wIAB+P3336HT6Qyhu27dunj66afx/PPPo1OnThbfq1mzZgCA/Px8p4zVGY4dO4ZFixZBLpdj/vz5mDBhgkPuu2DBAkP4fu211/Dxxx+XeX/IkCG47777MHXqVOh0Onz44YfYtm2bQ55NRFQdJVnYDbGoUIVR17Yhu5qFbwC4FNgSHz88x+h7u1sOQEhOKgac34xW6edQ659mNwVetXC+QRv80nqoYc337Ofqo2kjb/zw613sPmL7z+zBpTqNerdsisAPp1l8bUSIF6Y8UXPLP1LVJUkA//XXXw1fx8TE4Pnnn8fjjz8OhcL6Ivs+Pj7o0aOHYZNmVVdcXIwJEyaguLgYc+bMQfPmlq1Ls8Rff/0FoKRk46xZs4ye8/LLL2PevHnIzMzEwYMHHfZsIqLqKF9deTdEnU6HQ2tjUVh8r1NjdQnfQOX1s9P8GuPbri+ZPadhfQ883Lk2AOCdcYHIyL5l8dKe0lidhKorSTZhNmjQAK+99houXLiAvXv3YvTo0TaFbwBo1KgR4uPjsWfPHgeP0jk+//xznDhxAi1atMAbb9hez9SYwsJCAED9+vVNLt0RBMHwWwP9+UREZJyPovIfizKZDP6h3Qyvq1P4BoD4yD523yPIv+z83qRh/lB4WzdxxuokVJ1JMgOempoKT8+atxtcqVRi7ty5AIBvvvkGcrljN3q0bNkSZ8+exZ07d5CTk2MyhF++fNlwPhERmRYV5oUzyZXP1LbtuwiCzAOPZl/Hc7rqM0Nrb9lAvfJdJVtFyDF3QiDmLruNQgtW+bA6CVV3ksyA18TwDQAvvPAC8vPzMWrUKPTq1cvh9588eTKAks2s//nPf4ye89VXXyEzM7PM+UREdI8yTYvFP2Vi6me3sO+E8bXKOl3FpSlten8O3cMfolBWPQK4tfWzzSm9bhsoqa3+4285FoXv6Cg5FkxrgK5tazlkLERVUc1MxhJYu3Ytfv31V/j7+2PBggVOeUbfvn3xzjvvYP78+fjoo4+QlJSEUaNGITw8HGlpafj555+xcuVKAMDYsWMxbtw4q5+Rmppq9v20tDRbhk5EJAllmhY//HoXJy6ooSrQoajoXpWPhndT0TN5N8KzlGhfpEaUpwLX6kVgb2RvpPndW1JSVKjCoXW9EBQ5AC0emgcA6HDjGIae2YDAvHQIog4iAPfYmWScxkOOJQ9OdUjlkvLrtg+dLcC7yytvbAQA3l4Cnuznx5lvqvYkCeDvvfee1dcIggCFQoG6desiKioKnTt3NluisCrJzMzE9OnTAQAffvghGjRo4LRnffDBB4iNjcV//vMfbNq0CZs2bSrzfqdOnfDOO+9g+PDhNt0/LCzMEcMkIpJUolKDz9dlGq1qEpGZjBEJ6wxdLEuLunMRvS/9hgtBrbGhw0hcqhOMv9f2hLYgA9dPr0CjnBv4tlCEj9Z9KnFVRgcBce0ex5mQaLvv5e2FMuu29V1FLWlsBACFWhHvLs/AgmkNGMKpWpMkgM+bN8/uqiVeXl4YMmQI5s+f79BKIs4wc+ZMpKeno2vXrpg4caJTn3Xjxg2sWLHCZIWThIQErFy5Eq1bt0abNm2cOhYioqrg0NkCzFl6G1ojneDbpZ3E5IOLIC82v8675e3zeOGPdxGbcwpaTbbheJ/bF+BTjTZcAoAMIoadWY/kwBYWzYCH5Nz7zYGiSA31P785ONCiN158vVOZ4GxpV9HS1IUilsVlY8H0YKu/FyJ3IdkSFH39b0EQDF+bYuycwsJCbNiwAdu3b8fmzZvRu3dvu8bjiDKGK1aswLPPPlvmWHx8PFasWAEPDw8sWbIEMpnzltmfP38effv2xfXr11G/fn0sWLAAgwYNQnBwMDIzM/Hbb79h1qxZ2LZtG/bt24ctW7agZ8+eVj0jJSXF7PtpaWno0qWLPd8GEZHDJCo1JsN3RGYyJh/8AvLiyitCqXRFGJB1CPnVsM63MfJiDYYnrMNnscZL2gKW/eZAkR0N9bwXoejU2qquouWdTNLgapqWJQip2pJkE6ZOp4NSqUTXrl0hiiKGDRuGuLg4pKSkQK1WQ61WIyUlBXFxcRg6dChEUUTXrl2RnJyMrKws7N+/Hy+88AJkMhny8/MxYsQI3LlzR4qhW0Wj0WDSpEkAgKlTpyI6OtqpzxszZgyuX78OHx8fwz+jxo0bw8vLC8HBwRg9ejT+/vtvBAcHIycnB08//TQ0Guv+MmzcuLHZPyEhIU767oiIrLc0Ltto+AaAEQnrLA7f5TtczvRtUW3Dt16r2+fRMOe60ffapZ3EzPj5RsN3aeq/TuLG0KnI2/W3VV1Fjdli5/VEVZkkM+C5ubno168fLl++jPXr1xtdjxwaGorQ0FAMGTIEGzduxFNPPYV+/frh6NGj6N69O7p3747Bgwfj0UcfRU5ODr766ivMmWO8U5clzp83/5eIJcqHz02bNuHixYvw8vJCmzZt8OOPP1a45ty5c4avz5w5Yzina9euaNq0qcXPPnXqFI4ePQoAGDVqFFq3bm1yjFOmTMGsWbNw48YN/PrrrxgyZIjFzyEichfmZlxDclIrDY+A8fD9lk8EnpM7by9PVRKbvAs/dhpb5pg1vzkAADFfjVvPzUbO4/MA2L6P6JKFXUmJ3JEkAXzhwoW4ePEipkyZYtFmwOHDh2Py5Mn46quv8Nlnnxk2cfbv3x+jRo3C6tWrsWPHDrsCeKtWrWy+1hT97LJWq7Wo3fzGjRuxceNGACXLWawJ4KU/QNx3331mz+3cubPh68TERAZwIqqWzM249kzeXen1JsN3NZ/5Li08+2qFY5b+5qA0MV+NB3evwe4H37Z5LPmayruSErkrSZagrF+/HoIgYNiwYRZfow/q5at66MPjpUuXHDdAN1S6tnpRkYnft/5Dq733w6Sm1mQnourPWMUTvfAsZaXX7y7MrNHhGwAU2oIyry39zYExTVLPmlzSYgkfuSQRhcglJPmv+8qVKwBgVRlB/blXr5b9NN6kSRMAQE5OjoNG5zjPPvssRFE0+2fPnj2G8+fOnWs4Xn4zZ2VKz5bv37/f7Ll79+41eh0RUXWSrzY9Y6ooUld6/RBFA7zpEwEBNTN8A4Daq2zzG0t+c2BObPIum68t302TqDqRJIB7eZX8T3TmzBmLr9Gfq79WT9+NzN/f3zGDq6KUSiUEQYAgCIiNja3wfqdOnRAaWtIueOPGjdi92/hfksePH8eSJUsAAD4+Pk7pyElEVBX4KEz/SFN7Kiy6x/M+jfFXQJcaGb4B4Jp/kzKvLfnNgTnGlrRYqnw3TaLqRJIA3qFDB4iiiE8//dSiKhxqtRqffPIJBEFA+/bty7yXnJwMAAgKCnLKWN2FTCbDhx9+CAAoLi7Gv//9b0yZMgW//fYbTp48iT/++ANvvvkmYmJikJeXBwB46623qv0HFyKquaLMzJheqxdR4ZhKV4R3VckVjgfJvB05LLcSH9mnzGtLfnNgjr/MtjKE5btpElU3kgTw8ePHAwDOnj2LPn36IDEx0eS558+fR58+fXD27FkAwHPPPVfm/V27dkEQBHTs2NF5A3YTo0ePxieffAIvLy9otVp8+eWX6N+/Pzp16oTevXvjo48+Qn5+PgRBwPTp0/HOO++4eshERE4zyMyM6d7Isr0j9Bsu16jT8HR2grOH5hYSg1rjpl9omWOW/ubAlPqN6kDhbV3fDYW3UKabJlF1JMmOvDFjxmDTpk3YsmUL/vrrL7Rt2xadOnVC586dDTPZt2/fxrFjx3DixAnDdY8++ijGjBljeH337l2sX78eoijikUcekWLoVd7MmTMxaNAgLFu2DHv37sWlS5egUqng4+ODJk2a4KGHHsLzzz9fphIKEVF1okzTYuv+XNw5cRmjEn5FaMaVMh0a90b2RppfY1wIao2Wt89XqHZyuCgHhwvvoot3XRd/J66j8ZBjY4eRFY5fqxeBqDsXbb5vvX+1xNxnAvHu8gyoCyvviKnwFjB3QiDb0FO1J4iVtaV0kKKiIkydOhVLly6FKIomO1Hq35swYQIWL15cZg34nTt3DGvD//Wvf6F27dpSDJ3MSE1NRVhYSZ3XlJQUNG5cM9dNEpH0EpUaLI3Lxt1D50x2aNS7ENQaByJ6YOjR/2FAxl81vtpJaRoPOZY8OBVnQqIrvBeSk4r3dr5h873DDqyBd4sIJCo1WBaXjZNmOmNGR8kxcZg/wzfVCJIFcL2TJ09i2bJl2LVrV4VSgpGRkejduzcmTpxYaW1rqhoYwInIFQ6dLcC7yzPQ/OoJTD64CPLiytcaZwoe6JV9HCrtvXrhNT18Jwa1xsYOI6EMiDR5zsz4D2wqRajo3gmhPy8qc0z/24pLKVrka3TwkcvQPMwLg2N8ueabahTJA3hpGo0G2dnZAEqqmsjl/NTrbhjAiUhqiUoNpn+ejpBblzAz/gOb28tXxfCtg4BimQe8dOb7O1gix9sXZxt2QKOc6wjIvwNPXRGKZB644xOI5MAWiI/sU2HNtzH9aqdgxI/zIKgt35Ap+CjQ6OdFUHQy3qWZqKaTZA24vvTd6NGjMW7cOMNxuVyO4OBgKYZARETVxNK4bGi0osUdGt0lfAOADCIKBQ94wf4AfrdWPXzX9UXD67q1ZWgQ4IHwhiUzzUGZxaj9zyx0gwAPCABuZRYbmZkOR17/93HrudkQ8ysP4YKPAsHfvs/wTWSGJAF8//790Ol0mD17thSPIyKiakqZpsWpJI1VHRofyz7lFuFbT2HBchpLlG+q8/roADzYwceme9Xu8wAa/bwId979BuoDJ0yep+jeCfXnvsDwTVQJSQJ4gwYNcPPmTdagJiIiu2zdX7J+25oOjXPqNMP4nLPQoeqHb0cq31Tn2613bQ7gAKDo1BqhPy9C4YUryFm5GZrTSdCp8iGr4wN5+yj4jRsK7xYRdo6aqGaQJIB37NgRN2/exMWLF9GpUycpHklERNVQUkrJTLY1HRof8q6H7/za4mJRHsbXkPANAGeDyzayu3nH/mUtAODdsikCP5zmkHsR1VSSNOJ5/vnnIYqioSU6ERGRLfLVOgDmOzSqdEXIL7eJ8SHvem4VvtUe9hclaHvrdJnX2iKX1VwgonIkCeCPPfYYnnnmGezduxfjx483tEYnIiKyho+i5MeWqQ6N+g2X/bKPVwjh7sbeuByefbXM6+JiO29IRA4jyRKU1atXo3fv3khISMCqVauwefNmDBo0CB06dEC9evXg4eFh9vrS3TCJiKjmUaZpsXp7NhKVJRsUjXVoLFPtRAQezT6JPwL+5Yrh2s0RGzEV2oIyr3UicDVNy3rbRFWAJAH82WefLdP5MisrC2vWrLHoWkEQGMCJiGoofbfLU+U6KO6N7I3el34zvDZWanCUoqFk46yKyldBAYAt+3Mx5YkAF4yGiEqTJIADJS3mzb0mIiIqTd/tUl1Y8edFml9jXAhqjZa3z7tVnW8pZSv8Kxy7lKKteCIRSU6SAH7lyhUpHkNERNVEolKDecsyoNGanqzZ0GEkXvjjXQzIOsTwbUT9vNsVjuVrdC4YCRGVJ0kAb9KkSeUnERER/UPf7dKcS3WCEZtzCvkM30Y1y7qMhjnXy7Sb95FLUnuBiCrB/xOJiKhK0Xe7NKeoUIW/1/ZEvibbcIzhu6LY5F1lXjcP4wZMoqpAsjXgREREltB3uzSnSHMXxdp7JW0ju83C1ab/xu7k3QjPvoqGOTfgW1j5fSojAhAqPavqKl+KcHCMr4tGQkSlST4DnpSUhNmzZ6NPnz5o164dIiMjcenSpTLnnDlzBr/88gv27t0r9fCIiMjFkizYKKjwDUWXJ3+Hh7cvIrvNQnjHCUjza4wfO43Fxw/PwU3fELvHkRjUGum1g+2+jyuVLkUYHSVnCUKiKkKyGXCdTofXX38dX3zxBXQ6naEKiiAIKCwsLHPutWvX8Oijj8LT0xNXrlxBaGiosVsSEVEVpEzTYuv+XCSlaJGv1sFHIUNUmBcGxfgiwkgAPJCQjxVb7+LmnSIUFYnQWtgwppZfGB4alwCZrOJckrlOmZZIqxOCz2Jn4fU/3kVw3i277uVK+lKEggBMHObv2sEQkYFkAXzSpEn47rvvIIoiQkND8eCDD2LDhg1Gzx0wYACaNm0KpVKJDRs24JVXXpFqmEREZCNTNbsB4EyyBnHxKnSMkmPSMH+0ipDj14MqfLUhC3kFlZelLSpU4dTWZ9Cm72LU8gszHDcWvgHTnTItpVL4ATDe8MedXPMvKYLg6+POC2mIqh9JlqDs3r0b3377LQDg7bffhlKpxE8//WT2mscffxyiKOKPP/6QYohERGSHQ2cLMGNheqWbJ08laTDl01sY9noqPl6TaXH4/nttT+Skn8Dh/+uLgpyUSq+5Vi/C0qEbv/6f4Lo3srdd93G1+Mg+AICcPBEzFqbj0NmCSq4gIilIEsCXLVsGoGRm+4MPPqi09TwAdOnSBQBw9uxZp46NiIjso6/ZbaxhjjHFOuCuyrJ61PrwrS3IAADoigqQmbK/0uvsDc764Kpv+OOOEoNalylBqC4U8e7yDCQq7W9zT0T2kSSAHzx4EIIg4LnnnrP4msaNS0pJ3bx501nDIiIiB7CkZrctyodvoKTaSWjbpyu9Ns2vMfK9fGx6bp6XT5nguqHDSBTK3GvzosZDjo0dRlY4ri4UsSwuW/oBEVEZkgTw9PR0AEBERITF13h5lfxlV1RU5IwhERGRA1hSs9sWpsJ3eMcJFl0fkpMKH22+Tc+urc1Hw5zrhtfKgEh8020atG4SwjUecix5cCqUAZFG3z+ZpMHVNLakJ3IlSQJ47dq1AQC3b1dsi2tKamoqACAgIMApYyIiIvtZUrPbWvaGbwDombzbrjGUb2BzJiQaHz88G0r/CKvvlWfjTLwtEoNa49PYd3AmJNrseVuc8O+NiCwnSRWUZs2a4fjx4zh37hz69u1r0TU7duwAALRt29aZQyMiIjtYUrPbGo4I3wAQnqW0axzlG9gAJTPh8/vOR0hOKgac34xW6efgU5gHmViynl0UBBQLHtB6eOGOTyCSA1sgPrIPbvqFIiQnFT2Td6Obcj9qFTluI6TGQ45r9SJwzb+J4VmWuJSitbpcJBE5jiQBvF+/fjh27Bi++uorTJkyxWTZKL1z585h5cqVEAQBAwYMkGKIRERkg3y1ZZspLXVu9yt2h2/A/jrgpRvYlJfm1xjfdn3JqvvpmwS1uJ2IsLvX7Bpbael1gvHxw3Osvi4ppRDj30+rcNxYuUgicjxJlqBMnToVtWvXRnJyMiZPnmx2Xffvv/+Ofv36Qa1WIyAgABMmWP8XLxERScNH4dgfI237fgV57ZIulraGb8D+OuD6BjaOZu+4KtzPxnFWVrHmVJKGZQuJnEiSGfDg4GAsWbIEY8aMwbfffoudO3di4MCBhve/+OILiKKIAwcOIDExEaIoQiaTYeXKlahTp44UQyQiIhtEhXnhTLLlmzD1SzHCs5RQFKmh9lTgWr0I7I3sjTS/xvDwVKDr0/G4o9yFBs0ftXlc9jbQueNT3+ZrzXF0Yx99vXJn0JctXDCtAWfCiRxMEPU94SXw008/YdKkSbh79y4EoWJXLv1Q6tSpg1WrVmHYsGFSDY1slJqairCwkq50KSkphvKRRFQzKNO0RpcylBeRmYwRCevQ8vb5MsdVuiJcKMpDZ++6uBDUGhs6jDRZvcMaITmpeG/nG3bdw5Hj0etw4ximHFjgsPvN7v+xxeu+bRUdJceC6cFOfQZRTSNpAAeAO3fu4Ouvv8bWrVtx8uTJMstR2rZti8GDB+OVV15BgwYNpBwW2YgBnIimf37LbCnCdmknMfngIsiLy56j0hWhT9YxZIpafOvXFjHe9Qwl9Cqr4mGJmfEfVAj81tKP507tQLMz95Ux9QHEHolBrfFZ7CyH3c+cFbND0IQbM4kcRvIAXppOp0NmZiaKi4sREBBgqP1N7oMBnIgSlSXrhY2tK47ITMbM+A8gLy4sc1wfvjPEkioqnhBwMuBBKGQyaDzk+DT2HZMzz5UtYyn77PkVgr+1dBAgg+kflZXNlJv6AGKPyv4ZOdqw2DqY8gTLAhM5iksDOLk/BnAiAoDVv2Rj5bacCseNzUKXD98A8JZPBJ7zuff3h7HZXUtmkcuH4XZpJzHlz0/NBmhHMDVzb+oDiD0EHwWy587CrIRmlW6mdJT2kXJ88SqXoRA5iiRVUIiIqPpKVGqwbmfFxi4hOak2hW8AaHX7fJlulO3STmJm/PxKl3C0vH0eM+Pno13aSQDAndqBTg/fACAv1mDywUWIyEwuc3xEwjqHhm9F905o9PMi3De+JxZMa4DoKMs2Ryq8K+67ska+xrHlJolqOgZwIiKyy9K4bGi0FUNu+W6UloZvPX03yojMZEw++IXFSzhKh2F7O2JaQ16swfCEdYbXxj6AWOO6byOk1A1HUmBL7G7eD2snLkToz4ug6NQaANAqomRz5HezQzAstg6iwrxQt44MPgoBdWuXNNUZFlsHK2aHoHmYt13fm4+ccYHIkSQpQ6h3/vx5LFu2DPv378fly5eRm5sLnc78p2pBEMzWDSciItdRpmlNbsAs3Y3S2vAN3OtGacsssj4Me+iKrbrOXvqZ+5t+oXaH/8Tgdvix09h7B7KAMWnaCpshI0K8Kl2fbW25yPKah3GPFpEjSfaRdsGCBYiOjsaiRYtw/PhxZGdno7i4GKIoVvqHiIiqpq37Ky490dN3oyzU6awO30BJN0p7ZpFb3T4PX43p8TmLfua+9AcQW+g/gJS2xcw/b3MGxfjaNZbBdl5PRGVJMgP+66+/YubMmQBKZrQfeOABdO7cGQEBAZW2pScioqorKUVr8j1910dvmQzNPGoho6jkXEvCN1DS5dHeWWRFkfSdHPXBWf8BxFYKbcWxXzLzz9uciBAvdIySmy0XaUp0lJwlCIkcTJIAvnDhQgBAvXr1sGXLFnTv3l2KxxIRkZPlq00vIyzd9XGtfweMvXsGPbz8LQrfQEmXR3tnkeu4YAZcH5ztbTtvrM28PZshJw3zN1ku0hSFt4CJw/xtfiYRGSfJ9PPRo0chCALmzJnD8E1EVI34KEz/GNkb2bvM61V121kcvgEgPrKP3bPInqK0a8CBe8H5Wr0Iu+5jrM28PZshW0XIMXdCoMUVURTeAuZOCGQbeiInkCSA5+fnAwAeeughKR5HREQSiTKyOa+oUIW/1nTD38nbcCGotU33TQxqjZt+oXbPIruCPjiX/wBirfjIPhWO2bsZsmvbWhaVL4yOkmPBtAbo2rbiLDwR2U+SJSihoaG4fPkyCgsdVwuViIhcb1CML+LiVYbXRYUq/L22J7QFGbj894eY23ES1sg84aWzvJqVVuaJjR1GAii7jMVd6INzml9jXAhqbdMmUv0HkPIcsRlSX75QmabF1v25uJSiRb5GBx+5DM3DvDA4xpdrvomcTJIZ8EGDBgEADhw4IMXjiIhIIvrNfUDZ8K0n2PRj5t4SCXtnkaVWPjhv6DASGg/rlnBoPOSGDyClOXozpL584RevBmP52yH44tVgTHkigOGbSAKSBPCZM2ciICAAn332GW7evCnFI4mISCKThvnDE3kVwndkt1mYV1hk1ew3AHjptIaGNvpZZHdgLDgrAyKx5MGpFodwfUt7ZUBkmePcDElUvUgSwBs1aoTNmzejuLgY3bp1wy+//CLFY4mISAKNA7U49lNshfDdtem/7arhrW9Fb8ssstRMBWcAOBMSjU9j30FiJR8kEoNa49PYd3AmJLrMcW6GJKp+BFGCTje9evUCANy4cQMXL16EIAjw9/dHVFQUfHx8zA9QELB7t3SthMk6qampCAsLAwCkpKSgcWPLKxwQkftTqVRo3rw5bt26ZTgW2W0WwjtOwFMnVqH3pd9svvfu5v0MnSDbpZ3E5IOLLG5HL6XEoNbY2GGk0fBdXkhOKnom70Z49lUotAVQe9XCNf8miI/sY3TNd3SUHBOH+TN8E1UzkgRwmUwGQShZ02fp4wRBgCiKEAQBxcXSl5EiyzCAE9VcxsL3Z599hl3XHoe6UMTrf7xr1wbKpMCW+PjhOYbXEZnJGJ6wDq1snFV3BqV/BOb3nW/XPXrf7wO/2jJuhiSqQSSpgtKjRw9DACciourh22+/rRC+Z8yYgd+mXgPg+E6QyoBIfBY7q8wscniW0qWz4h6i7Y1x9NIzi/HOuEAHjIaI3IUkATw+Pl6KxxARkYReeeUVKJVKLFy4EC+/+iHOFYzEozNSUPjPnktndIIESjZm6pem2LvMxV6mxmgNe7pbEpF7kmQTJhERVU/9n3gfD4/bj7P5T+PydS3y1feWGTqjE2R5ri5TaMkYK2NPd0sick/8v56IiCyiUqnw7rvvGl6v/iUbH6/JBBThRs93RifI8lxdptCSMVbG3u6WROR+GMCJiKhS+g2X8+bNw9NPP41fD6qwcluO2WvsCcemOkEa46oyhdaM0RxHdLckIvfi8AD+2GOPYfjw4UhNTTX6fn5+Pvbt24d9+/aZvU9iYiICAgJQv359Rw+RiIisUL7aybp16/CfbyzrbOzITpCmWNvsxhGsHaMpju5uSUTuweEB/Oeff8bPP/+MnBzjMyNXrlxBbGysoTa4KcXFxcjOzkZ2drajh0hERBYyVmrw5Vc/hGcdy2a2HdUJsjKWNrsxJaVuGApllgVhW8dYHrtbEtVcklRBMUaC8uNERGQHU3W+zxWMBPK1Ft9HH44rq+FtTUMbY4yVKSzd7OZscHu0vXXaZBMcS+qMmxtjw/oeuJVZDEt+vLG7JVHN5vBGPPqmO6dPn0abNm0qvH/27Fm0b9++0gY7lp5HrsVGPETVk6nwPWPGDDw6I6VMtRNrlA/Hhd61cLVuE+wx0QnSFaztVqm3YnYICjQ6LIvLxskk07XJ2d2SiFw2A05ERFWTufANAEVFts/bpPk1xqYuY/HrF/cqp6zYmo30nTlAFSmHXbrOuKVKr+VeMD0YyjQttu7PZXdLIjKKAZyIqIY6kJCPFVvv4uadImiLRIgi4OEB7F8Vg7xs4+EbADw9BRTaEcKLioDFP2ViUIwv9p3Iw5od5qupSMEw652lhKJIDbWnAtfqRWBvZG+k+Zn/zZ63Jyqs5Y4I8cKUJwKcOGIicmcM4ERENcyvB1X4akMW8goqhuiiYiDsvmlI/KMkcEd2m4UUYRQSlRrDkomG9T1x+brla8DL04lAXLwKcfEqm+/hKBGZyRiRsA4tjaz7jrpzEb0v/YYLQa2xwczadBHA3bwqMn1PRG6BdcCJiGoQffMcY+FbL6TlcLTqtQCR3WYhvOMEnErSYMbCdBw6WwAAGDeorlTDdap2aScxM36+0fBdWsvb5zEzfj7apZ00+r62CHh3eQYSlabXfRMRlcYATkRUQ5hqnlOkzYdOV3YGN6TlcIR3nGB4rS4UDSGzewcf1JILTh+vM0VkJmPywS8gL7YsNMuLNZh8cBEiMpONvq8uFLEsLtuBIySi6owBnIiohvhqQ1aFY0WFKvz9QwyOrn+kQggvr3TIDKzr3j8+RiSsg7y40Kpr5MUaDE9YZ/L9k0kaXE2zfWkOEdUcTlsDPmvWLPj7+1c4Xrqxzvjx401ezwY8RESOcyAhv8Kyk6JCFf5e2xPaggxoCzJwfNNg/GvENrP3OZmkwV8J+UhJd9/ysCE5qZUuOzGl1e3zaJhz3WQ5wi37c7n5kogq5bQAvnnzZpPvCULJry5XrVrlrMcTEVEpK7beLfO6dPjWaxA1xKJ7fVfuXu6mZ/Juu66f+PeXUHsqjFZLuZTCGXAiqpxTAji7XBIRVS037xQZvq6fcRE7fx4KrTbPcGxEeH94N/030qy8lzsKz1LadX3Y3WtlXpeulrIfzwAItuv+RFT9OTyAX7lyxdG3JCIiOxUViYjITMaAk6vwwsW1yBXvzdS+5ROB5/LzgJ1vVFpyDwCKit17kkVRpHbKfVvePo+ITe/hc9mr+PfMXux0SUQmOTyAN2nSxNG3JCIiO7VNO4mn9i7AgIy/kFE+fPvcazSjL7m35MGpOBMSbfRenh4CCrXuG8LVngqn3VterEH/DZ9h4d1aGPdmV3RtW8tpzyIi9+Xe29iJiKhSF7afwsi9n1UavvUqK7nXsL5793C7Vi/CqfeXF2sw6Nha1gYnIpMYwImIqrlbc75GujYX2eK9tdumwreeuZJ74928Ec/eyN5Of0ar2+fhn5HK2uBEZJR7T2MQEZFZV/5MQui1c4Bnbayv2wFP3T2NGT7hZsO3nrGSe9FRcjQK8jK8DslJRc/k3QjPUhqtCmIvZ9w/za8xLgS1trkUoaVik3fhR7+xuJqmRZMQr8ovIKIagwGciKgau7xwIyL++bq9ly9OBTwAT5nlv/yMTd6FHzuNBQAovAVMHOaP1duzEZGZjBEJ64yG2NJVQSrb0GmKs++/ocNIzIyfb3EnTFuEZ18FwNrgRFQRl6AQEVVDKpUKffr0gfrC2TLHrQnfwL0QqfAWMHdCIO7m6XDnl4OYGT+/0hlk/YbOdmknrXpmu7STTr0/ACgDIrHkwanQeDivUolCWwAArA1ORBUwgBMRVTMqlQrNmzfH7t278fT5n5Cts67lemkKbQGio+RYMK0B6taW4bsPD2HSX19YPHNc2YbO8iIykzH5oPPuX9qZkGh8GvsOEoNaW32tJdReJRVQ8jU6p9yfiNwXAzgRUTWiD9+3bt0CAOQUF2KPJsv2G9bxwYLpwWgVIcfSuGwMOb4W8mLrAr25DZ3ljUhY59T7l6cMiMRnsbMwp/9H2N28H5ICWyKlbjiSAlsio4F9ZXWv+Zdc7yPnj1oiKotrwImIqony4RsAJrUfgGFpOTbfs7h5yfpqZZoW6ceSbd64aGxDZ3khOalOvb85aX6NDWvdS4/nvZ1v2HQ/AIiP7AMAaB7GDZhEVBY/lhMRVQPGwvenn36KN75eaNd9I6ePAABs3Z+Lnsm77bpXbPIus+87+/7WSvNrjDtR7Wy6NjGoteHDwOAYX0cOi4iqAQZwIiI3Zyp8v/rqq2j6UBSuh7ex6b6p4W0R0b05ACApRYvwLKVd49Rv6DT5vpPvb4vfuz0Dwce6zpkaDzk2dhgJoKRsI0sQElF5DOBERG7MXPjWC37vRRR6WBcCNR5y/PHQKEMnx3y1DooitV1j1VcFMfm+k+9vC2VAMwR/+z5EhWUhXOMhx5IHp0IZEGko20hEVB4DOBGRGxs0aJDZ8A0AjeUF8JKJFt9TK/PCkgenYr+2CWYsTMehswXwUcig9rRuJrg8fVUQk+87+f628JHLULvPA2i8ZREKozuaPTcxqDU+jX0HZ0KiDWUbW0U4r8whEbkvh27C7NWrlyNvBwAQBAG7d9u3LpCIqLpav349oqKikJ2dbTR8q0+cx63xsyBoi0zcoSIRgEpesm5ZXSji3eUZeKBdSQfKqDsXbR5rw24t0b65N05fMl7lxN7766uOOJJ+A6WiU2u0/v1LJP5xEec/2YBaystQaAug9qqFa/5NEB/Zx7DmOzpKjonD/Bm+icgkhwbw+Ph4CIIAUTQ90yIIQpnX+nMtPU5ERPcEBgYiKSkJW7duxbhx4yq8f2fe1xALrOv26K3TYnjCOnwWOwtASQi/mVGEC5G90fvSbzaPtc0bjyNvvemfD3vtvL++6ogjld9A2apXC7Tq9TaUaVps3Z+LSyla5Gt0CJLL8GCYFwbH+HLNNxFVyqEBvEePHmYD840bN5CUlASgJFhHREQgODgYAHDr1i0olUqIoghBEBAVFYVGjRo5cniSiY2Nxd69ey0619yHFVvpdDp0794df//9t1OfQ0TSU6lUuHr1Ktq2bWs4FhgYaDR8F164AvVfJ216TvmyfonXtGjZvhkuHG9tU6nAO1Ht4OEbipt3bpo8J82vMS4E2Xb/0lVHHMXcBsqIEC+2lycimzl8BtyUHTt2YNSoUfDz88M777yDcePGITAwsMw5GRkZWLFiBf7zn//g9u3bWLhwIf797387cog1wtdff10mfBNR9aDfcJmVlYVDhw4hOjra7Pk5Kzfb9bzY5F1lamOH1PfAls5PY+rvH1jcqRL4Z2Ni+BNQvp9W6bkbOozEzPj5Vt9fX3XEUbiBkoicSZJGPBcvXsQTTzwBT09PHDhwoMzMTWmBgYF47bXXMHDgQHTv3h1PPvkkjh49ihYtWkgxTIf717/+hRUrVkj6zOvXr+Ptt9+GIAioX78+MjIyJH0+ETlH+Won3bp1Q05ODjw9Tf81rkmwfT01ULGs3527Oox7syu+U7+C8fstaxdfuiqIJZQBkVjy4FRMPrjIKfe3BDdQEpGzSRLAP/vsM+Tl5WH+/Pkmw3dpbdq0weuvv4533nkHn376KZYtWybBKB2vdu3aaNfOtiYOtnr55ZeRm5uL8ePHIzk52eKlMERUdRkrNfj++++bDd8AoMvLt+u55cv65Wt06Nq2Fup+0g8/fOaPB3avQSszy0USg1pjY4eRVofjMyHR+DT2HQxPWOeQ+9epJUBVYNkyPG6gJCIpSBLAf//9dwiCYFWVlIcffhgAsGuXYzubVWebNm3Czz//jMDAQHz88ccYPny4q4dERHaypM63KbLaPnY9u3xZPx/5vcq1R2UROBA7CyE5qeiZvBvh2VdNVgWxhTIgEp8Zub9QxweX6jTB7ojehvsLADw8AE8PAXJvAX61ZfCv44HmpTZFlt40ma0qRr5aBCDCR1HxXCIiZ5MkgKelVb7urzz9Zs6bN01v2KF7cnJyMGXKFADAJ598gvr167t4RERkL3vCNwDIO7SA+vBpm59fvqyfviTf0rhsaLQlM8ppfo3LrBN3tPL396stQ4MADzwYKcegGF9EWBiYuWmSiKoSSRrx+Pv7A4BVyyH0Gzrr1q3rhBFVP2+++SZu3LiBHj164Nlnn3X1cIjITvaGbwDwe3aIXWMoX9ZvcIwvlGlanEqyrqyhI+Xk6XApRYu4eBXGv5+G6Z/fMnTrJCJyF5IE8JiYGIiiiP/+97+4eLHyTUEXL17ERx99BEEQ8NBDD0kwQudITExE165d4e/vD4VCgcaNG2PIkCFYvXo1tFqtw55z8OBBLFmyBF5eXvjmm28cdl8icg1HhG8A8G7ZFIpu0TaNoXxZP31Jvq37c226n7OcStIYunUSEbkLSQL4jBkzIJPJcPfuXTzwwANYuHAhMjMzK5yXlZWFL774At26dUN2djYEQbD6B05VcuvWLRw+fBh3796FRqPB9evXsWXLFowdOxbR0dE4f976WrflabVaTJgwAaIoYubMmWjTpo0DRn5Pamqq2T+2LC8iIvMUCgUCAu4tl7AlfOvVn/ciBB/rWryXL+tXuiRfUorjJg8cRd+tkzPhROQuBFGiDi2ff/45Xn31VcPabkEQ0LRpUzRo0ACCIODWrVu4cuUKRFE0NI359NNPMWPGDCmG51C9evWCTCbDgAED0LFjR9SvXx+5ubk4fvw4li5dagjewcHBOHz4MMLDw21+1gcffIDZs2ejadOmOHv2LGrVurdpqnRDIFv/NVvTiTQlJQWNGze26TlEVFZRURHuv/9+PPPMM3ZPROTt+hu3npsNMV9d6bn6sn5nQqIB3CvJ17Vtyd8tz89Pw+XrVS+EAyWz9AumB7t6GERElZIsgANAXFwcpkyZghs3btwbwD8Br/QwQkJCsHjxYjz22GNSDc2hsrOzDevey9PPWK9atQoAMGzYMGzatMmm5yQlJaFDhw5Qq9XYvn07BgwYUOZ9BnAi0lOfOI87734D9YETJs8pX9bPWEm+qZ/dwpnkqjvTvGJ2CCuZEFGVJ2kAB0oC6ObNm7Fr1y6cPn3asBSlXr16aN++Pfr06YOhQ4fCy8u5f4FaEyxNWbFihU0bHouKitCuXTtcuHABQMkyj9BQ68t19erVC3v27MHw4cOxYcOGCu87IoCnpqaafT8tLQ1dunQBwABOZCuVSoXo6Gi8//77GDnSsR0dyyu8cAU5KzdDczoJOlU+Cr1rQekXjv1RfXHdtxF85DKzJfkW/5SJuHiVU8doj2GxdVjthIiqPEnKEJbm5eWFESNGYMSIEVI/usrw9PTEc889h9dffx1ASXWYp59+2qp7fPfdd9izZw98fX3xxRdfOGOYAMBATeRkpTdcjho1CgCcGsK9WzZF4IfTyhyLBNDbwusHxfhW6QB+qQquUSciKk/yAF5VOGIDZEhIiM3Xlt4sef36dauv/+ijjwAAPXv2xP79+42ek56ebvj6xx9/BFDSnXPQoEFWP4+IHK98tRNRFKv8xuaIEC90jJK7tBShOfkanauHQERUKZcFcJ1Oh8zMTOTn5yM0NBQeHh6SPr9Vq1aSPq88e5fAaDQlP/y2bduGbdu2VXq+fkatSZMmDOBEVYCpUoPusPF80jB/zFiYDnWhpCsYLVK6WycRUVUl6d9UxcXF+PbbbxETEwMfHx8EBwejWbNmhrXQetu2bcPrr7+O+fPnSzk8SZ07d87wdaNGjVw4EiKSmqPqfLtKqwg55k4IhMLb/r00jqbv1klEVJVJNgOenp6OoUOH4tChQ5VuCIyIiMDgwYMhCAIGDhyI6OhoaQYpkaKiInz33XeG1z169LD6HkqlstJzHLEJk4gcy93Dt17XtrWwYFoDLIvLxkkzy1FkAqCT8K+fwTG+0j2MiMhGkgTw4uJiDBo0CEeOHIFMJsPjjz+OHj164OWXXzZ6frt27dC1a1ccPnwYcXFxbhXA9+zZg06dOlVahlC/Bn3QoEEICwurcJ5SqUTTpk0BlKzzjo+Pd9aQiUgi1SV867WKKKm7rUzTYuv+XFxK0SJfo6tQScXc+wt/zHTYenJ9t04ioqpOkgC+atUqHDlyBF5eXtiyZQv69+8PACYDOAAMHjwYhw4dwp9//inFEB1m1apVGDx4MAYPHozY2Fi0bNkSfn5+UKlUOHbsGJYtW2ZYftKgQQOnVjAhoqrlo48+qjbhu7SIEC+zpf/Mve+o9eSlu3USEVV1kgTwdevWQRAETJo0yRC+K9OpUycAqLA+3B2oVCqsXbsWa9euNXlO+/bt8eOPPxpmuYmo+nv//feRmJiIDRs2VJvwbS/9evJ3l2fYHML13TpLNwwiIqrKJAngCQkJAEpmtS3VoEEDAMCdO3ecMiZneeONNxAdHY2DBw/i3LlzuH37NjIzMyGXyxEcHIx//etfGDFiBIYNGyZ55Rcicr3169fj+PHjuO+++1w9lCrD0vXkxhjr1klEVNVJ0glTLpejqKgIR48eNcxsA4BMJoMgCDh9+nSZutgAcOTIEXTt2hU+Pj5Qqapu04eaLjU11bCGnZ0wicpSqVRYvXo1XnzxRVcPxW2UXi+erSpGvlqHomKgWCdCpwO8PAQ0CPBAu0i5yW6dRERVnSQz4AEBAUhPT0dKSkqZAG5OUlISACAoKMiZQyMicorSGy6vXr1qaJ5F5lW2npyIqDqQpA5427ZtAZTMalvq//7v/yAIAu6//35nDYuIyCnKVzv5+OOPceXKFRePioiIqgpJZsCHDh2KP/74A19++SVmzJiBevXqmT1/w4YN2Lp1KwRBwPDhw6UYIhGRQ5gqNeiOG671y0GSUrTIV+vgo5AhKswLg2J8EcGlH0RENpNkDbhGo0HLli2RkpKC++67D6tWrUKbNm0qrAFPT0/HF198gU8++QTFxcVo164dTp48aXfbdnIergEnuqe61PlOVGqwNC7bbH3ujlFyTOLmRyIim0gSwAHg1KlTiI2Nxd27dyEIAlq2bInExEQIgoCOHTtCpVLh8uXLEEURoiiifv36OHjwIJo3by7F8MhGDOBEJapq+C68cAU5KzdDk3AR6uw8ZBbJca1eBA636YuC0PAKM9qHzhZYXBJQX/6va9tazv42iIiqFckCOABcunQJY8eOxcGDB+8N4J/Z7dLD6NKlC9auXYtmzZpJNTSyEQM4UdUM3+oT53Fn3tdQ/3XS5DkXglpjQ4eRUAZEomOUHP0fqI0vfsyCRmv5jwWFt4AF0xpwJpyIyAqSBnC9P//8E1u2bMHRo0eRnp6O4uJi1K9fH506dcLgwYPRt29fqYdENmIAJwLCw8ORkpJieO3q8J2362/cem42xHx1pedqPORY8uBUnAmJhkwAdJX8RAjJSUXP5N0Iz1JCUaSG2lOB6/UjkNLj36jfqRnXhxMRWcAlAZyqDwZwIuDDDz/E22+/DcD14Vt94jxuDJkCscDyhjYaDzk+jX0HyoBIk+dEZCZjRMI6tLx93uQ5+hn1ul3bmFwfbsvGTm4GJaLqhgGc7MIATlTiww8/hLe3t8vXfF8fMsXsshNTEoNa47PYWUbfa5d2EpMPLoK8uPJQr59Rv9SkU5n14bZs7ORmUCKqriQJ4DKZDDKZDAkJCRU6XpqSnJyMqKgoyGQyFBUVOXmEZCsGcKqJCgsL4e3t7ephVFB44QpSHhpj8/Wz+3+Mm36hZY5FZCZjZvwHkBcXWnwf/Yz6zYbNsWBaA9zN01m9sRMAN4MSUbUlSSMeoOwmSymuIyJyBpVKhSZNmqBfv36uHkoFOSs323V9bPKuCsdGJKyzKnwDgLxYg+EJ66AuFLFwXSbmLbMsSAOAulDE3GW3MWfJbauueXd5BhKVli+7ISJyJckCuK1YA5yIqgp9tZObN2/i999/x8CBA109pDI0CRftuj48+2qZ1yE5qWbXfJvT6vZ5NMy5jospWquqqgBAoRbQFlv3PHWhiGVx2dZdRETkIlU2gGdkZAAAateu7eKREBEZLzXYq1cvF46oIl1evl3XK7QFZV73TN5t1/2Mzag708kkDa6maSV9JhGRLSQN4JbOZufl5WHx4sUAgMhI07vyiYikUBXrfBsjq+1j1/Vqr7JrqMOzlHbdr/yMuhS27M+V/JlERNbydMZNTTXQ6devH7y8zJeM0mg0SE9Ph06ngyAIGDRokDOGSERkEXcJ3wAg79AC6sOnbb7+mn+TMq8VRZXXETen/Iy6FC6lcAaciKo+pwRwpVJZ4Zgoirh+/bpV93nggQfw+uuvO2hURETWcafwDQB+zw7B3f9ttPn6+Mg+ZV6rPRV2jaf8jLoU8jU6yZ9JRGQtpwTwsWPHlnm9atUqCIKAwYMHw9/f3+R1giBAoVAgJCQE3bp1Q69evbgJk4hcwt3CNwB4t2wKRbdom+uA60sQ6jtiXqsXgag7tm/sLD+jLgUfeZXd2kREZOCUAL5ixYoyr1etWgUAmD9/vsV1wImIXOn48eO4ffu24XVVD9969ee9iBtDp1rUhl5P4yHHxg4jAZTU1J76ZD1s2ZuLvTm90fvSbzaPpfyMuhSah7EzJhFVfZJMFcydOxdz5sxBgwYNpHgcEZHdevToga1bt8LT09NtwjcAKDq1RvC370PwsWz5iL5zpTIg0tDQ5pEH6+DrN0Pg174ZLgS1tmkcpWfUpTQ4xlfyZxIRWYut6Mku7IRJ1V1+fj58fOyrLuIK6hPncefdb6A+cMLkOYlBrbGxw0goAyIRHSXHxHIt3ROVGiycfRBTf//Aojb0eqU7YYYFeyJJoo2R0VFyLJgeLMmziIjswQBOdmEAp+pCpVJhzJgx+PHHH6tkm3lbFV64gpyVm6E5nQR1dh4yi+S45t8ER9r2RUGjcDQP88LgGF80CTG+dOPQ2QKsn7cL4/d/YVEI18+oX2rSCXMnBKJubRlmLEy3uKslAHh7ARCBwiKLL4HCW8CCaQ3KfIAgIqqqJAngf/31F2JiYuDt7Y1Lly4hNNT8ryWvX7+OyMhIFBUV4dChQ+jcubOzh0g2YgCn6qD0hssmTZrg4sWL1SqE2ytRqcHPXx1F++2r0MpMZ0z9jLp/1zZlZtMPnS3Au8sta0evXwYDwOpruraVvuoKEZEtnLIJs7wff/wRoiji0UcfrTR8A0BoaCgGDRqEjRs3Yu3atQzgROQ05audXL16FVu2bMGIESNcPLKqo1WEHG9+0h3KGV2w56ez8NvxCwJuXIGnpgAaLwXSg5viYpdHUL9TU8w1MpvetW0tLJjWAMvisnEyyfQsevllMLZcQ0TkDiQJ4H/++ScEQcC///1vi68ZOHAgNm7ciH379jlxZERUk5kqNcjwbVxEiBfGvRINvBJt9bWtIkrWZyvTtNi6PxeXUrTI1+jgI5eZXAZjyzVERO5AkgCenJwMAFaVIGzVqhUA4NKlS04ZExHVbO5Y57s6iAjxwpQnApx+DRFRVSZJGUK1uqQerUJheVc1ubzk14l5eXlOGRMR1VwM30RE5EqSBPCAgJKZi2vXrll8TWpqKgCY7ZxJRGQthm8iInI1SZagtGnTBunp6diyZQsGDx5s0TU///wzAKBly5ZOHBkR1TQPP/www7cZ+vXWSSlaZOcWI1+tAwQBPnIB/r4eiArzwqAYX0Rw7TURkc0kCeADBgzAnj17sHr1aowdOxYxMTFmz9+3bx/WrFkDQRDw6KOPSjFEIqoh1q5di44dO6KgoMBtw7ehtnfCRejy8lHoXQtKvybYH9UHqXVC4aOQWR2UE5UaLI3LxikTFUcyAaSmF+FMsgZx8Sp0jJJjkhOqj5T+AJCv1tn0vRARVXWS1AFXqVRo1qwZ7ty5Ax8fH3z44Yd4/vnnK6wJV6vVWLZsGd555x3k5eUhICAAly9fhp+fn7OHSDZiHXByR0lJSdi1axdeeOEFVw/FKuoT53Fn3tdQ/3XS5DkXglpjwz/dLQFYFJStqdNdmreXgHcnOqb+dmUfAADLvhciIncgWSfMXbt2YcCAASguLgYA1K5dG507d0ZISAgAIC0tDUePHkV+fj5EUYSnpye2b9+Ovn37SjE8shEDOFV1KpUK2dnZbv/fZt6uv3HrudkQ89WVnqvvRnkmJBqA+UY1iUoNpn+eDo3Wth8FMgGY+UwAHnmwjk3XA7Y16mHTHSJyZ5K2ot+zZw9Gjx6NGzdulDxcEMq8rx9KaGgo1qxZg9jYWKmGRjZiAKeqTL/hMi8vD2fPnkV4eLirh2QT9YnzuDFkCsSCylvB62k85Pg09h3DTLipVu3TP79ldtbZEjIBmP9ikE2h2JYPAGw7T0TuTpIqKHoPP/wwkpOTsXTpUgwaNAihoaGQy+WQy+UIDQ3F4MGDsXz5cly6dInhm4jsUrraiUqlQnR0NHQ6nauHZZM78762KnwDgLxYg+EJ6wyv1YUilsVllzlHmaa1O3wDgE4E5i67jUSl9fdaGpdt9ey7se+FiMidSLIJszS5XI4JEyZgwoQJUj+aiGoIY6UG33nnHchkks45OEThhStm13yb0+r2eTTMuY6bfqEAgJNJGlxN0xq6R27dn+uoYaJQCyyLy8aC6cEWX2PPB4Dy3wsRkTtxv59GRERmVLc63zkrN9t1fWzyrjKvt5QK3UkpWrvuXZ4+FFvK3g8AWxz4AYKISEoM4ERUbVS38A0AmoSLdl0fnn21zOtLpUJ3vtrxS3KsCcX2fgC45OAPEEREUmEAJ6JqoTqGbwDQ5eXbdb1CW1Dmdb7mXuj2UTj+R4A1odjeDwClvxciInfi0DXgzZo1A1BS3SQ5ObnCcVuUvxcRUXnVNXwDgKy2j13Xq73KVibxkd8L3VFhXjiTbP8mzNKsCcX2fgAo/b0QEbkThwZwpVIJoGJ5Qf1xW5S/FxGRMaU3WFaX8A0A8g4toD582ubrr/k3KfO6edi9TYuDYnwRF6+y+d7GWBOK7f0AUPp7IaKqTalUomnTpgCAFStW4Nlnn3XtgFzMoQF87NixVh0nInKEOnXq4OLFi2jfvj1efvnlahO+AcDv2SG4+7+NNl8fH9mnzOvBMb6GryNCvNAxSu6QUoR61oRiez8AlP5eiGqi0qHWHhK2hKF/ODSAr1ixwqrjRESOUqdOHVy5csXVw3A475ZNoegWbVMpwsSg1oYShAAQHSWvULZv0jB/zFiYbnUbelOsCcX2fAAw9r0QEbkLyeuAExHZS6VSoUuXLvj6669rRNOu+vNexI2hUy1qQ6+n8ZBjY4eRhtcKbwETh/lXOK9VhBxzJwRa3AreHFtCsS0fAEx9L0Q1TWhoKE6fNr1ErX379gCAf/3rXy6fDI2IiOBMeymStqKn6oet6ElqpTdcenh4YNeuXTUihOft+hu3npttUQjXeMix5MGpOBMSDaAksM6dEGi2VXyiUoNlcdk4aeNyFHvawx86W2DxBwBLvhciKqHfR9ezZ0/Ex8e7djBUBreQE5HbKF/tpLi4GKdOnXLxqKRRu88DaPTzIii6dzJ7XmJQa3wa+44hfEdHybFgWoNKA2urCDkWTA/Gd7NDMCy2DiJCPGHpFnh9KLYlfANA17a1sGBaA0RHmb/e0u+FiKiqc+gM+L59+xx1qzJ69OjhlPuS/TgDTlKpzqUGrVV44QpyVm6G5nQSdKp8FHrXgtIvHPuj+uK6byP4yGVoHuaFwTG+dq2TtmRWPDpKjonD/G0O3+Up07TYuj8Xl1K0yNfoHPa9EOn/20pK0SJfrYOPQoaoMC8MivFFRDX9b8vcDHhsbCz27t1reC8pKQlffPEFdu7cievXr6OgoABXrlxBREQEACAtLQ1xcXH4448/cOrUKdy4cQNFRUUIDAzEv/71Lzz99NN4/PHHy1SkKq2yKijz5s3Du+++C6BkU6harcbixYuxbt06JCUlAQBat26NMWPGYPLkyfD0dO9V1A4dfWxsrMPLBgqCgKKiIofek4jcC8N3Wd4tmyLww2lljkUC6O3g5+hnxaUMxREhXpjyRIBD70k1W6JSg6Vx2UY3+55J1iAuXoWOUXJMcuAHSXezefNmjBo1Cnl5eUbfLy4uRuPGjaHTVazzf+PGDWzZsgVbtmzBt99+i02bNqFOnTp2jefWrVt45JFHcPLkyTLHjxw5giNHjuC3337Dzz//bDLsuwOHf3zgknIiciSGb9djKCZ3Zen+glNJGsxYmF4j9xdcu3YNzzzzDHx8fDB79mzExMTAw8MDR44cMQRpfbbr1asX/v3vf6N9+/YICgpCbm4uLl++jOXLl+PgwYP4/fff8dJLL2HVqlV2jemxxx7DuXPnMHXqVAwaNAgBAQG4cOEC3n//fZw/fx5bt27F8uXLMWnSJLu/f1dxaADfs2ePyfcKCwsxa9YsHDlyBEFBQXjiiSfQpUsXBAcHAyj5tHPkyBH89NNPSE9Px/3334/58+fDy6t6/lqIiCrH8E1EtkpUajBvWQY0WssmBtWFIt5dnmHzZmJ3deXKFTRq1AgHDx5EeHi44XjXrl0NX3t4eODChQto3rx5het79uyJcePGYe7cuXjvvfewZs0azJo1C1FRUTaPST/LXXqD/X333Yf+/fujTZs2uHXrFr7++msGcL2ePXsaPS6KIgYMGICjR4/iueeew8KFC1G7du0K540ePRr//e9/MW3aNPzvf//DggUL8MsvvzhyiETkRl577TWGbyKyydK4bIvDt566UMSyuGwsmB7spFFVTf/973/LhO/yBEEwGr5LmzNnDr7++mtkZGRgy5Ytdv1dPWXKFKPVrQICAjBu3Dj897//xenTp3H37l3UrVvX5ue4kiSLZ7799lvs3LkTffr0wfLly42Gbz0fHx8sW7YMffv2xc6dO7Fs2TIphkhEVdBXX32Fbt26AWD4JiLLKdO0Nnd4PZmkwdU0rYNHVHV5e3vj8ccft+oanU6HGzdu4MKFCzhz5gzOnDmD8+fPGwox2FudatSoUSbf69y5M4CSyV13br4mSQBfuXIlBEHAiy++aPE1L730EkRRtHsdERG5L5lMhv3792P37t0M30Rksa37c+26foud17uTqKgoKBSKSs8TRRHff/89Hn74YdSpUwehoaFo1aoV2rdvb/ij3zSZkZFh15hatWpl8r2AgHv7UXJz3fffkyQ1XBITEwHA7K83ytOXttNfS0TVn0qlwi+//IInnnjCcEwmk6FXr14uHBURuZukFPtmsC/Zeb07qVevXqXnqNVqPPbYY9ixY4dF9ywoKLBrTD4+PibfK135pLi42K7nuJIkM+BqdUnntpSUFIuv0Z+r0dj2KyQici/6DZdPPvkkFi9e7OrhEJEby1dXLJdn1fUa+653Jx4eHpWeM3/+fEP47tmzJ3766SdcunQJKpUKxcXFEEURoigiJiYGACviWUKSAK5fuL9kyRKLr9GfGxkZ6ZQxEVHVUb7aySuvvIL09HQXj4qI3JWPwr544yN33/rSjiaKIv73v/8BAGJiYvDHH3/g8ccfR2RkJGrXrl1mRjozM9NVw3Q7kvwX9sQTT0AURezcuRMvvviiYUbcGI1Gg5dffhm//vorBEHAU089JcUQichFjJUa/OSTT9CgQQMXjoqI3FlUmH0ljJvbeX11kpmZiZs3bwKA2U6XKpUKFy5ckHJobk2SNeAzZszA999/j8TERCxduhQ///wznnjiCdx///1o0KABBEEw1AFfv3694V90y5YtMWPGDCmGSEQuwDrfROQMg2J8ERevsvn6wTG+DhyNeyvdjdxUp0wA+N///sfO5VaQJIArFArs2bMHAwcOxPHjx3Hz5k2Tazz164Y6deqEbdu2QS6vOcXwiWoShm8icpaIEC90jJLbVIowOkqOJiGcAdcLCgqCv78/srOzsW7dOkyfPr1CNjty5Ahmz57tohG6J8kWOQUHB+PQoUNYvHgx2rRpY1iwX/5P69atsWjRIhw+fBghISFSDY+IJMTwTUTONmmYPxTeglXXKLwFTBzm75wBuSmZTGaoy52QkICHHnoI69atw9GjRw0lYnv06AGFQoEWLVq4eLTuQ5IZcD0PDw+89NJLeOmll3Dz5k2cPn3asGC/Xr16aN++PUM3UTWn0+nQokULhm8icqpWEXLMnRCId5dnQF1YeVUOhbeAuRMCa1QbekvNnz8fBw4cwMmTJ3H06FE8/fTTZd4PCAjAxo0bMWfOHFy8eNFFo3Qvkgbw0ho2bIiGDRu66vFE5CIymQxDhw7FN998A4Dhm4icp2vbWlgwrQGWxWXjpJnlKNFRckwc5s/wbULdunVx4MABLFiwAD/99BOSkpLg6emJsLAwDBw4EK+88oqhCyZZRhBZrJHskJqaamialJKSwv8ByWIvvfQSmjVrxvBNRJJQpmmxdX8uLqVoka/RwUcuQ/MwLwyO8eWab5Kc5AFcp9Nhz549OHjwIG7evIn8/HzMnz+/zNKTwsJCFBUVwcPDg5swqzgGcLKETqczWbqKiIioppH0J+K2bdvQvHlz9OvXD3PnzsU333yDVatWISsrq8x5//vf/+Dr64sGDRqYLXlDRFWfSqVC48aNMW7cOFcPhYiIqEqQLIAvX74cQ4YMgVKphCiKqF+/vslWpc8//zzq1q0LlUqFuLg4qYZIRA6mr3aSlpaGlStX4vnnn3f1kIiIiFxOkgCelJSEl156CQDQq1cvnDt3zmybaW9vbwwfPhyiKOK3336TYohE5GDGSg22bt3ahSMiIiKqGiQJ4J9//jmKiorQtm1b/PLLL2jVqlWl18TExAAATpw44ezhEZGDsc43ERGRaZIE8D/++AOCIGDatGnw9va26JrmzZsDKNnYR0Tug+GbiIjIPEkCeGpqKgCgY8eOFl9Tu3ZtAEB+fr5TxkREjsfwTUREVDlJArgglLSCtSZM37lzB0BJ8XciqvoYvomIiCwjSQAPDQ0FAFy+fNnia/78808AQLNmzZwyJiJyrN9//53hm4iIyAKSBPDY2FiIoohVq1ZZdP7du3f/v707D6uq2v8H/j7MMiiCMiiIaFgOKM7TVTCn0ouilqQ5oKKipenNbmkZmpV6vZmWUzhbXzRHVMoZQq5iilPmPIChCAKKyjyt3x/8zu4g5xyGMzG8X89znufIXnuvtc/ZyGevvdZnYd26dZDJZHj99dd13Doi0oZhw4Zh/fr1MDIyYvBNRESkhl4C8KlTp0ImkyEqKgpbtmxRWzYtLQ1+fn5ISkqCiYkJgoKC9NFEItKCwMBApKSkMPgmIiJSQy8BePv27fHBBx9ACIFJkybB398fO3fulLafPn0aoaGheO+99/DKK6/g5MmTkMlkmD9/Ptzc3PTRRCKqoIyMDEycOBFFRUUlfm5nZ2egFhEREVUPMqFqOUotE0Lg/fffx9q1a6VJmarKAcCsWbOwfPlyfTSNNPDgwQO4uroCKE4Z6eLiYuAWkT4oTrjs2LEjzp49CyMjvS2sS0REVK3p7S+mTCbD6tWrceTIEfj4+EAmk0EIUeIFAN27d8cvv/zC4Juoino528n58+dx7NgxA7eKiIio+jDRd4X9+/dH//798eLFC1y8eBGPHz9GYWEh7O3t4eXlhQYNGui7SURUTqpSDQ4cONCArSIiIqpe9BKAT5w4EQDw5ptv4u233wYA2NjYoHfv3vqonoi0gHm+iYiItEMvAbg8/aC/v78+qiMiLWPwTUREpD16GQPesGFDAICjo6M+qiMiLWLwTUREpF16CcBbtWoFALh//74+qiMiLerevTuDbyIiIi3SSwA+ZsyYCq2EWZNkZmZi9erV6Nu3Lxo3bgxzc3M4OjqiQ4cOmDFjBo4ePaqVek6fPo0xY8bAzc0NFhYWcHJywsCBA7F9+3atHJ9qr5CQEJiYFI9WY/BNRESkOb3kARdCYMCAAYiIiMD8+fMRHBysNhd4TREZGYkJEyao7flv164dLl26pFE9CxYswKJFi0otiCI3ePBg7N69GxYWFhrVowzzgNcOMTExiI2NxYwZMwzdFCIi+v/i4+Ph7u6u8XH0tCQMKdBLAH7y5ElkZ2fj448/xpUrV9CiRQv4+/ujbdu2qF+/PoyNjdXuXx2zpRw/fhy+vr7IycmBra0tgoKC4OPjAwcHB2RlZeH69esIDw9HcnIyYmJiKl3PDz/8gKCgIABA8+bNMW/ePHh6eiIxMRErV65EZGQkAGDUqFEIDQ3VyrkpYgBe82RkZKCoqAh169Y1dFOIiEiN6hqA+/j4ICoqCt7e3vjtt9/0WndVoZcA3MjIqNI93jKZDAUFBVpukW6lpKSgZcuWSEtLg5eXFw4fPqxyAmpeXh7MzMwqVc+TJ0/QrFkzPHv2DE2aNMH58+dL5FEvLCzEsGHDcPDgQQDFPfI+Pj6VqksVBuA1i3zCZVFREe7cucMgnIioCsvPz8fNmzdVbvf09AQAdOrUCZs3b1ZZrk2bNlpvmzoMwPW4EE9terwxd+5cpKWlwdLSEmFhYWqzv1Q2+AaADRs24NmzZwCApUuXllrEyNjYGGvWrMGvv/6KwsJCLFu2TOsBONUcL2c7ad26NRISEgzcKipL/KN8HIx+gdsJ+cjKKYKlhRE8XE3h28sGTZ1NDd08ItIhU1PTcgXPVlZWeg+yST29BODyYRC1wdOnT6WhHvJJkboSFhYGAKhbty6GDx+utIyLiwv69euHI0eO4MSJE3jx4gVsbGx01iaqnpSlGpw1a5bhGkRluhGfix/2pePy7dxS2/68m4t9v2WgnYc5pg6zxWtNzQ3QQiIiUkUvAbi3t7c+qqkSwsPDkZ2dDQAYMmSI9POsrCwkJibC2toajo6OGk9CzcvLw9mzZwEUp4lT15Pu7e2NI0eOIDc3F7GxsejTp49GdVPNwjzf1c/vV7OxcH0qcvLUP1m8fDsX/1rxGMGTG6Br6zp6ah1R1ZR3Mw7Pt+xH7h+3UJSZBSMrS5i3bYG6AUNh9qrm46irqwsXLiAkJASRkZF4+PAhhBBwcXHB66+/jtmzZ6NFixYq901PT8fq1asRHh6OGzduICMjA7a2tmjYsCFeffVVDBgwAMOHD5dGAgQEBJTIiBcVFVUqHnJzc0N8fLxOzrUq0dsQlNrizJkz0ntPT0+cO3cOn376KU6cOCFlKWnYsCFGjhyJ+fPnV3pxolu3bqGwsBAA8Nprr6ktq7j9+vXrDMBJwuC7+rkRn4sFIanIzS/fsL6cPIGF61OxfJYDe8KpVsq5eB1pC9Yg5/Sl0tvOXsGzDXtg0cML9gumw6J9S/030ECKioowZ84crFixotQw4Vu3buHWrVvYsGEDVq9ejSlTppTa//r16+jXrx8SExNL/Dw1NRWpqam4fv06wsLCUFhYiPfff1+n51Id6TQA/+WXX3D48GHcv38fhYWFaNSoEXx8fDBy5EiYmtbMsYnXrl2T3kdGRiIwMLDUJNKUlBSsXr0ae/bsweHDh9GuXbsK1/PgwQPpfVkTH+WTJAFUeEyvYj3KPHr0qELHo6qDwXf19MO+9HIH33I5eQIh+9KxfDZXI6baJfP4GSRPmg+RlaO2XM7pS0j0mwnHjYtg1a+bnlpnWDNmzMCaNWsAFGebCwgIQLNmzWBpaYnLly9jxYoVuHr1KqZOnQonJ6cST/UBYOzYsUhMTISpqSkmT56MN998E05OTigqKsKDBw9w5swZ7Nu3r8Q+X331FebMmYMJEyYgNjZW6eRQTebGVSc6CcCTk5Ph5+cnDZFQtGnTJnz++ecICwuTZufWJE+ePJHeBwUFQSaT4csvv8S4cePg6OiIO3fuYNmyZdiyZQuSkpLg5+eHy5cvVzjbxIsXL6T31tbWastaWVlJ7zMyMipUj2LwTjUHg+/qKf5RvtIx3+Vx6XYu7j/KhxsnZlItkXPxOpInfgaRXb7fGZGVg+RJ89Eo7Lsa3xN+7NgxKfjesGEDJk2aVGJ7586dMWbMGAwePBgRERGYOXMmBg0aJC3Kdu/ePZw/fx4AsHz58lI93F26dMHw4cOxdOlSpKenSz9v3LgxGjduLMUltXlyqNZXwiwsLMSQIUPw+++/Qwih9BUXF4eBAwciNTVV29UbXGZmpvQ+JycHGzduxKeffgpXV1eYmZmhVatW2Lx5s/Q4Jz4+HmvXrq1wPTk5f9/Nl3W3aG7+92Nn+fh0qt0yMjJKXAsMvquHg9Evyi6kxgEN9yeqTtIWrCl38C0nsnKQtrDif5OrmyVLlgAARowYUSr4lrOwsMCqVasAAPfv3y+RUCMpKUl6r26tFplMhvr162ujyTWO1gPwnTt34ty5c5DJZHjllVewceNGXLlyBTdu3MCuXbvQrVvxo53k5GR888032q6+3GQymcavLVu2lDqu4mqTbdu2xdixY5XW//XXX0uB8c8//1zh9ivWk5eXp7Zsbu7f/wHVqVOxiVgJCQlqX8qeclDV5+TkhJs3b8Le3p7BdzVyOyFfo/3vaLg/UXWRdzNO6Zjv8sg5dRF5t+K12p6q5Pnz51Lu7bfeektt2ZYtW0opjhUXDXR2dpbeK4uFqGxaH4Kyc+dOAEDTpk1x9uxZ2NraSttatGgBPz8/9OvXD1FRUdi1axcWL16s7SYYlGKKvwEDBqgsZ29vj06dOuHUqVO4fPlyhRfkUaynrGElir3yZQ1XeRkX1qm5nJyc8PjxYxgZaf0+nHQkK6dIs/1zNdufqLp4vmW/ZvtvDkODxbO005gq5uLFi1JSiFGjRmHUqFHl2k+x19vd3R29evVCdHQ0vv32Wxw5cgQjRoyAj48PunXrBktLS520vSbRegB+8eJFyGQyfPjhhyWCbzljY2MsXLgQPj4+iIuLM1he6uvXr2t8DMU7QDlXV1cpE0pZ46fl24uKivDkyRM4OTmVu27FwLisiZKKEy85prt2ysjIgLe3N7Zt24bWrVtLP2fwrRu6SndmaaHZ92Vpzu+baofcP25ptv+V21pqSdXz+PHjSu2XlZVV4t/bt2/H22+/jZiYGFy7dg3Xrl3DokWLYGpqim7dumH06NEICAgo8cSe/qb1ADwlJQVA8bKnqihuS01NNUgAXlbqvspq3bo1du3aBQBSmkBVFLfLJzaUV4sWLWBsbIzCwkLcuHFDbVnF7S1b1uyJJVSa4oTLjh074vz58yWCcNIeXac783A1xZ93KzcJEwBeceUETKodijKzyi6kbv8MzfavyhRjjx9++AE9evQo134vj+Vu3LgxTp8+jRMnTmDv3r2IiorCtWvXkJ+fj+joaERHR+O///0vfv31V7W5xGsrrQfg2dnZkMlkaoc6KD6aUJxMWBMoTka4d++e2rJ3794FUDye287OrkL1mJmZoUuXLoiJiUFMTIzaISxRUVEAiidjqrsxoprn5Wwnubm5iIqKYgCuA/pId+bbywb7fqtYJiNFQ3pxFVyqHYysNBsCYWRdc4dQ2NvbS+8tLS01zkLSt29f9O3bFwCQlpaG48ePIyQkBBEREbh79y78/f1x8eJFjeqoiQz+PPLl5O/VXe/evdGwYUMAwMGDB1X2gsfFxeHSpUsAgJ49e1ZqKICfnx+A4gkVe/fuVVrmwYMHOH78OIDiXxIuQ197qEo1OH36dAO2qmaS0p2VEXzLydOd5Vys2FC4ps6maOdRucV0vDzMmYKQag3ztpr1uJp7emipJVWPl5eXtPrkqVOntHpse3t7+Pv748SJE1Le8EuXLuH27ZJDejRdDbwmMHgAXtMYGxtjzpw5AIrT9ixatKhUmYKCAkyfPl2aBBEUFFSqTHx8vJRtxcfHR2ldgYGBqFevHgDgk08+QVpaWonthYWFmD59unQT8NFHH1X6vKh6YZ5v/dJnurOpw2xhYVaxP14WZjJMGWZb4bqIqqu6AUM123+Cn3YaUgU1bNhQykgXGhoqDR3WNnmvOIBSaafl48IVs7TVNjpbCXPNmjVwcHDQSrnPP/9cW83Si5kzZ+Lnn3/GhQsXsHDhQty8eRPjx4+Hg4MD7t69i2+//VZK5zNo0CCMGDGiUvXY2dlh6dKlCAoKwv3799G1a1d8+umn8PT0RGJiIlasWCHl7Rw1apTKQJ5qFgbf+qWNdGdmLZqWe5/XmpojeHIDLFyfipy8sp8gWpjJEDy5AZehp1rF7FV3WPTwqtTvpkXP9hX6nayOPvvsMwwePBjPnz/HW2+9hf379ytNnAEUB8nyxXrkgbP8Cb6Xl5fSfYQQ0tN3mUyGpk2bltguT2Jx7949CCFqZY+4TGh5DIiRkZHWP8iyJjNWRY8ePYKvr6+0UpQygwYNwo4dO5QOC4mPj4e7e3G2BG9vbylnpzLBwcFYtGiRyuE8gwYNwp49e3QyE/nBgwdSZpWEhASmLTQwBt/6lzp3BZ5t2FPp/esFjqhUurMb8bkI2ZeOS2pWxvTyMMeUYbYMvqlWyrl4HYl+M8s9NAwAZJYWNWolTHk8piyOmDVrFlauXAmgOC1tUFAQ/vGPf8De3h6ZmZm4c+cOoqOjsXfvXjx9+hQvXryQ5vdt2bIFEyZMQOfOneHr64sOHTrAyckJ+fn5iIuLw+bNm3Hs2DEAwNChQxEWFlai7g0bNmDy5MlSO8aMGSM90Tc1NYWbm5uuPpIqQycBuDbJZLJqGYADxUNNNm7ciO3bt+PatWtIT0+Hvb09unTpgoCAAAwbNkzlvhUJwAHg9OnTWL16NaKjo5GcnAxbW1u0a9cOEyZMKHeOz8pgAF61jBw5UsrCAzD41oeHg6cj5+yVSu9v0bUtGoevrvT+8Y/ycTD6Be4k5CMrtwiW5kZ4xdUUQ3rZcMw31XrlnRwNFAfflZkcXZWpC8CFEFi0aBEWLVqEgoICtcexsrJCSkqKtJifPAAvS48ePXDgwIESEz+B4s6idu3aKU1W4ebmhvj4+DKPXd1pPQCXZ9zQJm9vb60fk7SDAXjVUlBQgFatWuH27dsMvvUkwScAeVfvVnp/s9avwPW3zVpsEREpyrl4HWkL1yLnlOpMHBY928M+eFqN6fmWUxeAy8XFxWHdunWIiIjAvXv38OzZM1haWsLV1RXt27fHgAEDMGzYsBJP63NzcxEREYFjx47h3LlzePjwIZKTk1FQUAAHBwd06NAB/v7+eOedd1R2zCYnJ2Px4sU4evQo7t+/L+UZZwBOVA4MwKuegoIC/PLLLxg6VLNJSFQ+hu4BJ6LykRbIunIbRRlZMLK2hLmnB+pO8KvxY76p6tHZJEwi0r2MjAz8/vvvJWabm5iYMPjWI/O2LTQKwGtyujOiqsTsVfcau7w8VT9MQ0hUTcknXPbv3x979lR+EiBphunOiIiootgDTlQNvZztxN/fH+np6WpXoK0t5JMSbyfkIyunCJYWRvBwNYVvLxs01cGkRKY7IyKiimIATlTNKEs1uHTp0loffN+Iz8UP+9JxWUlavj/v5mLfbxlo52GOqTpIy2e/YHql0p3ZB0/TajuIiKh64BAUomqEeb6V+/1qNv614rHS4FvR5du5+NeKx/j9arZW67do3xKOGxdBZlm+XPvydGc1LeMCERGVDwNwomqCwbdyN+JzsSCkfKtCAkBOnsDC9am4Ea/dJZCt+nUrXsCjZ3u15Sx6tkejsO9qVK5hIiKqGA5BIaoGGHyr9sO+dOTmVyybak6eQMi+dCyf7ajVtli0b4nGYd8x3RkREanFAJyoiisoKICHhweDbyXiH+WXOexElUu3c3H/Ub5OVotkujMiIlKHQ1CIqjgTExN06dJF+jeD778djH6h0f4HNNyfiIioMtgDTlQN7N+/H8OHD0fPnj0ZfCu4nZCv0f53NNyfiIioMhiAE1UTe/fuNXQTqpysnCLN9s/VbH8iIqLK4BAUoiomIyMDLi4u+Pzzzw3dlCrP0kKz/8IszflfIBER6R//+hBVIfJsJw8fPsSiRYsYhJfBw1WzCZSvaLg/ERFRZTAAJ6oilKUarFevngFbVPX59rLRaP8hGu5PRERUGQzAiaoA5vmunKbOpmjnUbll5b08zHWSgpCIiKgsDMCJDIzBt2amDrOFhZmsQvtYmMkwZZitbhpERERUBgbgRAbE4FtzrzU1R/DkBuUOwi3MZAie3ACvNa1czzkREZGmGIATGQiDb+3p2roOls9ygFcZw1G8PMyxfJYDurauo6eWERERlcY84EQGsmPHDgbfWvRaU3Msn+2I+Ef5OBj9AncS8pGVWwRLcyO84mqKIb1sOOabiIiqBAbgRAYSGBiIBw8e4IsvvsCyZcsYfGtJU2dTzBhpZ+hmEBERqSQTQghDN4KqrwcPHsDV1RUAkJCQABcXFwO3qPr566+/0KRJE0M3g4iIiPSEY8CJ9CQjIwMff/xxqZ8z+CYiIqpdOASFSA8UJ1xevXoV4eHhhm4SERERGQh7wIl07OVsJ7/88gt+//13A7eKiIiIDIUBOJEOqUo12LVrVwO2ioiIiAyJATiRjjDPNxERESnDAJxIBxh8ExERkSoMwIm0jME3ERERqcMAnEjLOnTowOCbiIiIVGIATqRlS5cuhUwmA8Dgm4iIiEpjHnAiLRs2bBiKiooM3QwiIiKqorgUPWmkoKAASUlJAAAnJyeYmPCejoiIiEgdBuBERERERHrEMeBERERERHrEAJyIiIiISI8YgBMRERER6REDcCIiIiIiPWIATkRERESkRwzAiYiIiIj0iAE4EREREZEeMQAnIiIiItIjBuBERERERHrEAJyIiIiISI8YgBMRERER6REDcCIiIiIiPWIATkRERESkRwzAiYiIiIj0iAE4EREREZEeMQAnIiIiItIjBuBERERERHrEAJyIiIiISI8YgBMRERER6REDcCIiIiIiPWIATkRERESkRwzAiYiIiIj0yMTQDSCqiIKCAiQlJRm6GURERFTLODk5wcREO6EzA3CqVpKSkuDq6mroZhAREVEtk5CQABcXF60ci0NQiIiIiIj0SCaEEIZuBFF5aTIE5dGjR+jSpQsA4OzZs3B2dtZm06ia4/VBqvDaIHV4fdQeHIJCtZaJiYlWHv84Oztr7TES1Ty8PkgVXhukDq8PKi8OQSEiIiIi0iMG4EREREREesQAnIiIiIhIjxiAExERERHpEQNwIiIiIiI9YgBORERERKRHDMCJiIiIiPSIC/EQEREREekRe8CJiIiIiPSIATgRERERkR4xACciIiIi0iMG4EREREREesQAnIiIiIhIjxiAExERERHpEQNwIiIiIiI9YgBORERERKRHDMCJiIiIiPSIATgRERERkR4xAKcaLzMzE6tXr0bfvn3RuHFjmJubw9HRER06dMCMGTNw9OhRrddZVFSE7t27QyaTSS+qGnx8fEp8L+pe2nT27FlMnz4dLVu2RN26dWFtbY3mzZtj8ODBWL58OVJSUrRaH2nPoUOHSlwXCxYs0MpxHz9+jC+++AI9evSAnZ0dTE1NYWtri44dO+Lf//434uPjtVIP6Zaurg8AOH36NMaMGQM3NzdYWFjAyckJAwcOxPbt27VWBxmIIKrBIiIihJubmwCg8tWuXTut1/v999+XqoeqBm9vb7XXg7a/s5ycHBEYGChkMpnauvbt26eV+ki7MjIySv0fEhwcrPFxjxw5IurXr6/2mqhTp47Ytm2b5idBOqOr60MIIYKDg4WRkZHK62Pw4MEiOztbK3WR/pnoIKYnqhKOHz8OX19f5OTkwNbWFkFBQfDx8YGDgwOysrJw/fp1hIeHIzk5Wav1Pnz4EPPmzYNMJoO9vT1SU1O1enzSjk6dOmHz5s06rSMvLw/Dhg3DoUOHAACvv/463n33Xbz22muwsLBAYmIiTp8+jd27d+u0HVR58+fPx/379+Hg4IDHjx9r5Zj37t2Dn58fsrOzAQBDhw7F2LFj0aRJEyQmJiIsLAxbt25FdnY2AgIC0KxZM/Ts2VMrdZN26eL6AIAffvgBCxcuBAA0b94c8+bNg6enJxITE7Fy5UpERkbil19+wcSJExEaGqq1ekmPDH0HQKQLjx8/Fvb29gKA8PLyEklJSSrL5ubmarVuPz8/AUBMnDixRG8rVQ3y78Tb21vndc2fP18AEDKZTKxdu1Zt2by8PJ23hyomNjZWGBsbC3Nzc7F+/Xqt9XC+99570rE++ugjpWW+++67Ej2dVPXo6vpIS0sT9erVEwBEkyZNREpKSontBQUFwtfXV6ovMjJSo/rIMDgGnGqkuXPnIi0tDZaWlggLC4Ojo6PKsmZmZlqrd+/evQgLC0ODBg3wn//8R2vHpern3r17WLJkCQBg+vTpCAoKUlve1NRUH82iciosLMTkyZNRWFiIefPm4ZVXXtHasU+fPg0AkMlk+Oyzz5SWef/992FnZwcAiImJ0VrdpB26vD42bNiAZ8+eAQCWLl2KBg0alNhubGyMNWvWwNjYGACwbNkyrdVN+sMAnGqcp0+fSo/k5JNX9OH58+eYMWMGgOL/EO3t7fVSL1VNISEhyM/Ph5GREebOnWvo5lAFffvtt7h48SJatGiBjz/+WKvHzsvLAwDY29ujbt26SsvIZDI0a9asRHmqOnR5fYSFhQEA6tati+HDhyst4+Lign79+gEATpw4gRcvXmi1DaR7DMCpxgkPD5fGVg4ZMkT6eVZWFu7cuYOkpCQIIbRe7yeffILExET07t0bAQEBWj8+VS+7du0CAHTo0AGNGzcGAAgh8OjRI9y7dw+ZmZmGbB6pER8fj+DgYADA2rVrYW5urtXjv/rqqwCAtLQ0PH/+XGW5e/fulShPVYMur4+8vDycPXsWANC9e3e1T2i9vb0BALm5uYiNjdVaG0g/GIBTjXPmzBnpvaenJ86dO4cBAwbAxsYGHh4ecHZ2hqOjI95//32tTcCMiYnBunXrYGpqirVr12rlmKRbN27cQNeuXWFrawsLCwu4uLhg6NCh2LZtG/Lz8zU6dkpKihQ8eXp6Ii8vDwsXLkSjRo3QqFEjNG/eHHXr1kWPHj2wd+9ebZwOadG0adOQlZWFd999F6+//rrWjy8fjiSEwNdff620zOrVq/HkyZMS5alq0OX1cevWLRQWFgIAXnvtNbVlFbdfv35dq+0g3WMATjXOtWvXpPeRkZHo0aMHjh07hqKiIunnKSkpWL16Nby8vHD58mWN6svPz8fkyZMhhMCcOXPQqlUrjY5H+pGcnIyzZ8/i2bNnyM3NxcOHD3HgwAGMHz8eXl5eGv1BU7wGLS0t4e3tjQULFiApKUn6eVFREWJiYjBixAhMnz5do3Mh7QkNDcXhw4dha2uL5cuX66SO/v3749NPPwVQPMZ3xIgR2Lt3L2JjY3Hw4EFMmjQJM2fOBACMHz8eEyZM0Ek7qOJ0fX08ePBAeu/i4qK2rKurq/Q+ISFB620h3WIATjWOvNcIKO45kslk+PLLL/HXX38hNzcXV69elYaIJCUlwc/PT+1j4LIsXboUV69ehbu7O+bPn69p80nHjIyM0LdvX3zzzTc4fvw4Ll68iJMnT2LFihVo2bIlgOIAuk+fPvjrr78qVYfiNbhx40acOXMGXbp0QUREBLKysqR5Co0aNQJQ/Bj7u+++0/zkSCNPnjzB7NmzAQCLFy+Gg4ODzur68ssvcezYMfTp0wd79+7FiBEj0LlzZwwZMgSbNm1Cu3btsHv3bmzZskWabEeGpY/rQ3Est7W1tdqyVlZW0vuMjAytt4V0zLBJWIi0r3nz5iUWK1C1kMWUKVOkMkuWLKlUXbdu3RIWFhYCgPjll19KbWcawqrn6dOnKrfl5eWJ8ePHS9/ZsGHDKlXHjz/+WOIabNOmjcjMzCxV7tatW8LKykoAEHZ2dkrLkP5MmDBBABBdu3YVhYWFJbZFRkZqdaGVhw8fitGjR0v/f7z8MjY2Fv/85z/F1atXNa6LtEMf18e2bduk42zcuFFt2bt370plJ02aVOk6yTDYA04GU97lwNW9tmzZUuq4FhYW0vu2bdti7NixSuv/+uuvpckzP//8c6XOYerUqcjJycGIESMwaNCgSh2DStPVtQEAtra2Kus1NTXFhg0bpElv+/btw8OHDyvcfsVrEAC++OILWFpalirn4eGBadOmASjuXTt+/HiF66ptdHVt/Pbbb9i8eTOMjY2xbt06GBnp7s/j9evX0aVLF4SGhsLKygpr1qxBQkIC8vLykJSUhG3btqFx48YIDw9H9+7dERUVpbO21DTV/fpQ/L+jrOw3ubm50vs6deropD2kOwzAqcaxsbGR3g8YMEBlOXt7e3Tq1AkAcPny5Qqn+tq0aRMiIyNhY2ODlStXVq6xVOWYmJhg0qRJ0r8rE/woXoMymUxKF6bMwIEDpffnzp2rcF2kudzcXEydOhUAMHPmTHh5eem0vnHjxuHhw4ewtLREdHQ0pk2bBhcXF5iamsLR0RFjx47FmTNn4OjoiOfPn2P06NElgi3SL31eH4r/d5Q1rEQxk1JZw1Wo6uFS9GQw2pi17ezsXOpnrq6uUiYUxUkqysi3FxUV4cmTJ3Bycip33UuXLgVQnAoqOjpaaRnFpYl37NgBoHjcnq+vb7nrqY10dW2Ul+JE2sr0gCted7a2tiX+qKorm5KSUuG6ahtdXBt79+7FrVu3YGpqilatWkm/q4oUJ9b++eefUpmuXbvC3d293HVfvnxZShn37rvvSvMOlLVxxowZ+Oyzz5CYmIjDhw9j6NCh5a6ntqru14fixEvFCZnKKE68LOtvHVU9DMDJYMpKsVRZrVu3lnIwy9M5qaK43cSkYr8O8h6p8PBwhIeHl1l+1KhRAAA3NzcG4GXQ1bVRXjKZTKP9PTw8YGpqivz8fJ1eg7WRLq4N+e+yPKNRWfbs2YM9e/YAADZv3lyhAEsxQOzQoYPash07dpTe37hxgwF4OVT366NFixYwNjZGYWEhbty4obas4nZVN3JUdXEICtU4vXv3lt7LczGrcvfuXQDF4+7kyz4TKfZmyTOVVISpqSm6d+8OoHiF1NTUVJVl5dcgAGnBHqq5FG+yCgoK1JZVzEfPm7PawczMDF26dAFQvL6EuqGR8uFx5ubm0nBKqj4YgFON07t3bzRs2BAAcPDgQZU9kHFxcbh06RIAoGfPnhWeVBMfHw8hhNqXfKUyANLP4uPjK3VepB8FBQXYtGmT9G/FG7qKGDFihPRevrS0MooL8fTq1atSdZFmAgICyvxdjoyMlMoHBwdLP6/oqreKvaGqhq7JKc4/qEgvKmmXPq8PAPDz8wNQfPOuaqGuBw8eSJO2+/btq3aYG1VNDMCpxjE2NsacOXMAAPfv38eiRYtKlSkoKMD06dOlxXmUrTQXHx8vzYr38fHRaZtJPyIjI5Genq5ye35+PgIDA6VhAr6+vkrHVpbn2pg4caKUJzg4OBiJiYmlykRFReHHH38EALRp0wY9e/as4BlRVVPWtdG+fXvpSceePXtw4sQJpce5cOEC1q1bB6B4MSddrMhJ+lee/zsCAwNRr149AMAnn3yCtLS0EtsLCwsxffp0qXPpo48+0mmbSTcYgFONNHPmTGl85cKFCzFq1CgcPnwYFy5cwK5du9C7d28cPnwYADBo0KASvZVUc23duhWurq549913sX79epw8eRKXLl3C//73P6xcuRJeXl7YunUrAMDBwUGj7DbW1tb47rvvIJPJkJiYiM6dO2PVqlU4d+4c/ve//+Gzzz7Dm2++icLCQpiYmGDdunUajz2nqs/IyAiLFy8GUBxIvfnmm5gxYwaOHj2KS5cuISIiAp988gl69eolZbmYO3eu2vSZVLPY2dlJk/zv37+Prl27YvPmzYiNjcWBAwfQv39/HDx4EEDx3CJ2EFVPHFRGNZKFhQXCw8Ph6+uL8+fPY8eOHUpnrg8aNAg7duxg4FOLZGRkIDQ0FKGhoSrLeHp6YseOHRo/9vf390dqair+9a9/ITExETNmzChVxtraGj/99BN7v2uRsWPHIjk5GfPmzUN+fj5WrVqFVatWlSonk8kwa9Ysadl6qj2mTp2KxMRELFq0CHfv3sXEiRNLlRk0aFCJ4XJUvTAApxrL2dkZZ86cwcaNG7F9+3Zcu3YN6enpsLe3R5cuXRAQEIBhw4YZupmkRx9//DG8vLwQExODa9euISUlBU+ePIG5uTkcHR3RqVMnvPXWWxg2bJjWlv9+77334OPjg9WrV+PYsWN4+PAhjI2N0axZM7zxxhuYNWuWRikTqXqaM2cOfH19ERISgqioKNy5cwcZGRmwtLSEm5sb/vGPfyAwMLBEJhSqXRYuXIiBAwdi9erViI6ORnJyMmxtbdGuXTtMmDBByqxF1ZNMCCEM3QgiIiIiotqCY8CJiIiIiPSIATgRERERkR4xACciIiIi0iMG4EREREREesQAnIiIiIhIjxiAExERERHpEQNwIiIiIiI9YgBORERERKRHDMCJiIiIiPSIATgRERERkR4xACciIiIi0iMG4EREREREesQAnIiIiIhIjxiAExERERHpEQNwIiIiIiI9YgBORERERKRHDMCJiIjKkJaWBjs7O8hkMpw7d87QzaFqIisrCw4ODpDJZPjtt98M3RyqQhiAE5HOxMfHQyaTafyqaRYsWFDhzyAsLMzQza7VPv/8czx9+hSDBg1C586dtXbcnJwc2NraQiaToWnTphBCVGj/0aNHS9fIxYsXNWrLlClTIJPJMHToUI2OQ3+ztLTEv/71LwDArFmzKvz9Us3FAJyIiEiN+/fvY/369QCKA3FtsrCwwNtvvy3VEx0dXe59X7x4Id2YtWnTBu3bt690O4QQCA8PBwD4+vpW+jiVFRAQIN2E1DTvvfce7OzscPnyZezatcvQzaEqwsTQDSCimqtx48a4cuWKyu2enp4AgE6dOmHz5s36alaVsmnTpnL1qLq5uemhNaTM0qVLkZ+fj549e6Jr165aP/64ceOwYcMGAMCPP/6I3r17l2u/PXv2IDs7WzqGJs6fP49Hjx5BJpNh8ODBGh2LSrKxscGUKVOwZMkSfPnllxg5cqShm0RVAANwItIZU1NTtGnTpsxyVlZW5SpXE7m7u9fac68O0tPTsW3bNgDAmDFjdFLHP/7xD7i7uyMuLg67du3CqlWrYG5uXuZ+P/74IwDA2NgY7777rkZtOHjwIIDim2FnZ2eNjkWljR49GkuWLMGVK1fw22+/wcfHx9BNIgPjEBQiIiIVduzYgczMTJiamkpDRbRNJpNh7NixAIBnz55JwbA6Dx48kCb19e3bF40aNdKoDfI6//nPf2p0HFLO09NTeuK3ceNGA7eGqgIG4ERUJfn4+EAmk0k9Rbdv38b7778PDw8PWFpaQiaTIT4+HgCwZcsWaSKa/GfKKE4K3bJli9r6w8LC8Pbbb6NJkyawsLCAra0tOnXqhIULF+Lp06faOUkNKDuXY8eOwdfXF05OTjA3N4e7uzumTZuGBw8elOuYkZGRGD9+PJo1awZLS0vUrVsXnp6e+Oijj5CYmKhyP8VJpUBxELlo0SK0b99emmD48uedlpaGf//733j11VdRp04dODo6on///ti3bx8A1d/pgQMHpJ/v2LGjzHP68MMPIZPJYGJiovYcVNm5cyeA4uvR3t6+zPI5OTlYtWoV+vbtCycnJ5iZmcHBwQH9+vXDxo0bUVBQoHQ/xSEk8p5tdf7v//4PRUVFpfatjIcPH0oTOCsz/ruoqAgRERGYM2cOevbsiQYNGsDU1BS2trbw8vLCnDlz8NdffyndV37tbN26FUDxOPjyTsSOj4/H7Nmz0bp1a9jY2MDS0hIeHh6YOnWq2qFvAKTjLliwAEDxte/n54dGjRqhTp06aNmyJRYtWoTMzMwS+/36668YNGiQVK5Vq1ZYvHgx8vLyyvycRowYAaD4/5acnJwyy1MNJ4iIDASAACC8vb1LbfP29pa2hYWFCSsrK6m8/BUXFyeEEGLz5s2lfqZMXFycVG7z5s1Kyzx58kS8/vrrpepSfDk4OIiYmJhKn3dwcLB0rMjIyEod4+Vz+eSTT1S2t2HDhuLatWsqj5WdnS3eeecdtedsZWUlDhw4UOb53Lp1SzRt2rTU/oqf9x9//CEcHR1V1jVlyhSV32lBQYFwdnYWAMTAgQPVfkb5+fnCwcFBABCDBw+u0OcrhBA5OTnC3NxcABDz588vs/ylS5eEm5ub2s+xc+fOIikpSen+PXr0EACEqampSE1NVVtX69atBQBhY2MjMjMzK3xuitatWycACBcXl0rtr/j9q3pZWlqKvXv3VmpfZaHK1q1bpe9G2cvY2Fh8/fXXKtssLxccHCwWL14sZDKZ0uP06NFDZGRkiKKiIjFz5kyV9b3xxhuioKBA7ed0+PBhqfzRo0cr/kFTjcIAnIgMpjwBuLu7u7C2thYNGzYUS5YsEadOnRJnzpwR33//vUhJSRFCaC8Az8nJER06dJD+gI8dO1Zs375dnDlzRkRHR4uvvvpK2NvbCwCifv36Ij4+vlLnre0AXB64eXt7i9DQUBEbGyuOHz8uxo0bJ5Xp1q2b0uMUFRWJwYMHS+V8fX3Fjz/+KE6dOiViYmLEypUrRZMmTQQAYWZmJs6dO6f2fNq2bStMTU3FjBkzxLFjx0RsbKzYvn27OH36tBBCiKdPn4pGjRpJ5ceOHSsOHTokYmNjxY4dO0T37t0FANG1a1eV3+ncuXMFAGFkZCQSEhJUfkZhYWHSMfbs2VPhzzg6OlraX9XNh9zt27dFvXr1BABRt25dMXfuXLFv3z4RGxsrjhw5It577z1hYmIinVteXl6pY8gDYQBi1apVKuu6cOGCVC4gIKDC5/Uy+fcfFBRUqf0//fRT4ezsLKZPny5dO+fPnxdhYWHi3//+t7C2thYAhIWFRakbweTkZHHlyhUxdOhQAUA0atRIXLlypdRLUXh4uBQwW1tbi+DgYBEdHS1iYmLEN998Ixo0aCB9PmvWrFHaZvn2Ll26CACie/fu0u/O4cOHxZtvvimV+fTTT8U333wjAIg333xT7NmzR5w/f17s379fdOvWTSq3du1atZ/TkydPShyTajcG4ERkMOUJwOV/lO/fv6/yONoKwOfNmycACFtbWxEbG6v0GPHx8VIP7OjRo8s6RaUUA9ZNmzYpDTgUXzdv3lR7LgDE5MmTRVFRUalygYGBUpkLFy6U2h4SEiL1uh46dEhpe588eSL1uPbs2VPt+RgZGYkjR46oPPdZs2ZJZVesWFFqe0FBgRSMqfpO79y5IwVgX331lcq6hgwZIoDiJwDKAt6yLF26VGqDukBfiL97r9u3by/dGL7s0KFDwsjISAAQISEhpbY/ffpU6tVVdcMkhBCzZ8+W2hUREVGxk3pJVlaWqFOnjgAgwsPDK3WMuLg4tZ9vQkKCaNy4sQAgxowZo7TM+PHjBQDh5uamtq68vDzpBs7a2lpcvHixVBnF31FLS0ul34fi9TVixIhSvdcFBQVScG1jYyMsLCzErFmzSh0nMzNTeurRtm1btW0XQgh3d3epx5xqNwbgRGQw5Q3At23bpvY42gjAX7x4IfVgfv/992rrW7NmjRS0ZmRkqC2rTHkfu8tfyoISxXNxdnYWOTk5Suu6ceOGVG7lypUlthUVFYnmzZsLAOLDDz9U2+Zff/1VOs6tW7dUns/EiRNVHiMnJ0fY2toKoHgohipJSUnCwsJC7Xfap08fAUB4eHioPIa8x3n27Nlqz02VGTNmSG3Izc1VWe7kyZNSuT/++EPtMUeOHCk9tVDmrbfeko51+/btUtsLCgqEk5OTACCaNGmi9KarIg4cOCAFqtnZ2RodS50VK1ZITweUtbm8AfjPP/8sfT5LlixRWe6nn36Syv3nP/8ptV2+zdLSUqSlpSk9xqZNm6Ryrq6uKm8yPv/8c6lcenq62vbLn+y0aNFCbTmq+TgJk4iqNDMzM51ln1AUFRWFZ8+eAQDeeusttWXleZrz8/Nx/vx5nbetLG+99ZbKtHWvvvoqrK2tAQD37t0rse3atWu4e/eudAx1FHNTx8TEqCynLh1ebGws0tPTAahP6efo6IiBAweqbU9gYCCA4sm5//vf/0pt/+mnn6QJjxMnTlR7LFVSUlIAFK9maGZmprLcgQMHABR/1vJMF6rIP8dz584pnZA5fvx46b2yyZjHjh1DUlISgOLPUNOVYuXZT/r16wcLCwuNjiX3/PlzxMXF4erVq/jzzz/x559/wtLSssS2yjp+/DiA4kmU6r7Xt99+G/Xq1SuxjzL9+/eHnZ2d0m3t2rWT3g8fPhympqZllivr3OR1yb9Dqr0YgBNRlebh4aG1wECd2NhY6b2zs7PaZeEV83Zr+oc0MjISovhppMqXuswuAPDaa6+p3V6/fn0AxSsnKlI85+7du6s9Z3kQD6g/57Zt26rc9ueff0rvO3bsqLbNnTp1Urt9+PDh0nkpW8RJ/rPOnTtXOs/6kydPAPz9+aki/xxv3ryp9jOUyWR4//33ARTfvMmPr+iNN96Ag4MDgOJMJy9TDMrlqQsrS2hx9cv79+9jxowZaNq0KerVq4dmzZqhTZs2Uvq9KVOmSGVTU1MrXY/8GnJ3d0fDhg1VljMzM5NWBlW87l7WokULldtsbW0rXO7l37GXya+ll7OrUO3DAJyIqrSygh9tefz4caX2y8rK0nJLKk7eu6iKkVHxf/WFhYUlfq6Lc1b3fSmmb1QXPJVnu4WFhdSLvnPnzhIBzdmzZ3H16lUAle/9ltcBQFptUhVtfo4mJiYYNWoUAODu3bs4ffq0tC0jI0Naer5z585l3niVRVurXx46dAitWrXCqlWrcP/+/TLLl/V5qiO/aZHfpKjj5ORUYh9l1P3uyH9vKlLu5d+xl8nPXVVvOtUeXAmTiKo0Y2NjvdSj+IfzwoUL5f4D6eLioqsm6ZziOR88eBBNmzYt137qgh99fV9A8TCU77//HhkZGdi9e7c0fEPe+12nTh0pmK0M+U1Aeno6hBAqh3vIP8d27drhp59+KvfxGzdurPTn48aNw8qVKwEU93j36NEDQPHS8/KgXdPc3wCk3u+OHTtWevXL1NRUjB49GllZWbC2tsacOXMwcOBANG/eHPXq1ZOG7kRERKBv374AinveNaXp0BtDkd8MKPaaU+3EAJyIqj3FHij54iTKqHvsq7jISsOGDat1YF1eiudsa2tb6aEa5aXYO56SkqL2sb58/LU6bdu2RefOnXHu3Dls3rwZ48ePR05OjrRAz/Dhw6VxwJUhD8CLiorw7NkzlUGT/HPMyMjQymfYoUMHtG7dGlevXsXOnTuxcuVKmJmZScNPTE1NNbqxkJOP/9Zk+Mnu3bulcf379u1Dv379lJZT1wtdEfIx1MnJyWWWlQ+VUjXG2xDkT4GaNGli4JaQoXEIChFVezY2NtJ7datU3rp1S+U2+XhRADh16pR2GlbF6fucW7duLb0va/Kq4vh0deSTMU+ePIl79+5h7969UkCoyfATACUmVJbn2rl3757WJtfJe7ifPHmCX3/9FQ8fPkRkZCQAYNCgQeValVOdhw8f4sKFCwA0C8DlQ33s7OxUBt9A2d9neXu05Tc4cXFxam/S8vPzpdU9dX1jWV5FRUXSRGjF3wWqnRiAE1G15+7uLr1X94d++/btKrf169dPGuf53XffaeUxeVXXoUMHqac/JCRE58tjd+rUSeqRVjdUIzk5GUeOHCnXMUeNGgUrKysIIbBlyxZp+Im7uzv69OmjUXt79eolvT937pzKckOGDAFQPLRCPnREU2PGjJGe7Pz4449aXXoe+Hv4iYuLS4kbsYqSZ3LJyclR+fQpKytLaUYXRfLx9rm5uWrLyYN8IYTSybdyu3fvlrIaqbsx0Kdr164hIyMDANC1a1cDt4YMjQE4EVV7bdq0kR4zr1q1Sukf8Z07d2LXrl0qj2FraytlqDh9+jRmz56tdjhLcnIyNmzYoGHLDcvIyAjz5s0DUNx7O27cOLUB0PPnz7Fq1apK12dhYSEFj+fOnVMarBYVFWHq1KnlvhmwsbHByJEjAQA//PADIiIiAAABAQEajxN2dXWFm5sbgOKJnaoMGDAAXbp0AQAsW7YMO3fuVHvcK1euSMM/VGnUqJE0Zjo8PBzr168HUNzT/M9//rPc56CKvH5NJl8CxVmKgOIgW9l5FxYWIjAwEImJiWqPIx+D/vjxY7WZRPz8/NCoUSMAwFdffYUrV66UKpOQkIA5c+YAKJ48OWHChPKdjI4pXkMDBgwwYEuoKmAATkTVnomJCaZOnQqgOOXY66+/jv379+PixYs4fPgwJk2ahFGjRkmT2VT54osvpJ6plStXokOHDli9ejVOnTqFS5cuITIyEqtWrYKfnx+aNGmCdevWadz2uLg4KVeyupeu8gYHBQVh2LBhAIBdu3ahdevWWLZsGaKionDp0iWcPHkSISEhGD16NBo1aoQFCxZoVN+CBQuk7BSzZs3CuHHjcOTIEVy4cAE7d+5Er169sH//fimgBcoeniAfhvL48WMUFRXByMgIAQEBGrVTbujQoQD+ThepSmhoKOzs7FBYWAh/f38MGTIE//d//4ezZ8/i/PnzOHToEL7++mt0794dbdu2RVRUVJl1y29W8vLycOfOHQCAv7+/2pzk5ZGdnS3dqGiafnDkyJFSDvoJEybgk08+wYkTJxAbG4utW7eia9eu2L59O3r27Kn2OPLfzaKiIgQFBeHMmTO4c+eO9JIzMzNDSEgIZDIZnj9/jp49e2LRokU4ffo0fv/9d3z77bfo1KmTFPD/97//RYMGDTQ6R205ceIEgOLJuopP7aiW0v/aP0RExVCOlTCVbVMmMzNTWjpa2cvHx0f8+eefKlfClHv+/LkYPnx4uVao7NOnT6XOu6IrYQIQH3zwQYljqFvV82XypbLHjx+vdHteXp6YNm2atLy7upe7u7va8ymPS5cuiYYNG6qsIyAgQGzcuFH6d1JSUpnHbNWqlVS+f//+5WpHeVy5ckU6blRUlNqyN2/eFG3atCnX97lw4cIy687MzBTW1tYl9ouJidH4nLS9+uWmTZuEkZGRynP19/cXx48fl/4dGRlZ6hiFhYVqf39ftmXLFmFubq6yvLGxsfj6669VtlleLjg4WGWZ8v6ORUZGqj03IYq/SysrKwFALFu2TOWxqPZgDzgR1QiWlpaIiIjAV199BU9PT9SpUwd169ZF586dsWrVKhw/fhxWVlZlHsfGxgZ79uxBdHQ0AgMD8eqrr8LGxgYmJiaws7ND586d8d577+HXX3/FsWPH9HBmumdqaoo1a9bg8uXLmDFjBjw9PVGvXj0YGxujXr168PLywqRJk7B7925cv35d4/ratWuHa9eu4cMPP4SHhwfMzc3RoEED9OnTB6Ghodi8eTOeP38ulS9PJhPFlTU1nXypqE2bNujevTuA4l5udVq0aIFLly4hNDQUI0aMQJMmTVCnTh2YmZnB2dkZPj4++Oyzz3D+/Hl8/vnnZdZtaWlZYoVSDw8PdOvWTbMTgvZXv5wwYQKio6Ph5+eHhg0bwtTUFM7OznjjjTfw888/Y8eOHWWmpzQyMsLRo0fx2WefoV27drC2tlb75GP8+PG4ceMGPvjgA7Rs2RJWVlaoU6cOmjdvjsmTJ+PixYuYO3euxuemLfv370dmZiYsLCyqzJAYMiyZELVgphEREVUrgYGB2LhxI1xcXJCQkFBm+XfffRehoaGoX78+Hj16JA2L0IadO3fC398f9evXx19//VViVdDqRggBFxcXJCYmIiQkBJMnTzZ0k2qFfv364cSJE5g6dapWhq5R9ccecCIiqlKys7Oxf/9+AChXj296ejr27dsHoDgQ12bwDQBvv/02OnbsiKdPn2o0CbUquHDhAhITEyGTybQymZPKdubMGZw4cQJmZmbSpGciBuBERKRXd+/eVTmhsbCwENOmTUNqaioASKtbqvPdd99JS3wHBQVpr6H/n0wmw9KlSwEAy5cvV7ugU1VXWFiI4OBgfPvtt5Ve/ZIqZuHChQCADz74gAvwkIRDUIiISK8CAgJw9uxZvPPOO+jatSscHByQnZ2NP/74A+vXr5cWiOnXrx+OHj1aaixwQUEB4uPjkZubi8jISMyZMwe5ubkYMmSI1HOuC99//z3S0tIwcuRItGrVSmf1UM2RlZWFZcuWAQBmz56NunXrGrhFVFUwACciIr0KCAjA1q1b1Zbp2bMn9u/fr3TFx/j4+FJp3OrVq4fz58+jefPmWm0rEZEumBi6AUREVLvMnTsXLVq0wPHjxxEfH4+UlBTk5+fD3t4enTp1gr+/P9555x1pJUh1HBwc0L17d3z11VcMvomo2mAPOBERERGRHnESJhERERGRHjEAJyIiIiLSIwbgRERERER6xACciIiIiEiPGIATEREREekRA3AiIiIiIj1iAE5EREREpEcMwImIiIiI9IgBOBERERGRHjEAJyIiIiLSIwbgRERERER6xACciIiIiEiPGIATEREREekRA3AiIiIiIj1iAE5EREREpEcMwImIiIiI9IgBOBERERGRHjEAJyIiIiLSIwbgRERERER69P8AJBPw3pdtcyMAAAAASUVORK5CYII=",
                        "text/plain": [
                            "<Figure size 350x300 with 1 Axes>"
                        ]
                    },
                    "metadata": {
                        "image/png": {
                            "height": 308,
                            "width": 368
                        }
                    },
                    "output_type": "display_data"
                }
            ],
            "source": [
                "analyse_model(model)"
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "## 3. Training loop\n",
                "\n",
                "As a simple and somewhat arbitrary example, below we implement a custom training loop with gradient accumulation. We make use of the [GraphDataLoader](https://jla-gardner.github.io/graph-pes/data/loader.html#graph_pes.data.loader.GraphDataLoader) class to create data-loaders for the training and validation sets - these automatically handle the batching of many [AtomicGraph](https://jla-gardner.github.io/graph-pes/data/atomic_graph.html#graph_pes.AtomicGraph) objects into a single [AtomicGraph](https://jla-gardner.github.io/graph-pes/data/atomic_graph.html#graph_pes.AtomicGraph) batch."
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 9,
            "metadata": {},
            "outputs": [
                {
                    "name": "stdout",
                    "output_type": "stream",
                    "text": [
                        "Epoch 01   ->   Val. loss: 0.6773 (eV) *\n",
                        "Epoch 02   ->   Val. loss: 0.6021 (eV) *\n",
                        "Epoch 03   ->   Val. loss: 0.4627 (eV) *\n",
                        "Epoch 04   ->   Val. loss: 0.4079 (eV) *\n",
                        "Epoch 05   ->   Val. loss: 0.3291 (eV) *\n",
                        "Epoch 06   ->   Val. loss: 0.3297 (eV) \n",
                        "Epoch 07   ->   Val. loss: 0.2644 (eV) *\n",
                        "Epoch 08   ->   Val. loss: 0.2699 (eV) \n",
                        "Epoch 09   ->   Val. loss: 0.3282 (eV) \n",
                        "Epoch 10   ->   Val. loss: 0.2494 (eV) *\n",
                        "Epoch 11   ->   Val. loss: 0.2546 (eV) \n",
                        "Epoch 12   ->   Val. loss: 0.2673 (eV) \n",
                        "Epoch 13   ->   Val. loss: 0.2476 (eV) *\n",
                        "Epoch 14   ->   Val. loss: 0.2041 (eV) *\n",
                        "Epoch 15   ->   Val. loss: 0.2162 (eV) \n",
                        "Epoch 16   ->   Val. loss: 0.2434 (eV) \n",
                        "Epoch 17   ->   Val. loss: 0.2577 (eV) \n",
                        "Epoch 18   ->   Val. loss: 0.2200 (eV) \n",
                        "Epoch 19   ->   Val. loss: 0.2087 (eV) \n",
                        "Epoch 20   ->   Val. loss: 0.1956 (eV) *\n",
                        "Epoch 21   ->   Val. loss: 0.1983 (eV) \n",
                        "Epoch 22   ->   Val. loss: 0.2091 (eV) \n",
                        "Epoch 23   ->   Val. loss: 0.1972 (eV) \n",
                        "Epoch 24   ->   Val. loss: 0.2212 (eV) \n",
                        "Epoch 25   ->   Val. loss: 0.1868 (eV) *\n"
                    ]
                },
                {
                    "data": {
                        "text/plain": [
                            "<All keys matched successfully>"
                        ]
                    },
                    "execution_count": 9,
                    "metadata": {},
                    "output_type": "execute_result"
                }
            ],
            "source": [
                "from graph_pes.atomic_graph import number_of_structures\n",
                "from graph_pes.data.loader import GraphDataLoader\n",
                "\n",
                "# hyperparameters\n",
                "accumulate_steps = 4\n",
                "n_epochs = 25\n",
                "batch_size = 32\n",
                "optimizer = torch.optim.Adam(model.parameters(), lr=3e-2)\n",
                "\n",
                "# data-loaders\n",
                "train_loader = GraphDataLoader(\n",
                "    train_graphs,\n",
                "    batch_size=batch_size,\n",
                "    shuffle=True,\n",
                ")\n",
                "val_loader = GraphDataLoader(\n",
                "    val_graphs,\n",
                "    batch_size=batch_size,\n",
                "    shuffle=False,\n",
                ")\n",
                "\n",
                "# keep track of the best model\n",
                "best_state_dict = {}\n",
                "best_loss = float(\"inf\")\n",
                "\n",
                "# training loop\n",
                "for epoch in range(n_epochs):\n",
                "    model.train()\n",
                "    optimizer.zero_grad()\n",
                "\n",
                "    for index, batch in enumerate(train_loader):\n",
                "        prediction = model.predict_energy(batch)\n",
                "        ground_truth = batch.properties[\"energy\"]\n",
                "\n",
                "        loss = (prediction - ground_truth).abs().mean()\n",
                "\n",
                "        loss.backward()\n",
                "\n",
                "        # accumulate gradients for `accumulate_steps` steps\n",
                "        if (index + 1) % accumulate_steps == 0:\n",
                "            optimizer.step()\n",
                "            optimizer.zero_grad()\n",
                "\n",
                "    # validate every epoch\n",
                "    model.eval()\n",
                "\n",
                "    # NB\n",
                "    # since we are only making energy predictions, we can save time and\n",
                "    # memory by turning off gradient calculations - if you want to\n",
                "    # calculate forces and stresses (using autograd internallly), you\n",
                "    # can't use a torch.no_grad() context!\n",
                "    with torch.no_grad():\n",
                "        losses, batch_sizes = [], []\n",
                "        for batch in val_loader:\n",
                "            prediction = model.predict_energy(batch)\n",
                "            ground_truth = batch.properties[\"energy\"]\n",
                "            loss = (prediction - ground_truth).abs().mean()\n",
                "            losses.append(loss.item())\n",
                "            batch_sizes.append(number_of_structures(batch))\n",
                "\n",
                "        avg_loss = sum(\n",
                "            loss * batch_size for loss, batch_size in zip(losses, batch_sizes)\n",
                "        ) / sum(batch_sizes)\n",
                "\n",
                "    # save the best model\n",
                "    if avg_loss < best_loss:\n",
                "        best_loss = avg_loss\n",
                "        best_state_dict = model.state_dict()\n",
                "        is_best = True\n",
                "    else:\n",
                "        is_best = False\n",
                "\n",
                "    # log metrics\n",
                "    print(\n",
                "        f\"Epoch {epoch + 1:02d}   ->   Val. loss: {avg_loss:.4f} (eV)\",\n",
                "        \"*\" if is_best else \"\",\n",
                "    )\n",
                "\n",
                "\n",
                "# load the best model\n",
                "model.load_state_dict(best_state_dict)"
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "Unsurprisingly, training has improved this toy model's performance:"
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 10,
            "metadata": {},
            "outputs": [
                {
                    "data": {
                        "image/png": "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",
                        "text/plain": [
                            "<Figure size 350x300 with 1 Axes>"
                        ]
                    },
                    "metadata": {
                        "image/png": {
                            "height": 308,
                            "width": 368
                        }
                    },
                    "output_type": "display_data"
                }
            ],
            "source": [
                "analyse_model(model)"
            ]
        }
    ],
    "metadata": {
        "kernelspec": {
            "display_name": "graph-pes",
            "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": "3.9.21"
        }
    },
    "nbformat": 4,
    "nbformat_minor": 2
}