Model_testing.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## **Functions and Classes**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **Packages and Utilities**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### **Import Some Packages**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# PyTorch\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.optim as optim\n",
    "from torch.utils.data import Dataset, DataLoader, SubsetRandomSampler\n",
    "from torchmetrics import R2Score\n",
    "\n",
    "# Data process\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import math\n",
    "import csv\n",
    "import os\n",
    "from sklearn.preprocessing import StandardScaler, MinMaxScaler\n",
    "from sklearn.model_selection import KFold\n",
    "\n",
    "\n",
    "# Plot figure\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.pyplot import figure"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### **Reproducibility**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "myseed = 870529  # set a random seed for reproducibility\n",
    "torch.backends.cudnn.deterministic = True\n",
    "torch.backends.cudnn.benchmark = False\n",
    "np.random.seed(myseed)\n",
    "torch.manual_seed(myseed)\n",
    "if torch.cuda.is_available():\n",
    "    torch.cuda.manual_seed_all(myseed)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### **Switch to GPU if available**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_device():\n",
    "    ''' Get device (if GPU is available, use GPU) '''\n",
    "    return 'cuda' if torch.cuda.is_available() else 'cpu'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **Dataset**\n",
    "\n",
    "We have two kinds of datasets:\n",
    "* `train`: for training and validation\n",
    "* `test`: for testing \n",
    "\n",
    "The `TrajDataset` below does:\n",
    "* read `.xlsx` files\n",
    "* extract features\n",
    "* normalize features\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "class TrajDataset(Dataset):\n",
    "    def __init__(self, path, mode='test'):\n",
    "\n",
    "        # Read data to dataframe\n",
    "        df = pd.read_excel(path)\n",
    "\n",
    "        # Split data to feature and target & Transfer to numpy \n",
    "        feature_np = df.iloc[:, 1:10].to_numpy()\n",
    "        target_np = df.iloc[:, 10:13].to_numpy()\n",
    "        \n",
    "        # Transfer data to tensor\n",
    "        self.feature = torch.Tensor(feature_np)\n",
    "        self.target = torch.Tensor(target_np)\n",
    "\n",
    "        # Transfer feature data to (0,1) -> MinMaxScalar\n",
    "        self.feature[:, 1:4] = (self.feature[:, 1:4]   150) / 300\n",
    "        self.feature[:, 4:7] = (self.feature[:, 4:7]   830) / 1660      \n",
    "        self.feature[:, 7:10] = (self.feature[:, 7:10]   980) / 1960\n",
    "\n",
    "        # Get sample number of data\n",
    "        self.dim = self.feature.shape[1]\n",
    "\n",
    "        print('Finished reading the Trajectory Dataset ({} samples found, each dim = {})'\n",
    "              .format(len(self.feature), self.dim))\n",
    "\n",
    "    def __getitem__(self, index) :\n",
    "\n",
    "\n",
    "        return self.feature[index], self.target[index]\n",
    "\n",
    "    def __len__(self) :\n",
    "\n",
    "        return len(self.feature)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **Deep Neural Network**\n",
    "\n",
    "`NeuralNet` is an `nn.Module` designed for regression.\n",
    "The DNN consists of 7 fully-connected layers with ReLU activation.\n",
    "This module also included a function `cal_loss` for calculating loss."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "class DDNN(nn.Module) : \n",
    "    def __init__(self, layers):\n",
    "        super(DDNN, self).__init__()\n",
    "\n",
    "        input_dim = 9\n",
    "        output_dim = 3\n",
    "\n",
    "        h1, h2, h3 = layers[1], layers[2], layers[3]\n",
    "        h4, h5  = layers[4], layers[5]\n",
    "\n",
    "        # Layer Strucutre\n",
    "        self.l1 = nn.Linear(input_dim, h1)\n",
    "        self.l2 = nn.Linear(h1, h2)\n",
    "        self.l3 = nn.Linear(h2, h3)\n",
    "        self.l4 = nn.Linear(h3, h4)\n",
    "        self.l5 = nn.Linear(h4, h5)\n",
    "        self.l6 = nn.Linear(h5, output_dim)\n",
    "\n",
    "\n",
    "        # Define Activation Function \n",
    "        # Use Softplus for smoother result (compared with ReLU)\n",
    "        self.act1 = nn.Softplus()\n",
    "        self.act2 = nn.Softplus()\n",
    "        self.act3 = nn.Softplus()\n",
    "        self.act4 = nn.Softplus()\n",
    "        self.act5 = nn.Softplus()\n",
    "\n",
    "\n",
    "    def forward(self, x):\n",
    "        \n",
    "        x = self.l1(x)\n",
    "        x = self.act1(x)\n",
    "        x = self.l2(x)\n",
    "        x = self.act2(x)\n",
    "        x = self.l3(x)\n",
    "        x = self.act3(x)\n",
    "        x = self.l4(x)\n",
    "        x = self.act4(x)\n",
    "        x = self.l5(x)\n",
    "        x = self.act5(x)\n",
    "        x = self.l6(x)\n",
    "\n",
    "        return x\n",
    "    \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### **Get hidden layer value**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Hook():\n",
    "    def __init__(self, module, backward=False):\n",
    "        if backward==False:\n",
    "            self.hook = module.register_forward_hook(self.hook_fn)\n",
    "        else:\n",
    "            self.hook = module.register_backward_hook(self.hook_fn)\n",
    "    def hook_fn(self, module, input, output):\n",
    "        self.input = input\n",
    "        self.output = output\n",
    "    def close(self):\n",
    "        self.hook.remove()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### **Inverse Dynamics**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def InvDyn(tk_space):\n",
    "\n",
    "    samples = tk_space.shape[0]    \n",
    "    # Parameters\n",
    "    unit = 1000\n",
    "    L_init = (518.067753315)/unit\n",
    "    D = (450 - 160)/unit\n",
    "    Norm_Z = 429.295/unit\n",
    "\n",
    "    # Task space data\n",
    "    x_raw, y_raw, z_raw = tk_space[:, 0], tk_space[:, 1], tk_space[: , 2]\n",
    "    x_d_raw, y_d_raw, z_d_raw = tk_space[:, 3], tk_space[:, 4], tk_space[: , 5]\n",
    "    x_dd_raw, y_dd_raw, z_dd_raw = tk_space[:, 6], tk_space[:, 7], tk_space[:, 8]\n",
    "\n",
    "\n",
    "    x, y, z = (300*x_raw-150)/unit, (300*y_raw-150)/unit, (300*z_raw-150)/unit\n",
    "    x_d, y_d, z_d = (1660*x_d_raw-830)/unit, (1660*y_d_raw-830)/unit, (1660*z_d_raw-830)/unit\n",
    "    x_dd, y_dd, z_dd = (1960 * x_dd_raw - 980)/unit, (1960 * y_dd_raw - 980)/unit, (1960 * z_dd_raw - 980)/unit\n",
    "\n",
    "\n",
    "    L1 = torch.sqrt( x**2   (y-D)**2   (z-Norm_Z)**2 ) \n",
    "    L2 = torch.sqrt( (x   math.sqrt(3) * D / 2) **2   ( y D/2 )**2   (z-Norm_Z)**2 ) \n",
    "    L3 = torch.sqrt( (x - math.sqrt(3) * D / 2) **2   ( y D/2 )**2   (z-Norm_Z)**2 ) \n",
    "    \n",
    "    L1_d = x / L1 * x_d   ( y - D ) / L1 * y_d   ( z - Norm_Z ) / L1 * z_d  \n",
    "    L2_d = (x   math.sqrt(3) * D / 2 ) / L2 * x_d   ( y   D / 2 ) / L2 * y_d   ( z - Norm_Z ) / L2 * z_d\n",
    "    L3_d = (x - math.sqrt(3) * D / 2 ) / L3 * x_d   ( y   D / 2 ) / L3 * y_d   ( z - Norm_Z ) / L3 * z_d \n",
    "\n",
    "    L1_dd =  x / L1 * (x_dd - x_d * L1_d/L1)   x_d**2 / L1   \\\n",
    "        ( y - D ) * (y_dd - y_d * L1_d / L1)   y_d**2 / L1   \\\n",
    "        ( z - Norm_Z ) * (z_dd - z_d * L1_d / L1)   z_d**2 / L1 \n",
    "    L2_dd = (x   math.sqrt(3) * D / 2 ) / L2 * (x_dd - x_d * L2_d / L2)   x_d**2 / L2   \\\n",
    "        (y   D / 2) / L2 * (y_dd - y_d * L2_d / L2)   y_d**2 / L2   \\\n",
    "        ( z - Norm_Z ) / L2 * (z_dd - z_d * L2_d / L2)   z_d**2 / L2\n",
    "    L3_dd = (x - math.sqrt(3) * D / 2 ) / L3 * (x_dd - x_d * L3_d / L3)   x_d**2 / L3   \\\n",
    "        (y   D / 2) / L3 * (y_dd - y_d * L3_d / L3)   y_d**2 / L3   \\\n",
    "        ( z - Norm_Z ) / L3 * (z_dd - z_d * L3_d / L3)   z_d**2 / L3\n",
    "\n",
    "    L1, L2, L3 =  torch.reshape(L1, (samples, -1)), torch.reshape(L2, (samples, -1)), torch.reshape(L3, (samples, -1))  \n",
    "    L1_d, L2_d, L3_d =  torch.reshape(L1_d, (samples, -1)), torch.reshape(L2_d, (samples, -1)), torch.reshape(L3_d, (samples, -1))\n",
    "    L1_dd, L2_dd, L3_dd =  torch.reshape(L1_dd, (samples, -1)), torch.reshape(L2_dd, (samples, -1)), torch.reshape(L3_dd, (samples, -1)) \n",
    "            \n",
    "    jt_space = torch.cat((L1, L2, L3, L1_d, L2_d, L3_d, L1_dd, L2_dd, L3_dd),1)\n",
    "    \n",
    "    return jt_space"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### **Custom Loss**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "class CustomLoss(nn.Module):\n",
    "    def __init__(self, mode='mse'):\n",
    "        super(CustomLoss, self).__init__()\n",
    "        self.mode = mode\n",
    "\n",
    "    def forward(self, output, target, tk_space, jt_space):\n",
    "\n",
    "        if self.mode == 'phys1':\n",
    "\n",
    "            jt_target = InvDyn(tk_space) \n",
    "\n",
    "            mse_loss = torch.mean((output-target)**2)\n",
    "\n",
    "            pen_loss = torch.mean((jt_space-jt_target)**2)\n",
    "\n",
    "            penalty = 10**(2)\n",
    "\n",
    "            total_loss = mse_loss   penalty * pen_loss\n",
    "\n",
    "            return total_loss, mse_loss, penalty * pen_loss\n",
    "\n",
    "        elif self.mode == 'phys2':\n",
    "\n",
    "            # Parameters\n",
    "            L_init = 518.067753315/1000\n",
    "            D = (450 - 160)/1000\n",
    "            Norm_Z = 429.295/1000\n",
    "            # Task space data\n",
    "            x_raw, y_raw, z_raw = tk_space[:, 0], tk_space[:, 1], tk_space[: , 2]\n",
    "            x_d_raw, y_d_raw, z_d_raw = tk_space[:, 3], tk_space[:, 4], tk_space[: , 5]\n",
    "            x_dd_raw, y_dd_raw, z_dd_raw = tk_space[:, 6], tk_space[:, 7], tk_space[:, 8]\n",
    "\n",
    "            L1_pred, L2_pred, L3_pred = jt_space[:, 0], jt_space[:, 1], jt_space[: , 2]\n",
    "            L1_d_pred, L2_d_pred, L3_d_pred = jt_space[:, 3], jt_space[:, 4], jt_space[: , 5]\n",
    "            L1_dd_pred, L2_dd_pred, L3_dd_pred = jt_space[:, 6], jt_space[:, 7], jt_space[:, 8]\n",
    "\n",
    "            x, y, z = (300*x_raw-150)/100, (300*y_raw-150)/100, (300*z_raw-150)/100\n",
    "            x_d, y_d, z_d = (1660*x_d_raw-830)/100, (1660*y_d_raw-830)/100, (1660*z_d_raw-830)/100\n",
    "            x_dd, y_dd, z_dd = (1960 * x_dd_raw - 980)/100, (1960 * y_dd_raw - 980)/100, (1960 * z_dd_raw - 980)/100\n",
    "\n",
    "            L1_target = torch.sqrt( x**2   (y-D)**2   (z-Norm_Z)**2 ) \n",
    "            L2_target = torch.sqrt( (x   math.sqrt(3) * D / 2) **2   ( y D/2 )**2   (z-Norm_Z)**2 ) \n",
    "            L3_target = torch.sqrt( (x - math.sqrt(3) * D / 2) **2   ( y D/2 )**2   (z-Norm_Z)**2 ) \n",
    "            \n",
    "            L1_d_target = x / L1_target * x_d   ( y - D ) / L1_target * y_d   ( z - Norm_Z ) / L1_target * z_d  \n",
    "            L2_d_target = (x   math.sqrt(3) * D / 2 ) / L2_target * x_d   ( y   D / 2 ) / L2_target * y_d   ( z - Norm_Z ) / L2_target * z_d\n",
    "            L3_d_target = (x - math.sqrt(3) * D / 2 ) / L3_target * x_d   ( y   D / 2 ) / L3_target * y_d   ( z - Norm_Z ) / L3_target * z_d \n",
    "\n",
    "            L1_dd_target =  x / L1_target * (x_dd - x_d * L1_d_target/L1_target)   x_d**2 / L1_target   \\\n",
    "                ( y - D ) * (y_dd - y_d * L1_d_target / L1_target)   y_d**2 / L1_target   \\\n",
    "                ( z - Norm_Z ) * (z_dd - z_d * L1_d_target / L1_target)   z_d**2 / L1_target\n",
    "            L2_dd_target = (x   math.sqrt(3) * D / 2 ) / L2_target * (x_dd - x_d * L2_d_target / L2_target)   x_d**2 / L2_target   \\\n",
    "                (y   D / 2) / L2_target * (y_dd - y_d * L2_d_target / L2_target)   y_d**2 / L2_target   \\\n",
    "                ( z - Norm_Z ) / L2_target * (z_dd - z_d * L2_d_target / L2_target)   z_d**2 / L2_target\n",
    "            L3_dd_target = (x - math.sqrt(3) * D / 2 ) / L3_target * (x_dd - x_d * L3_d_target / L3_target)   x_d**2 / L3_target   \\\n",
    "                (y   D / 2) / L3_target * (y_dd - y_d * L3_d_target / L3_target)   y_d**2 / L3_target   \\\n",
    "                ( z - Norm_Z ) / L3_target * (z_dd - z_d * L3_d_target / L3_target)   z_d**2 / L3_target\n",
    "\n",
    "\n",
    "            pos_target, pos_pred = torch.stack((L1_target, L2_target, L3_target), -1), torch.stack((L1_pred, L2_pred, L3_pred), -1)\n",
    "            vel_target, vel_pred = torch.stack((L1_d_target, L2_d_target, L3_d_target), -1), torch.stack((L1_d_pred, L2_d_pred, L3_d_pred), -1) \n",
    "            acc_target, acc_pred = torch.stack((L1_dd_target, L2_dd_target, L3_dd_target), -1), torch.stack((L1_dd_pred, L2_dd_pred, L3_dd_pred), -1) \n",
    "\n",
    "            mse_loss = torch.mean((output-target)**2)\n",
    "            pos_loss = torch.mean((pos_pred-pos_target)**2)\n",
    "            vel_loss = torch.mean((vel_pred-vel_target)**2)\n",
    "            acc_loss = torch.mean((acc_pred-acc_target)**2)\n",
    "\n",
    "            pos_pen = 1\n",
    "            vel_pen = 1\n",
    "            acc_pen = 1\n",
    "\n",
    "            pen_loss = pos_loss   vel_loss   acc_loss\n",
    "\n",
    "            total_loss = mse_loss   pos_pen * pos_loss   vel_pen * vel_loss   acc_pen * acc_loss\n",
    "\n",
    "            return total_loss, mse_loss, pen_loss\n",
    "        \n",
    "        else :\n",
    "\n",
    "            jt_target = InvDyn(tk_space) \n",
    "\n",
    "            mse_loss = torch.mean((output-target)**2)\n",
    "\n",
    "            pen_loss = torch.mean((jt_space-jt_target)**2)\n",
    "\n",
    "            penalty = 0\n",
    "\n",
    "            total_loss = mse_loss   penalty * pen_loss\n",
    "\n",
    "            return total_loss, mse_loss, pen_loss\n",
    "\n",
    "       "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **Train/Dev/Test**\n",
    "\n",
    "We have three kinds of dataloaders:\n",
    "* `train`: for training\n",
    "* `valid`: for validation\n",
    "* `test`: for testing "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### **Hyper-parameters Setup**\n",
    "\n",
    "`config` contains hyper-parameters for training and the path to save your model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "device = get_device()\n",
    "os.makedirs('models', exist_ok=True) \n",
    "\n",
    "config = {\n",
    "    'n_epochs': 50,                    # maximum number of epochs\n",
    "    'batch_size': 16,                   # mini-batch size for dataloader\n",
    "    'optimizer': 'Adam',                # optimization algorithm (optimizer in torch.optim)\n",
    "    'optim_hparas': {                   # hyper-parameters for the optimizer (depends on which optimizer you are using)\n",
    "        'lr': 0.0005433177089293607                     # learning rate of Adam\n",
    "    },\n",
    "    'layers': np.array([9, 189, 119, 9, 53, 85, 3], dtype='int64'),    # layers of NN architecture\n",
    "    'early_stop': 200,                  # early stopping epochs (the number epochs since your model's last improvement)\n",
    "    'save_path': 'models/model.pt'      # your model will be saved here\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **Read Records**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle\n",
    "\n",
    "record_path = 'records/record_DDNN.pkl'\n",
    "f = open(record_path, 'rb')\n",
    "DDNN_noisefree = pickle.load(f)\n",
    "\n",
    "record_path = 'records/record_PGNN.pkl'\n",
    "f = open(record_path, 'rb')\n",
    "PGNN_noisefree = pickle.load(f)\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **Plot Learning Curve**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAEWCAYAAACXGLsWAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8qNh9FAAAACXBIWXMAAAsTAAALEwEAmpwYAABPZklEQVR4nO3dd3iUZdbA4d ZzEx6QggtIXTpXRAr0sSCnVXXsisq1nUta0NXPzv2spZ1LWvBhoVVQKUogqgI0kWKSIAgIaQQIL3P8/3xvgmTkGQmkDIJ5/aaKzNvPTPBOXm6GGNQSimlauNo6gCUUkoFPk0WSimlfNJkoZRSyidNFkoppXzSZKGUUsonTRZKKaV80mShmhURGSkim5s6DuUfEZkrIpOaOg51 ETHWSh/iUgScLUxZkFTx3IkEZEo4GFgItAaSAO AB41xuxpytjUkUNLFiqgiEhQU8dwuOrzPYiIG/gW6A cDkQBxwOZwIj6uk99E4t v7Qg stUh01EHCJyt4hsFZFMEflERFp77f9URFJFJEtEvheR/l773hGR/4jIHBHJA8aISJKI3CEi6 xzPhaREPv40SKS7HV jcfa  8Skd0ikiIiV4uIEZGjangfrUXkbfvYfSIy095 hYj8WOXYiutU8x7usN9vkNfx54vIOn8 ryouBzoD5xtjNhpjPMaYdGPMI8aYOfb1 orIdyKyX0Q2iMg5VT7fV zqoFwRWSIiHUTkX/Z7/E1Ehlb5PO8RkY32/re9PvsYEflSRDLsfV KSILXud JyFQRWQLkA93tbVfb 48SkcX272mPiHzsde4JIrLC3rdCRE6oct1H7NhzRORrEWlTw elGogmC1UfbgLOA0YB8cA 4N9e  cCPYF2wGrggyrnXwpMBSKB8i/li7D ku4GDAKuqOX 1R4rIqcDtwGnAEcBo328j/eAMKy/4tsBz/s4vqb38AKQB4ytsv9D 7mvz8vbKcA8Y0xudTtFxIVVJfW1HfNNwAci0tvrsIuA 4A2QBGwFOv30AaYATxX5bKXAacBPYBe9rlgfV 8DXTBSmAFwMtVzv0rcK39Oeyosu8RO84YIAF4yX4PrYGvgBeBWDuer0Qk1uvcS4Er7ffoBu6o7vNQDcgYow99 PUAkoBTqtm CRjn9ToOKAGc1RzbCjBAtP36HeDdau7zF6/XTwGv2s9HA8l HvsW8LjXvqPsex9VTVxxgAeIqWbfFcCPVbZVXKeG9/Ao8Jb9PBIreXQ5hM/rG CJWn4nI4FUwOG1bTrwoFdsb3jtuwnY5PV6ILC/yud5vdfrCcDWGu49BNjn9fo74OEqx3yH1c4F8C7wOpBQ5Zi/AsurbFsKXOF1jfu89v0NK4E2 f8TR9JDSxaqPnQBPrerQfZjfRmWAe1FJEhEnrCrXLKxvozA qu23M5qrpnq9TwfiKjl/jUdG1/l2tXdp1wnYK8xZl8tx9Sm6rU/BCaKSDBWw/RqY0z5X9o1fl7VXDcTK5nUJB7YaYzxeG3bAXT0ep3m9bygmtdVP1vv97LDvgciEiYir4nIDvt3 T3QSiq30dT2Gd8FCLDcri67yus9VC2FVH0Pdfn3oBqAJgtVH3YCZxhjWnk9Qowxu7CqD87Fqk6JBrra54jX Q3VJW83VnVHuU61HLsTaC0irarZl4dVPQWAiHSo5phK78EYsxHrC 8MKldBld rps rqgXAaSISXkPcKUAnqdyY3Bmo7lr 8v6cOtv3ALgd6A0ca4yJAk62t/v1uzTGpBpjrjHGxAPXAa/Y7T4pWAnU2 G B1XPNFmounKJSIjXwwm8CkwVkS4AItJWRM61j4/EqifPxPrCfawRY/0EuNJuAA4D/q mA40xu7HaVl6xG3JdIlL ZfgL0F9EhtiNvQ/6ef8PgVuwvlQ/9dpe2 dV1XtYyeV/ItLHbhyPFZF/isgE4Gesv7TvsmMeDZwNfORnjNW5UUQS7LaEe4HyhuhIrJLIfnvfA3W5qIhc6NUgvg8rsXiAOUAvEblURJwi8megH/DlYbwHVc80Wai6moP1hVH eBCrQXc28LWI5ADLgGPt49/F gt7F7DR3tcojDFzsRpNFwGJXvcuquGUv2K1HfwGpAO32tf5HWucwwJgCwca4X2ZjtWIvdBUHg9R2 dV9T0UYZXKfsNqv8gGlmNV4/1sjCnGSg5nAHuAV4DLjTG/ RljdT7EaojeBmzFan8B BcQat9nGTCvjtc9BvhZRHKx3v8txphtxphM4CyskksmVnXVWUbHkAQUHZSnjhgi0hdYDwQbY0qbOp5AJDrwUtVASxaqRbPHNwSLSAzwJPCFJgql6q7BkoWIvCUi6SKy3mtbaxH5RkS22D9j7O0iIi KSKJYg6uO9jpnkn38FtE5ZlTdXYdVpbQVq8fRDU0bjlLNU4NVQ9mNg7lY/c8H2Nuewuqe ISI3I3Vp32K3VB3E1af7mOBF4wxx9qNaCuB4ViNYauAYYfRvVEppdQhaLCShTHme2Bvlc3nAtPs59OwRrGWb3/XWJZh9d2OwxpF o0xprz/ zdYI3WVUko1Imcj36 93UURrEE25YOQOlJ5ME yva2m7QcRkWuxphkgPDx8WJ8 feot6ILSArZlbaNLZBci3DWPBSrbt5 SXbsI7tULcbvq7f5KKdUYVq1atccY07a6fY2dLCoYY4yI1FsdmDHmdaypBBg fLhZuXJlfV2a9XvWc8lXl/Dy2JcZ1WlUjcdlffEFKXfeRY/Zs3B37Vpv91dKqcYgIlVH0ldo7N5QaXb1EvbPdHv7LiqPGk2wt9W0vVGJPUDV BhoLE4r95qSkgaPSSmlGlNjJ4vZQHmPpknALK/tl9u9oo4DsuzqqvnAqfaI2hjgVHtb47InM/DVGUBcVtWTKdWemUqplqXBqqFEZDrWDKFtxFp/4AHgCeATEZmMNar3IvvwOVg9oRKxpi64EsAYs1dEHgFW2Mc9bIyp2mje4Bx2TvVVskBLFkqpFqrBkoUx5pIado2r5lgD3FjDdd7Cmmq6yYhIeSy1H6clC3UEKSkpITk5mcLCwqYORdVRSEgICQkJuFz d8Rpsgbu5sT/Ngs7WRRryUK1fMnJyURGRtK1a9eKP6hU4DPGkJmZSXJyMt26dfP7PJ3uow58JgstWagjSGFhIbGxsZoomhkRITY2ts4lQk0WfvC/Gqq8zaK4wWNSKhBoomieDuX3psnCD/42cGvJQinVUmmy8IPfJQu7NxTaG0qpBpeUlMSAAQMqbXvwwQd55pln6vU 77zzDg6Hg3Xr1lVsGzBgAElJSbWeN2HCBPbv339Y987IyODYY49l6NCh/PDDD4d1rcOlycIPfjdwa8lCqRYpISGBqVOn1umcOXPm0KpVq8O677fffsvAgQNZs2YNI0eOrLSvrKzssK5dV5os/OHvoDwdZ6FUwBg9ejRTpkxhxIgR9OrVq Iv87KyMu644w4GDBjAoEGDeOmll3xe66yzzmLDhg1s3rz5oH3Tp09n4MCBDBgwgClTplRs79q1K3v27CEvL48zzzyTwYMHM2DAAD7 2FqldtWqVYwaNYphw4Zx2mmnsXv37krXXbt2LXfddRezZs1iyJAhFBQUEBERwe23387gwYNZunQp77//PiNGjGDIkCFcd911FQnk66 /5vjjj foo4/mwgsvJDc395A/x3LaddYP/pYsKC9ZlGjJQh1ZUh97jKJNh7OS68GC /ahwz//eVjXKC0tZfny5cyZM4eHHnqIBQsW8Prrr5OUlMTatWtxOp3s3et7nK/D4eCuu 7iscceY9q0aRXbU1JSmDJlCqtWrSImJoZTTz2VmTNnct5551UcM2/ePOLj4/nqq68AyMrKoqSkhJtuuolZs2bRtm1bPv74Y 69917eeuvAkLIhQ4bw8MMPs3LlSl5  WUA8vLyOPbYY3n22WfZtGkTTz75JEuWLMHlcvG3v/2NDz74gAkTJvDoo4 yYMECwsPDefLJJ3nuuee4//77D uz1GThB4fUsYFbSxZKNbiaevR4b584cSIAw4YNq2hjWLBgAddffz1OuyagdevWft3v0ksvZerUqWzfvr1i24oVKxg9ejRt21oTtV522WV8//33lZLFwIEDuf3225kyZQpnnXUWI0eOZP369axfv57x48cDVmknLi7OZwxBQUH86U9/AqwqqlWrVnHMMccAUFBQQLt27Vi2bBkbN27kxBNPBKC4uJjjjz/er/dYG00WfqgoWegIbqWqdbglgEMRGxvLvn2V10Hbu3dvpYFmwcHBgPUlW3qY/186nU5uv/12nnzyyTqd16tXL1avXs2cOXO47777GDduHOeffz79 /dn6dKldbpWSEgIQUFBgPV9NGnSJB5//PFKx3zxxReMHz e6dOn1 navmibhR901lmlAk9ERARxcXEsXLgQsBLFvHnzOOmkk2o9b/z48bz22msVycOfaqhyV1xxBQsWLCAjIwOAESNGsHjxYvbs2UNZWRnTp09n1KjKyxikpKQQFhbGX/7yF 68805Wr15N7969ycjIqEgWJSUlbNiwwe84AMaNG8eMGTNIT0 veB87duzguOOOY8mSJSQmJgJW1dXvv/9ep2tXR5OFP o866wmC6Uaw7vvvssjjzzCkCFDGDt2LA888AA9evSo9Zyrr76azp07M2jQIAYPHsyHH34IwP3338/s2bNrPdftdnPzzTdXfEHHxcXxxBNPMGbMGAYPHsywYcM499xzK53z66 /VjRCP/TQQ9x333243W5mzJjBlClTGDx4MEOGDOGnn36q03vv168fjz76KKeeeiqDBg1i/Pjx7N69m7Zt2/LOO 9wySWXMGjQII4//nh   3w25MabA3uplTfix8l5yRzxmdn8MiJj3DeUefVeuymfv2Jve5a2t1yS73dX6lAtGnTJvr27dvUYahDVN3vT0RWGWOGV3e8liz8UNHA7UdiFacTtM1CKdXCaLLwg99dZ7GqonTWWaVUS6PJwg/ TvcBVslCG7iVUi2NJos68KdkgdulXWeVUi2OJgs/1KkayunSkoVSqsXRZOGHOjVwu7RkoZRqeTRZ EHbLJQKTKmpqVx88cX06NGDYcOGMWHChHoZgOZNpyi3aLKoA797Q mgPKUanDGG888/n9GjR7N161ZWrVrF448/TlpaWr3fS6co12Thl7q1WWjJQqnGsGjRIlwuF9dff33FtsGDBzNy5EiMMdx5550MGDCAgQMHVkwL/t133zF69GguuOAC vTpw2WXXeZXjYFOUa4TCfqlTtVQLhfoFOXqCPPk8if5bW/9TlHep3UfpoyYUuP 9evXM2zYsGr3ffbZZ6xdu5ZffvmFPXv2cMwxx3DyyScDsGbNGjZs2EB8fDwnnngiS5Ys8TmflE5RrsnCL/6uwQ1aslAqEPz4449ccsklBAUF0b59e0aNGsWKFSuIiopixIgRJCQkANYXclJSks9kATpFuSYLP9SpZOF24SkobOiQlAootZUAGkr//v2ZMWNGnc8rn7Yc6jZ1uU5Rrvzm16A8LVko1SjGjh1LUVERr7/ esW2devW8cMPPzBy5Eg /vhjysrKyMjI4Pvvv2fEiBGHfU dolzVqm5tFm4dZ6FUIxARPv/8cxYsWECPHj3o378/99xzDx06dOD888 vmIJ87NixPPXUU3To0KHW6 kU5bXTKcr9kFOcwwnTT CO4Xcwqf kWo9NvuVWihIT6fHVl/V2f6UCkU5R3rzpFOUNoO4juLUaSinVsmiy8EOdpyjXNgulVAujyaIO/O06q Ms1JGiJVZjHwkO5femycIPdR2UpyULdSQICQkhMzNTE0YzY4whMzOTkJCQOp2n4yz8UNFm4Vc1lFN7Q6kjQkJCAsnJyRXdSFXzERISUjEw0V aLPxQ0WahJQulKrhcLrp169bUYahGotVQfqhLAzdOLVkopVqeJkkWIvIPEdkgIutFZLqIhIhINxH5WUQSReRjEXHbxwbbrxPt/V0bP2Drh98TCZaVYRp5 mCllGpIjZ4sRKQjcDMw3BgzAAgCLgaeBJ43xhwF7AMm26dMBvbZ25 3j2vcmOu4rCqgpQulVIvSVNVQTiBURJxAGLAbGAuUzwo2DTjPfn6u/Rp7/zgp757USOrWwG0nC 0 q5RqQRo9WRhjdgHPAH9gJYksYBWw3xhT/g2bDHS0n3cEdtrnltrHx1a9rohcKyIrRWRlfffOqFMDt9PqM2BKius1BqWUakpNUQ0Vg1Va6AbEA HA6Yd7XWPM68aY4caY4eVzy9eX8oKMx3h8H u2ShZoNZRSqgVpimqoU4DtxpgMY0wJ8BlwItDKrpYCSAB22c93AZ0A7P3RQGbjhgyhzlAKSgt8HnegZKHdZ5VSLUdTJIs/gONEJMxuexgHbAQWARfYx0wCZtnPZ9uvsfcvNE0wZDTcFU5eSZ7P4yraLLRkoZRqQZqizeJnrIbq1cCvdgyvA1OA20QkEatN4k37lDeBWHv7bcDdjR0z J8s0JKFUqoFapIR3MaYB4AHqmzeBhy0lJUxphC4sDHiqo2WLJRSRzIdwe0nv5NF TiLYi1ZKKVaDk0Wfgp31rVkoclCKdVyaLLwU7i7jslC2yyUUi2IJgs/hTvDyS/N93mcuOxmIG2zUEq1IJos/BTuDie3ONfncVqyUEq1RJos/BTuDKfYU0yJp/YkUDEoT0sWSqkWRJOFn8Jd4QDkl9ReFaUlC6VUS6TJwk/lySK3xEdVVMWgPC1ZKKVaDr8G5YlIO6z5m KBAmA9sNIYP2bWayHKk4WvHlHicgNaslBKtSy1JgsRGYM1vUZrYA2QDoRgrTXRQ0RmAM8aY7IbOM4m53 yKG z0GShlGo5fJUsJgDXGGP qLrDngH2LGA88L8GiC2g J8stM1CKdXy1JosjDF31rKvFJhZ3wEFKr ThVPHWSilWh6fDdwi0kdExolIRJXth71gUXOiJQul1JGs1mQhIjdjrStxE7BeRM712v1YQwYWaOpastBkoZRqSXy1WVwDDDPG5IpIV2CGiHQ1xrwA9sLURwh/k4V2nVVKtUS koXDGJMLYIxJEpHRWAmjC0dYsnA6nAQHBfsuWYiAy6UlC6VUi KrzSJNRIaUv7ATx1lAG2BgA8YVkOqyAJJO96GUakl8JYvLgVTvDcaYUmPM5cDJDRZVgPJ/ASSnliyUUi2Kr66zybXsW1L/4QS2CFdEHUoWmiyUUi2HX3NDiciDDRxHsxDmCvM/WWjJQinVgvjqOusQkTeB4EaKJ6DVpRpKB UppVoSXyWLL4G9xph7GiOYQFenBm4tWSilWhBfyWIY8FljBNIc1K2BW0sWSqmWw1eyGAO8LiLHNkYwgc7/dbi1ZKGUallqTRbGmI3AacDTjRNOYAt3h1NQWkCpx0epweXUcRZKqRbFZ28oY0wKcGYjxBLwwp320qo ShdaslBKtTR dZ01xuQ0dCDNQcX8UMW JhPUEdxKqZbF32VVewF3Al28zzHGjG2guAJSuNv/aco9 b7bNpRSqrnwK1kAnwKvAm8AZQ0XTmArr4bKK/U9TbmO4FZKtST JotSY8x/GjSSZsDvaiiXC7TNQinVgvjVZgF8ISJ/E5E4EWld/mjQyAJQRbLwp2Sh4yyUUi2IvyWLSfZP7zW5DdC9fsMJbHVZWlV7QymlWhK/koUxpltDB9IcRLisZch9juJ26RTlSqmWxd/eUC7gBg6sYfEd8Jox5oj6RqxTyUK7ziqlWhB/2yz gzVP1Cv2Y5i97ZCISCsRmSEiv4nIJhE53m4H UZEttg/Y xjRUReFJFEEVknIkcf6n0PlyvIhcvh8p0snFoNpZRqWfxNFscYYyYZYxbajyuBYw7jvi8A84wxfYDBwCbgbuBbY0xP4Fv7NcAZQE/7cS2HkaTqgz8LIGnJQinV0vibLMpEpEf5CxHpziGOtxCRaKzqrDcBjDHFxpj9wLnANPuwacB59vNzgXeNZRnQSkTiDuXe9cGfBZB0WVWlVEvjb2 oO4FFIrINEKyR3Fcd4j27ARnA2yIyGFgF3AK0N8bsto9JBdrbzzsCO73OT7a37fbahohci1XyoHPnzocYmm/ TFMuLheUlmKMQUQaLBallGos/iaLH7GqgXrbrzcf5j2PBm4yxvwsIi9woMoJAGOMERFTl4saY14HXgcYPnx4nc6tC3 roQBrYJ7b3VChKKVUo/G3GmqpMabIGLPOfhQBSw/xnslAsjHmZ/v1DKzkkVZevWT/TLf37wI6eZ2fYG9rEn5VQ7msHKztFkqplsLXGtwdRGQYECoiQ0XkaPsxGgg7lBsaY1KBnSJSXkoZB2wEZnNg8N8kYJb9fDZwud0r6jggy6u6qtH5XQ0F2m6hlGoxfFVDnQZcgfXX/HNe27OBfx7GfW8CPhARN7ANuBIrcX0iIpOBHcBF9rFzgAlAIpBvH9tk/KmGwqklC6VUy1JrsjDGTAOmichEY0y9rcVtjFkLDK9m17hqjjXAjfV178PlXzWUliyUUi2Lv20WT4vIByJyvYj0b9CIAly4y1qH22M8NR4jTjtZaMlCKdVC Jss gGvAbFYiWOriHzecGEFrvL5ofJLal7cqKJkUawlC6VUy D3oDygxP7pweqplF7rGS1UmMtq16 tKkoq2iw0WSilWgZ/x1lkA79iNXK/YYzJbLiQAps/M8 KW9sslFIti78li0uA74G/AR JyEMiclBj9JHAn5lny0sWaJuFUqqF8Hc9i1nALBHpgzWx363AXUBow4UWmMKcdjVULavlaW8opVRL41fJQkT JyKJWLPFhgGXAzENGViginDb1VC1rMMtOs5CKdXC Ntm8TiwxhhzSDPNtiThTt/rcGvJQinV0vjbZrERuEdEXgcQkZ4iclbDhRW4/OkNRfk4ixItWSilWgZ/k8XbQDFwgv16F/Bog0QU4CqqoWpr4NaShVKqhfE3WfQwxjyFNdYCY0w 1roWRxy3w41TnD6Shd1moclCKdVC JssikUkFDAA9qp5RQ0WVQATEZ/zQ1WULHRQnlKqhfC3gfsBYB7QSUQ AE7Emo32iORr5lmthlJKtTT jrP4RkRWA8dhVT/dYozZ06CRBTCfJQsdlKeUamFqTRYiUnUx61/tn2Ei0tkY80fDhBXYwl3h5Jbk1rhfSxZKqZbGV8niK6x2Cu/GbAO0BdoBQQ0UV0CLcEWQXZxd4/6KQXnadVYp1UL4WvxooPdrEekKTAFOAR5ruLACW5grjN15Na/sqiULpVRL4 90Hz1F5B1gLrAK6GeMeakhAwtkEa6IWquhcOniR0qplsVXm8UA4F6gP/AUMFmn/LBXy6tt8SMRcDq1ZKGUajF8tVn8AuzEarsYAYwQOdB8YYy5ueFCC1zlvaGMMXh/Ht7E6dRxFkqpFsNXspiMPRBPHRDhisBgKCgtqJgrqipxubRkoZRqMXw1cL/TSHE0K94LINWYLJxOHWehlGoxam3gFpE37HaL6vaFi8hVInJZw4QWuPxah1tLFkqpFsRXNdS/gftFZCCwHsgAQoCeQBTwFvBBg0YYgPxah9vp1HEWSqkWw1c11FrgIhGJAIYDcUABsMkYs7nhwwtMfq3DrSULpVQL4u/cULnAdw0bSvPh1wJILqeOs1BKtRj TlGuvJRXQ9U P5RbSxZKqRZDk8UhKK GqnVgno6zUEq1IHVOFiLiEJGohgimuQhz2tVQpdpmoZQ6Mvg7N9SHIhIlIuFYvaI2isidDRta4Ap1huIQB7nFPqYp195QSqkWwt SRT9jTDZwHtZkgt2AvzZUUIFORAh3hpNf6qMaSksWSqkWwt9k4RIRF1aymG2MKeEInwYk3B3us2ShyUIp1VL4myxeA5KAcOB7EekC1Lz6zxHAZ8lCu84qpVoQf8dZvAi86LVph4iMaZiQmodwV7gOylNKHTH8beC xW7gFhF5U0RWA2MbOLaA5msdbpxaslBKtRz VkNdZTdwnwrEYDVuP3E4NxaRIBFZIyJf2q 7icjPIpIoIh LiNveHmy/TrT3dz2c 9YXnwsgaclCKdWC Jssylf4mQC8Z4zZ4LXtUN0CbPJ6/STwvDHmKGAf1loa2D/32duft49rcuULINVEnC4tWSilWgx/k8UqEfkaK1nMF5FIwHOoNxWRBOBM4L/2a8Gq1pphHzINq cVwLn2a z946Sm5ekaka91uB2hoXjyrNX0lFKqufM3WUwG7gaOMcbkA27gysO477 AuziQcGKB/caY8j/Fk4GO9vOOWEu7Yu/Pso vRESuFZGVIrIyIyPjMELzT3k1VE3JwN21C6aggNLU1AaPRSmlGppfycIY4wESgPtE5BngBGPMukO5oYicBaQbY1Ydyvm1xPi6MWa4MWZ427Zt6/PS1QpzhVFmyigqK6p2v7t7dwCKErc2eCxKKdXQ/O0N9QRWG8NG 3GziDx2iPc8EThHRJKAj7Cqn14AWolIeVfeBGCX/XwX0MmOwwlEA5mHeO9642vm2eCjjgKgeJsmC6VU8 dvNdQEYLwx5i1jzFvA6cBZh3JDY8w9xpgEY0xX4GJgoTHmMmARcIF92CRglv18tv0ae/9CEwANAb5mnnW2bk1Qq1YUbd3WmGEppVSDqMuss628nkfXcxwAU4DbRCQRq03iTXv7m0Csvf02rLaTJufPAkjuHj0o2qolC6VU8 fXCG7gcWCNiCzC6jJ7MvXwpW2M Q57BT5jzDZgRDXHFAIXHu696ps/CyAF9 hBzvz5GGMIgA5cSil1yPyd7mO6iHwHHGNvmmKMOaK7 fizAFJwj 7sz8qibO9enLEHdeBSSqlmo9ZkISJHV9mUbP MF5F4Y8zqhgkr8JUni1qrobr3AKBo61ZNFkqpZs1XyeLZWvYZjuD5ocqTRa3VUEdZyaJ42zbCRxxUw6aUUs1GrcnCGHNEzyxbG3 qoZwdOuAIC9OxFkqpZq/Oa3ArS6gzFEFqLVmICO4ePXSshVKq2dNkcYgc4vA5mSBAcPfuWrJQSjV7miwOg6/V8sAaa1Gank5ZTk4jRaWUUvWv1mQhIn/xen5ilX1/b6igmgtf63BD5UZupZRqrnyVLG7zev5SlX1X1XMszU64M5y8Ut/VUKATCiqlmjdfyUJqeF7d6yOOr9XyAFwJCYjLRZE2ciulmjFfycLU8Ly610ccn twA J04u7WjWKdUFAp1Yz5GpTXR0TWYZUietjPsV93b9DImgF/ShYA7h7dKVy/oREiUkqphuErWfRtlCiaKX 6zgIEd 9Bzrz5eAoLcYSENEJkSilVv2qthjLG7PB ALnA0UAb /URrbZ1uPcU7GFdhlUQCz6qBxhD8fbtjRmeUkrVG19dZ78UkQH28zhgPVYvqPdE5NaGDy whbvCKfWUUlxWfNC e3 8l2u voZST6nXhILabqGUap58NXB3M8ast59fCXxjjDkbOBbtOlvjzLObMjfxU8pP5Jfmk5SVhLtbV3A4dNoPpVSz5StZlHg9HwfMATDG5ACehgqquahp5tm3179NkAQBsHHvRhxuN 5OnXSsRYAxxnD1/KuZu31uU4eiVMDzlSx2ishNInI VlvFPAARCQVcDR1coKtu5tmdOTuZv2M l/a9lJCgEDZlbgLsJVa1ZBFQsouz Tn1Z35K ampQ1Eq4PlKFpOB/sAVwJ NMfvt7ccBbzdcWM1DdSWLaRum4RAHk/pNonfr3mzM3AhYS6wWJ 3AlJRUey3V Hbn7QYgJTeliSNRKvD56g2Vboy53hhzrjHma6/ti4wxzzR8eIGtaptFZkEmMxNncnb3s2kf3p5 sf3YtHcTHuPB3aM7lJZSvHNntdcyZWXs  QTynJrH Sn6k95ktiVu6uJI1Eq8PlaVnV2bfuNMefUbzjNS9VqqA9/ 5DismKuGHAFAH1b92V66XSSspOI63FgidXy aK87f/sM1Lvf4CyrCzaXHNN47yBI1x5ySI1L5VSTylOh19L0it1RPL1f8fxwE5gOvAzOh9UJd7VUPkl Xz020eM7TyW7tFWMugX2w wekd16TYagOKtW2H8 ErX8RQWsuflf1vXWvSdJotGkpqXCkCZKSMtP42OER2bOCKlApevNosOwD BAcALwHhgjzFmsTFmcUMHF i8q6Fm/D6D7OJsrhpwoEdx91bdcTvcbMzcSFBEOM64uGrHWux7/31K09IIP EECtasoXTv3kZ7D0cy77aKXTlaFaVUbXy1WZQZY YZYyZhNWonAt/pWhaWMGcYAFlFWby78V2Gtx/OoLaDKva7HC56t 7Npr1Wj6jgHj0o2ppY6RplWVnsef0NwkedTLs7bgdjyF38feO9iSNYal4qCREJgLZbKOWLz5XyRCRYRCYC7wM3Ai8Cnzd0YM1BkCOIUGcos7bOIi0/rVKpolzf1n3ZlGk1cgf36E7xtu0Yz4EhKpn//S enBza3XYbwX374uzQgdyFCxvzbRyxUvJSGNpuKA5xaLJQygdf0328CyzFGmPxkDHmGGPMI8YY/T/LFu4KJz0/nV4xvTip40kH7e8X24/cklySc5Jxd  BKSykJMVqWC1JTWXvu 8Rfc7ZhPTujYgQMWY0uUuW4CkqauR3cmQpLitmT8EeOkV2okNYB 0 q5QPvkoWfwF6ArcAP4lItv3IEZHshg8v8JW3W1w54EpEDm7/7xtrTdy7MXPjgSVW7aqoPf/ N3g8tLnp5orjI8eOxeTnk//zzw0d hEtLS8NgLiIOOIj4rVkoZQPvtosHMaYSPsR5fWINMZENVaQgSzKHUV8eDyndz292v09W/XE6XCyce9G3OVLrG7dRtG2bez/32e0uuRi3AkHeuGEHXssjrAwcrQqqkGl5FklibjwODpGdCQ5N7mJI1IqsGnH8sN077H34nQ4a yj7wpy0bNVTzZmbsQ5LIag1q0p2raVgjVrcISG0ub66ysd73C7CT/pJHIXLsI88EC1pRV1 MrHWJQni4z8DIrLinEHuZs4MqUCk88GblW7/m3607t171qP6Rfbj02ZmzDGENyjB7mLviPnm29ofdWVOFu3Puj4iLFjKE1Pp3DDxoYK 4hXniw6hHegY2RHDKZim1LqYJosGkG/2H5kF2ezK3cX7h7dKcvMJCg2ltgrrqj2 IhRo8Dh0F5RDWh37m7ahLbBHeQmPjwe0O6zStVGk0UjqBjJvXcTwT2OAqDN327AER5e7fHOmBhChw4lZ9GiRovxSLM7b3dFkkiI1LEWSvmiyaIR9IzpiVOcbMzcSNSZE2h3x 3EXHhhredEjh1D0aZNlKRol86GsDtvNx3COwDQNrQtTodTR3ErVQtNFo0gOCiYHq16sClzE87WrYm9 mrEXXtDasSYsQBaumgAxhh25 4mLjwOsAZXxoXH6VgLpWrR6MlCRDqJyCIR2SgiG0TkFnt7axH5RkS22D9j7O0iIi KSKKIrBORoxs75vrQL7YfGzM3Yozx6/jg7t1wd 1K7kJNFvVtb Feij3FxEXEVWzrGNFRq6GUqkVTlCxKgduNMf2w5pu6UUT6AXcD3xpjegLf2q8BzsAaGNgTuBb4T OHfPj6xvZlX9E 0vLT/D4nYuxY8pYv1zUu6pl3t9lymiyUql2jJwtjzG5jzGr7eQ6wCegInAtMsw bBpxnPz8XeNdYlgGtRCSOZqa8kXtD5ga/z4kcOwZKSsj78ceGCuuIVJ4s4iPiK7Z1jOhIZmEmBaUFTRWWUgGtSdssRKQrMBRrrYz2xpjyju6pQHv7eUesNTXKJdvbql7rWhFZKSIrMzIyGi7oQ9QrphcOcVSsye2P0CFDCGrVitwa2i1K0tMp3bevvkI8YpS3TVQtWYDVpVYpdbAmSxYiEgH8D7jVGFNpniljVez7V7l/4JzXjTHDjTHD27ZtW4 R1o9QZyjdo7tXrMldnZKyyutzi9NJxKhR5H63GFNaCkDp3r3smz6dHX/5K4mjRvPHVZP9bgdRltS8VEKdoUS5D8xYU17K0Gk/lKpekyQLEXFhJYoPjDGf2ZvTyquX7J/p9vZdQCev0xPsbc1O Zrc1flq21cc9 FxfJ9ceS2LiDFjKMvKIuOFF/nj6mvYMvJkUh96mNK9e4kYN5aiTZsoWLu2EaJvOcrHWHhPpaJjLZSqXVP0hhLgTWCTMeY5r12zgUn280nALK/tl9u9oo4Dsryqq5qVfrH92FOwh/T89Erbv9z2Jf/88Z8Ue4r5attXlfaFn3QS4naT cYbFCclETt5Mt1mzaT7V1/S8ckncYSFsf TTxvzbTR7KbkpdIjoUGlbbEgswUHB2n1WqRo0xUSCJwJ/BX4VkbX2tn8CTwCfiMhkYAdwkb1vDjABa5W fODKRo22HvVtbU1XvilzE 3C2gFWorj3x3sZ1n4YMcEx/LjrR0o9pRUTEwZFhNP5nXeQIAchgwZV mtYwsOJOvtssmbNov09dxMUpRMB yM1L5UBbQZU2iYiOlW5UrVoit5QPxpjxBgzyBgzxH7MMcZkGmPGGWN6GmNOMcbstY83xpgbjTE9jDEDjTErGzvm tKndR8EqWi3 GLrF9z7470Mbz cl8e zGldTyO7OJtfMn6pdF7Y0UMJHTy42hloY/58EaawkKzZXzTKe2ju8kvy2Ve0r1LjdjlNFkrVTEdwN6IwVxhdo7uyce9Gvtj6BfctuY/h7Yfz0tiXCHOFcUL8CTgdThbvXOz3NUP69SNkwAD2f/yxNnT7ITU/FaBiqg9vCREJmiyUqoEmi0bWL7YfP / uSJRvDzuZcJcYQBEuCMY3n44i5P9TxYArS66kKItWwK6obs4KYlt555H0fbtTRpHaq6VLLzHWJTrGNGRrKIscot1EKRSVWmyaGR9W/eloLSgIlGEOkMr7R VMIptWdvYmb2zhiscLPrMM62G7o8/qe9w682 jz haPPmJm M914hr6ryBKKlC6UOpsmikU3sOZE7h99ZbaIAK1kAdSpdOOyG7uy5cynLyqq3WOuLKS0l6wurTSX7yy8xZWVNFsvuvN04xFHRwcBbQoR2n1WqJposGlmkO5LL 19ebaIA6BTVie7R3etcFRXz54swRUUB2dCd  OPlO3ZQ9Q5Z1OakUHesmVNFktqXirtwtpVuwxueclCu88qdTBNFgFoVKdRrExbWae684qG7k8 CbiG7qzPZxIUE0OH x/AERlJ9uzZTRZLSm5KtVVQAK2CWxHmDNOShVLV0GQRgEYljKLUU8pPKT/VetyO7B1s3b 14nWrP19kNXSvWdvAEfqvbP9 chcuJOqsswiKCCfq9NPJ/mYBnvz8Jolnd97uGpOFiNAxsqNO aEChikp8X1QI9FkEYAGtx1MlDuq1qqonOIcrph3BZfPvZzMgkwAoidMsEd0B05Dd9acOZiSElqdfx4A0eeeg8nPJ2fBgkaPpcxTRlpeWo3JAqBjeEethlIBoWDtWjYPG07BBv9nqm5ImiwCkNPhZGTCSH5I/oEyT/WNwS teYm9hXvJL83n6ZVPA4HZ0J01cxbBvXoR3NcavR569NG44uPJmtX4VVF7CvZQakqr7TZbrmOkta5FoFXlqSPP/lmzMMXF7J8xo6lDATRZBKxRCaPYV7SPX/f8etC DXs28NFvH3Fx74u5ZuA1fLXtK5bsWgLUvaE7tziXdze8y0VfXFRxjfpStHUrhevWEX3  RWjz8XhIOqcs8lbupSS9HQfV6hf5etYVDcgr1x8eDx5JXlkF2fXeIxSDc2UlZHzjVX6zpkzF1Nc3MQRabIIWCd2PJEgCTqoKqrMU8bDyx4mNjSWvw/9O1cPvJquUV15ZNkjFJQWVDR0Z77 Orv/737S//Uv9r73Ptlz55L383KKtm/HlJWRkpvC0yue5pQZp/D0yqdJyk7iwaUPkl9Sf20JWTNnQlAQ0WefVWl79DnngsdD9ldz6u1e/qhuhbyqOkZa61pou4VqSvkrV1G2Zw/REydSlpVF7vff z6pgTXFRILKD1HuKI5ufzSLkxdzy9G3VGz/5PdP2Ji5kadOfopIdyQA9x9/P1fNv4pXf3mVfwz7B21v jvpzz1PzqJFlO3dCx5Pxflb4uGr490s6 lBEMaGDmHSsMmURkdwxfwreO2XV7m552TK9u2jbP9 yrKyCO7TB1f79gfFWBtTVkbWrNlEnHQSzjZtKu0L7t6NkIEDyZo9m9grrzj0D6mO/EoW9iJIKbkp9I/t3yhxKVVVzvx5SEgI7e 5m9zFi8ma/QWRp5zSpDFpsghgoxJG8czKZ0jJTSE Ip6M/AxeXP0ix8cdz ldT6847pgOxzCx50SmbZjGhG4T6D1qFBGjrMF9pqzM sskbRdP//YKs7N/JNwTxHlbYzjl20za7F0BrCC0dWvGjHIyrewtBtzyBgmZB IQt5uYv/6FNtdcQ1CrVn7FnvfTUkrT04n 5z8rthljKqqjos85h7SpUyn8/XdCevU67M/KHym5KUS6I4lwR9R4THmy2JVTt 6zRYmJ7Jv EW1vuVln/1WHxZSVkf31N0SMGkVQZCRRZ05g//SPKMvKIig6usni0mqoAFZ1NPfTK5 muKyYe4 796AZaG8bdhvRwdE8tPShSo3iEhTENtnDpM338UX2EiYPmMzCv/zIQ498x4nfraLrJx/T/r77iBgzmutDTyNUgnnv2m7EPfE4nV57lS7vvUvUhAnsfettEk89jT1vvIGnsNBn7FkzZ KIjiZi7BjrPexczOhPRlcs7hQ14QwICmrUMRepeanEh9fcuA3WoMkod1SdqqFKMzL449pr2ffBB6Q  ujhhklhaWG9Vgeq5iV/lVUFFXX6aYBVbWtKSsieNx ArKIsTvn0FL7Y2rgDcDVZBLCu0V3pEtWFxcmLWZqylLnb5zJ54GS6RHU56Njo4GimHDOFX/f8ykebPwKsv Q/2fwJl351KVlFWbw2/jVuHXZrxcSFDreb0EGDaP2Xy4ifOpV 9z/BP06Ywhp2smSQi4hRowg75hjin3icbjNnEjZ0KBnPPsfW005n/4wZFUu9VlWWk0POggVEnzkBh9tN4r5E7vr LvYW7uXu7 /mj w/cMbGEnHSSWR98SXGq5rMH/krVrBv vQ6n1fbGAtvHSP87z7rKShg5w1/o2zffqLPPYfs2V QPXduneLyZozh gXXc8W8K7RH1hEqZ/7XSHAwESefDEBI/364e/Qgy/7DambiTNLy03hnwzuN m9Ek0WAG5UwiuW7l/PoskfpHNmZyQMn13jsGd3O4MT4E3lx9Yv8vu93bl98O48se4Th7Ycz45wZHB9/vM/7/annn gf259nVj5DTnFOxfaQ3r0qShquDh3Yfd//se3Ms9j7wQd48vIqXSN77lxMURHR553H/sL93LTwJsJcYbx7xrs4HA5uWXQL SX5RJ97DqVpaeQvX 7355H74xL mHw1qQ89zM4bbqhTF Hdubtr7QlVrmNER79GcZuyMlLuuovCDRvo OyzxE2dSsigQex 8CFK0tL8jsvb/KT5rEpbxaa9m/g59edDuoZqvozHQ87XXxNx8sk4wsMBa7Bo9DnnULBqFQU7d/DRbx8REhTC7/t Z92edY0WmyaLADcqYRQlnhL yPmDe4 7l Cg4BqPFRHuO 4 PMbDhV9cyKI/FnHbsNt45ZRXaBPapsbzvAU5gvi/4/6PzIJMXln7ykH7w445hi4fTafjSy/iaBVN2iOPsmX0GNKeepqSXdYXbNbMWbi7dyeofx/uWHwHaflp/GvMvxjabihPjXyKrfu38uDSBwkfMwZHRITfYy7yli0j cYbcXfvTru77iLvp6Vsv BCCjdvBiA9P53ZW2fjMQeXOHKKc8gpyal1jEW5 Ih4UnJTfP7Vlv7Ms R8s4D299xD5NgxiNNJx6eexBQXs/uef9a55FNYWshzq56jd0xvWoe05r2N79XpfNX8FaxeTWlGBpF2FVS58h6FC a8QnJuMvccew9hzjA 2dx4A3A1WQS4oe2HEhsSyxndzuCE BN8Hp8QmcCdx9xJr5heTDtjGlcOuBKH1O3X3L9Nfy7qfREf/vYhv 397aD9IkLU PF0 /hjun40nYiRJ7F32jQSx5/KzutvoGD1aqLPP49nVz7Lz6k/88DxDzC47WAATuh4AjcNvYm52 cyffv/iDztVHLmz8dTUFBrTPkrVrDzhr/h7tyZzm /RexVV9Ll3WmYoiKS/nwx62a yWVzLuPeH /l/Y3vH3S Pz2hynWM6EhhWSGZhZk1HrNv nT2vv02MX/5C60v/2vFdnfXrrSfMoW8n35i3wcf ryXt3c3vsvuvN1MGTGFP/f M98nf8/2rKZd/6MhaPVazbLnzUeCg4kcPbrSdld8PGEjRjBj77e0DW3L2T3O5qzuZzE/aT5ZRY0zAFeTRYBzOVx8fu7nTD1pqt/nXNT7Ij49 1MGtR10yPe9aehNtApuxaPLHq32L/VyoUOG0PG55zhqwTfETr6K/DVrEJeLhUOdfPjbh1ze73LOPercSudMHjiZMZ3G8OzKZ9k2rg e/HzSHnuc4uTqG5XzV6/mj uuxxUfT d33sYZEwNA2NChdPvfDP44oRvX736ewux9DGt3NC sfoEt 7ZUukZqnrXoUVyE72SREFn7VOW5339P6iOPEjF6NO3vuRuwGh3/9/v/KCorotWfLyJi1CjSn3mGosREn/cDq1T031//yymdT2GIowsTW4/F5XDxwaYP/Dq/ufAUFrLj4kvYftGf/f5sjhQHqqBGVlRBecs  0RWxxdxfvTJuBwuLux9IUVlRY3W0K3JohmICYnB5XA16j2jg6O5bdht/JLxC9M2TPP516ArLo52t99Oz0ULyZr FI9vepET40/kH8P cdCxDnEw9aSpJEQmcF/Gm5RecDr7//c/to4/lT uvZachQsr1rwoWLuWnddci6tdOzq//RbO2NhK11pevIX7Tkwm3B3Jw2/kccsHuYR7XExZdAfFZQdGve7O9b9kUd5jqrpG7vxVq9h16z8I7tObjs8 gwQFkV Szw0LbuDBpQ9yzdfXWGt8P/oIjrAwUu6a4tfo25fWvERpWSl/Xeoicdwp7D/zIsYVdWdW4qxG 8uxoRljSH3wIQrWraN4xw62T/wTmW 93aTrmwSSgjVrKE1PJ/LU06rd/1V8OkFlMG6lNblgn9Z9GNRmEJ/83jgzTWuyUDU6p8c5jE4YzXOrnuP2xbf79aW1rSiFuzY8TkJEAk NeqradSPA6qL6/OjnyS/N58kTMuiyYB5tbriBok2/kfy3G0k8ZTzpzz7LH1dfQ1BsLJ2nvYOrXeUFi Zun8uNC2 kc2RnPrxkNkff9wwRKfu49tNstuRs49GHTiXjlVco3LSJlNwUnA4nse4YSjMzKdqyhbxlP5M9b/5BS71Wt2Je4ebf2Xnj39lx2V9wtIqm03/ gyM8nOKyYm5ddCsbMzdyRf8r2Ji5kcu uow/3DnEPfIwhRs3kvHvg9t vK3bvoxZW2ZyxvJSQj cS6uJE4kcP55x0zZQWFbI9O9e9Pm5Nwf7P/qIrJkzaXPjjfT46kvCR44k/amn2DFpEsU7/V8ZsqXKnj8fcbuJGDPmoH35JfnM3jGHkdntcX65qGI22gt6XcD2rO2sSlvV4PFJS6w/HD58uFm5cmVTh9EilHnKmLZxGi teYnWIa2ZetJUjos77qDjdubs5NVfXuXLbV8S4Yrg/Qnv0y26m8/rz0uax52L7 SYDsfw2EmP0d4dS86iRez/6CPyflqKKyHB6oEVV7lE8OGmD3li RMc3f5oXhz7IlFuayCcMYaizZt5eMmDfBm0gQc LKPfH4YXLwxlS9tSXvpPKVTzbz64b1 izjiDqAln4E5IYNTHoxjbeSz3xE0i46WXyZ4zB0dEBLFXXUnMXy8nKCKcMk8ZU36Ywvyk Txy4iOcd9R5rMtYx00Lb6LEU8Lzo5 n08uzyfp8JhEnn4wrPg5XfDzOuDhccfE428Syf bn/D33TXa18vDOztPo8rdbcXfqBEDOokVct x2dkUU80HOxcTf8o9qqydq48nPR0JDDxqX01CMMWzP2s7CnQtZmrKUC3tdyOndTid/zRp2XD6J8BOOp9N//oM4HBhjyJo5i7SpUzEeD 3vuotWf76o0WINJMbjIXHMWEIGDKDTv18 aP8nmz/hkWWP8Fr7W4m59RkSXnmFyLFjKCgtYNwn4zgp4SSeOvmpw45DRFYZY4ZXu0 ThfLHxsyNTPl CknZSUzqN4mbj74Zd5Cb1LxUXlv3GjO3zCTIEcQlfS7hygFX0jqktd/X/nzL5zy /HGcDif3H39/xej04uRkgiIiKo0aT8lN4bV1r/HZls8Y02kMT538FCHOkIOumV Sz0VfXkRRSSFvO67ijoxXcYqT54vPI6h1LM7WMQS1jsURGUHBqlVkfzWHgl9 ASBk4EDuPTObJE86Fy4sZvwmF 0u/Suxk6 qiMUYw9Sfp/Lx5o 5fdjtXDHgiop778rdxd //TtJWUncd/QUjn9vLYWbN1OSkoInu/IEhct6C89NDOKfR93AJSf 7aD3sWjLPG7 6U5unuVhzP4O1gjxNm0QlwtxuqyfLid4PBTvTKY4Kcl6bN9OcVISZfv3EzpkCG1uuJ7wk09ukC/iMk8Z6/asY EfC1m0cxE7sncAEBMcQ05xDs8Nf4SE659GgoPpNuPTg0Yhl6SkkHLvveQvXUb4SScR9/BDuOJ991prSfJXr2HHpZcS//RTRJ99dqV9xhgmzp6Iy Hio9PeJ3H0GMJGjCDhX88D8PjPj/Pp75 y4MIFdfr/rjqaLFS9KCgt4NmVz/Lx5o/pFdOLoe2G8tmWzzAYLux1IdcMvIa2YW0P6do7sndwzw/38OueXzmnxzncM KeStNy7MrdxRvr3mDW1lkIwqV9LuXWYbfWWM0F8GvGr/x17l85o9sZrEhdwbFxx9baUaBk1y6y580j 6s5JKZt5M3TnWzoDJ3DErhtxB2M7Ty24sv232v/zau/vMqVA67ktmG3HXStnOIc7lh8Bz l/MRVA67iqgFXER0cTVluLiUpKZTu3k3u7l1MMv8lIjyGT876hCBH0EHX8RgP5848l7ASYeo7pZRs8907ytmuHe6uXXF37UpQbGuyZs2iNGU3wf360ub664k85RTEUT810CtSVzDl ylkFGTgdDgZ0WEEYzuNZXSn0YS7wrly3pVs37OZ//vUwWnPf0RInz7VXsd4POybPp30Z59DgLZ33E7MxRfXW5yBLu3xx9k3/SN6/rSEoIjK09GsSF3BVfOv4uETHub8nueT uhU9n/yCT1//IGgqCgS9yVy/uzzuW3YbVw54MrDikOThapXi3cu5v6f7ierKIvzjjqPawdd69f4BV9KPCW8vu51Xl/3OnHhcTwx8gliQ2P576//ZXbibESEP/X8E5MHTvZrcB3Af375T8V4kesGXcffh/7dr/OKk5PB5WJZyWaeW/kcW7O2MrTdUG4ffjvr96znieVPcP5R5/PQCQ/V Nd6qaeUJ5Y/wcebPwasZVu7RnWtGJmfkpvCp79/yn9P/S/Hxh1bYywf/fYRU3 eyrRT3qTv3lBMcTGmpBRTUoIpLamov3YnJODq3IWgiHCMMaTlp/FH9h9EO8KJ/mEduW 8Q8mOPwjueRSx111PSP9 lKalUZqWRklqGqVpqZSkpUNZGa6EBNydEnAlJOBK6IQ7oSOO8HBMSQll2dmUZWWzJPkH7kz6F3GOGK7t9hdGDz2PqLCYSrH/9vRD3OT8lLzYMN49 0OOijnKx e i9T7/4 8n5YSNnw4cY8 grtrV79 Z U8hYXkr1iJJz8fR1gojtBQJDQUR2gYjrBQglq3xuF21 ma/iratp3sOXMQp5OQAQMIHdDf53xqxuMhcew4Qvr1o9Mr/z5o/23f3cby1OUsuGABIc4QCn79laQLLyKkf38iTxlH MiTuW7nk2QWZPLF V/Uuau8N00Wqt5lFWVRUFrg95d2XaxNX8vdP9zN7rzdCEKQBHFBrwu4asBVtA v2 y3pZ5SJs2bxLqMdTx0wkNM7DmxzvGUekqZmTiTl9e8XDH2YmynsTw7 tlaSzZgVSGsSF3Bpr2bSMpOIikriR3ZO8goyABgTKcxvDi29gbs/JJ8TplxCsfFHcdzo587aH9haSGr0laRuD RxP2JbNu/ja1ZW8krqTyyvkNYBzoWh9P2tzTaJ2UxMMnQOePAfkdUlDW7sMNByc6dBy19K6GhGHs8zLLewgvnOuicAfd VEZUAUhICCG9exPSvz8h/fvjKSgg7dFHKbz8HG47ajkOcfD Ge/77L5sjCHrs89Ie8Ia4Nj25ptoPWkS4qz5sy5JTSX3u8XkfvcdecuWYWqbvywoCHeXLgT37ImjZ3eSuoTwa1Q2aa48okNa0Sq4FdHB0bQKtp63DmlNXERcjT0SS/ftI3vuXLJmzqJw3ToQqdQu5urcmdABAwgZOBBXx3gwWPuNB4yhJDWN9KeeIv7JJ4g t3I389S8VE7/3 lc3u9ybht W8Xns  998maPZvC9esBWHJsJC MLeD5qMmMGXOF3xN VqXJQjU7ucW5vPrLqwBc3v9y2oW183FGzf7I/oP7ltzHQyc85Feje03yS/KZtmEau/N2 xxN70tucS47c3bSJapLxVxdtXlu1XNM2zCNORPn0DGiI2WeMpanLuerbV x4I8FFYkhNiSWHq16WI/oHnSO6kxWcRY7snawI9t6bM/eXjGVy/CwvlzW7QJG952AM/xA9YcxhrL9 ylJTqZk506KdyZTtncvjqhIFkQm84T5in7BXXmh5xSiwlpTlJhI4foNFGxYT9HGTRWJJnToULpMe4ffc7dz5bwriQ2N5d0z3iUmJObgN1lFSVo6qY88TO6Cb3F37467c2ckNARHSCiO0BAkJBSMIe/nZRRt3ASAKyGBiDFjiBg1CmebWDz5BXgK8jGFhXjyC8jPy KXzF9Yk7WRdUG7 T22iGKXVTKMzDfkhQieav4wd3qE NIIOpsYujjb0SW0I52d7SlbsYa8lSugtBRXly6EjxlN2MiTyKaAXVvXsmvXZtL27SStcA Z7iJicg1/WuKhd5UhPI6wMI5a/B1BkZGVtr 05iXeWPcGcybOqRj/4600M5O8H39k7/eLuKznt/RP8nD//pPp9Op/fH6 1dFkoVQzV/4X5mldTyM2NJZ52 eRUZBBhCuCU7qcwuldT6d/bH9ahbTyeS1jDJmFmcxKnMWHv31Ien46XaO6clnfyzinxzm1Jq/yXjnHdjiWF8e WO2xxuOhOCmJot9/J zYYysGUa5KW8V131xHz1Y9efO0N/1KksYYcubNY9 H0/Hk5eEpLMRTUIApKMBTWIgpKSF0yBAix4wmYvRo3D16VKoWNMawZf8Wftr1Ez m/MjqtNWUeEpwiIPeMb05OnYwg8ri6ZvuImJ3FmUFeeQV5rC/KIv9pTlkleWy1 SS7MxmZ2gBuyJLSG1lKAvyv6NA65DWtA9rT1tnDOv3bmBvaRYntjqaGzpfQp oniBCUKtWOFtXbpzOLMhk4uyJDGoziJfGveTzPk8vf5oPNn3AF/1foNPwUX7H502ThVItwF2L72Ju0lxcDhcjO47kzO5ncnLCydX2BvNXiaeEBTsW8N7G9/h1z69EuiMZ22ks8RHxtA1rS/uw9rQLa0e7sHbMTpzNs6ueZXTCaJ4Z/cwhlawW/bGIf3z3D6LcUQxuN5jBbQczqM0gBrQZUCl5FJYWWqWgrO1sz9rO7rzdBDmCcDlcBx5BLpziJMgRhFOcOMRBkCOIILE6Cqzfs56lKUtJL7CW7z2q1VGcEH8Cx8Ydy9B2QysWD6sLYwzFeTnsSN3Ejsyt0DYWR5BdPSYg9n9RwVEVn5076ED7SH5JPh/ 9iFvr3 b7OJsxncZz41DbqRHqx4UlBawOm01y3YvY9nuZRVT7fhq0yqXlJXE2TPP5uahN3PNoGvq/N5Ak4VSLcKegj2sSF3BCfEnEB1cv4vgGGP4JeMX3t/0PqvSVpFZkInh4O G07uezmMjHzusGQV 2vUTX23/inUZ60jKTgKsUf09W/WkTWgbkrKTrIkc7fsLQpvQNhgMJZ4SSspKrJ eklrvE WO4vj44zkx/kSOjz  QdrXDlVOcQ7vbXyPdze S35JPn1j 7Jl3xZKPCW4HC6GthvKcXHHcVLHk gb29fv616/4Hp6RPfgzmPuPKS4NFkopeqkxFNCZkEmaflppOenk56fjjvIzcSjJlbbxfdQ7S/cz697fmXdnnX8kv4L 4v20zW6K92iu1mPqG50iepSbenJGEOpp5QyU4bHeCg1pXg8HspMGWWmjNiQ2HqNtSHsK9zH2xveZm36Woa0HcJxcccxtP1QQp2hh3Q9j/Fob6i60GShlFJ1V1uyODJGvCillDosmiyUUkr5pMlCKaWUT80mWYjI6SKyWUQSReTupo5HKaWOJM0iWYhIEPBv4AygH3CJiPRr2qiUUurI0SySBTACSDTGbDPGFAMfAef6OEcppVQ9qX0WtMDREfBeSisZqDSkUUSuBa61X aKyObDuF8bYM9hnN8YNMb60xzibA4xQvOIU2OsWZeadjSXZOGTMeZ14PX6uJaIrKypr3Gg0BjrT3OIsznECM0jTo3x0DSXaqhdQCev1wn2NqWUUo2guSSLFUBPEekmIm7gYmB2E8eklFJHjGZRDWWMKRWRvwPzgSDgLWPMhga8Zb1UZzUwjbH NIc4m0OM0Dzi1BgPQYucG0oppVT9ai7VUEoppZqQJgullFK GWP0YT A04HNQCJwdwPd4y0gHVjvta018A2wxf4ZY28X4EU7nnXA0V7nTLKP3wJM8to DPjVPudFDlQ1VnuPGmLsBCwCNgIbgFsCNM4QYDnwix3nQ/b2bsDP9rU/Btz29mD7daK9v6vXte6xt28GTvP1b6Kme9QSaxCwBvgygGNMsn8na4GVAfo7bwXMAH4DNgHHB1KMQG/78yt/ZAO3BlKMh/zdVZ8Xa84PrP ZtwLdATfWF1C/BrjPycDRVE4WT2H/Tw7cDTxpP58AzLX/QR0H/Oz1j2Kb/TPGfl7 j2 5fazY555R2z1qiDGu/B8tEAn8jjXNSqDFKUCE/dyF9cV4HPAJcLG9/VXgBvv534BX7ecXAx/bz/vZv 9grC/Yrfa/hxr/TdR0j1pivQ34kAPJIhBjTALaVNkWaL/zacDV9nM3VvIIqBirfKekYg10C8gY6/TdVZ8Xa84PrL9Q5nu9vge4p4Hu1ZXKyWIzEGc/jwM2289fAy6pehxwCfCa1/bX7G1xwG9e2yuOq kefsY7CxgfyHECYcBqrJH9ewBn1d8rVm 64 3nTvs4qfq7Lj upn8T9jnV3qOG2BKAb4GxwJe1nd9UMdrHJHFwsgiY3zkQDWzH/ks6EGOsEtepwJJAjrEuD22zOKC6KUU6NtK92xtjdtvPU4H2PmKqbXtyNdtru0etRKQrMBTrr/aAi1NEgkRkLVbV3jdYf2XvN8aUVnPtinjs/VlA7CHEH1vLParzL AuwGO/ru38pooRwABfi8gqe/ocCKzfeTcgA3hbRNaIyH9FJDzAYvR2MTDdx/lNHaPfNFkEGGP9WWAC4R4iEgH8D7jVGJN9KNc4HP7cwxhTZowZgvXX wigT0PGVFcichaQboxZ1dSx OEkY8zRWLM73ygiJ3vvDIDfuROrCvc/xpihQB5WdUsgxQiAPXj4HODTQzn/cDXEPTRZHNCUU4qkiUgcgP0z3UdMtW1PqGZ7bfeoloi4sBLFB8aYzwI1znLGmP1YjfLHA61EpHzAqfe1K Kx90cDmYcQf2Yt96jqROAcEUnCmi15LPBCgMUIgDFml/0zHfgcK/kG0u88GUg2xvxsv56BlTwCKcZyZwCrjTFpPs5v8v93/KXJ4oCmnFJkNlbPB yfs7y2Xy6W44Asu5g5HzhVRGJEJAarbnS vS9bRI4TEQEur3Kt6u5xEPvcN4FNxpjnAjjOtiLSyn4eitWusgkraVxQQ5zl174AWGj/BTYbuFhEgkWkG9ATqxGx2n8T9jk13aMSY8w9xpgEY0xX /yFxpjLAilG /MLF5HI8udYv6v1BNDv3BiTCuwUkd72pnFYPfYCJkYvl3CgCqq285syxrqpzwaQ5v7A6pnwO1a9970NdI/pwG6gBOsvpclY9cvfYnV5WwC0to8VrEWftmJ1lRvudZ2rsLrOJQJXem0fjvU/ VbgZQ50q6v2HjXEeBJWEXYdB7oATgjAOAdhdUddZ1/rfnt7d6wv0kSsaoBge3uI/TrR3t/d61r32rFsxu5dUtu/iZru4eN3P5oDvaECKkb72F840A353tp H034Ox8CrLR/5zOxegoFWozhWCW7aK9tARXjoTx0ug llFI aTWUUkopnzRZKKWU8kmThVJKKZ80WSillPJJk4VSSimfNFkoVQ0RybV/dhWRS v52v s8vqn ry Ug1Bk4VStesK1ClZeI2arkmlZGGMOaGOMSnV6DRZKFW7J4CRIrJWRP5hT1z4tIisEJF1InIdgIiMFpEfRGQ21qhiRGSmWJPybRB7Yj4ReQIIta/3gb2tvBQj9rXXi8ivIvJnr2t/JyIzROQ3EfnAHr2LiDwhIhvtWJ5p9E9HHTF8/QWk1JHubuAOY8xZAPaXfpYx5hgRCQaWiMjX9rFHAwOMMdvt11cZY/baU5GsEJH/GWPuFpG/G2vyw6omYo1QHgy0sc/53t43FOgPpABLgBNFZBNwPtDHGGPKpz5RqiFoyUKpujkVay6ftVjTtsdizdMEsNwrUQDcLCK/AMuwJoXrSe1OAqYbaybdNGAxcIzXtZONMR6s6Ve6Yk1fXgi8KSITgfzDfG9K1UiThVJ1I8BNxpgh9qObMaa8ZJFXcZDIaOAUrIWMBmPNYRVyGPct8npehrWoUSnWzLAzgLOAeYdxfaVqpclCqdrlYC0tW24 cINYU7gjIr3sWVqrigb2GWPyRaQP1jKY5UrKz6/iB DPdrtIW6wleJfXFJhY641EG2PmAP/Aqr5SqkFom4VStVsHlNnVSe9grUXRFVhtNzJnAOdVc9484Hq7XWEzVlVUudeBdSKy2ljTlZf7HGs9jl wZv29yxiTaieb6kQCs0QkBKvEc9shvUOl/KCzziqllPJJq6GUUkr5pMlCKaWUT5oslFJK aTJQimllE aLJRSSvmkyUIppZRPmiyUUkr59P9oJznt6X 6AgAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "total_steps = len(PGNN_noisefree['fold1']['tr_mse'])\n",
    "x_1 = range(total_steps)\n",
    "x_2 = x_1[::len(PGNN_noisefree['fold1']['tr_mse']) // len(PGNN_noisefree['fold1']['vd_mse'])]\n",
    "\n",
    "up_limit=1000.\n",
    "\n",
    "figure(figsize=(6, 4))\n",
    "\n",
    "\n",
    "\n",
    "plt.plot(x_2, DDNN_noisefree['fold1']['vd_mse'], c='tab:red', label='Unc. Noise free')\n",
    "plt.plot(x_2, PGNN_noisefree['fold1']['vd_mse'], c='tab:green', label='Con. Noise free')\n",
    "\n",
    "# plt.plot(x_2, DDNN_noise['fold1']['vd_mse'], c='tab:blue', label='Unc. noise')\n",
    "# plt.plot(x_2, PGNN_noise['fold1']['vd_mse'], c='tab:cyan', label='Con. noise')\n",
    "\n",
    "\n",
    "plt.ylim(0.0, up_limit)\n",
    "plt.xlabel('Iterations')\n",
    "plt.ylabel('MSE loss (Newton^2)')\n",
    "plt.title('Learning curve Comparison')\n",
    "plt.legend()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX4AAAEWCAYAAABhffzLAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8qNh9FAAAACXBIWXMAAAsTAAALEwEAmpwYAAAoZElEQVR4nO3deZxU9Znv8c/T1SsN1Sy2AuKAgisoKIsybgSXRGLiEjOKmmgyGZM4Y2LGqDF6zXLdZ KdzDUTw1VHjYaQYFwSjTqYGJW4sAgIIiqIEdmaXfbuquf cU4V1U11d0F3Lc35vl v86pTZ/n9nqrT/ZxTv3PO75i7IyIi0VFW7ABERKSwlPhFRCJGiV9EJGKU EVEIkaJX0QkYpT4RUQiRolfcmZmJ5vZomLHUSrM7EQze8/MNpvZucWOB8DMHjSzWzqpLDOz/zaz9Wb2RmeUuQd1/9HMLitknVGixN9FmNlSMzu9mDG4 8vufngxYygxPwbucffu7v5Ey5nhNtsW7hhWhUm5e6GCM7NxZrasA0WcBJwBDHD3MZ0U1m7M7Idm9kjmNHc/y90fyledUafEL2lmFit2DB1V4M8wEFjQzjKfc/fuwHHAKOCmvEfVeQYCS919S7EDkc6lxN/FmVmZmX3PzBab2Voz 42Z9c6Y/1szW2lmG83sJTMbmjHvQTP7uZk9Y2ZbgE FR6nfNbN54TpTzKw6XL7ZEWRby4bzrzOzFWa23My ZmZuZkNa Ry9w2aF5WHTwhPh9MvN7JUWy6bLyfIZvht 3ljG8ueZ2bxcvq8scf2Tmb1vZuvM7Ckz6x9OXwwcAvw PKKvams7ufvHwB BYeH6J5jZX81sg5nNNbNxGXW aGb/28ymm9knZva8me2XMb/VbZqxTG1YX/8wvs1m1t/MtppZn4zljjOzBjOraLH PwL3AWPDdX U47b4mZk9Hcb9upkNzlh2qJn9T/hdrjKz75vZZ4DvAxeG9czN A6 lrHNbjKzD81stZk9bGZ14bxBYQyXmdnfzGyNmd3Y1rYQJf59wVXAucCpQH9gPfCzjPl/BA4F9gdmA4 2WP9i4FagB5D6p/4H4DPAwcAxwOVt1J912fAf l B04EhwLh2PscvgW7A0DDW/9PO8q19hp8CW4DxLeb/Khxv7/tKM7PxwO0En7Ef8CHwawB3Hwz8jfCI3t13tBWgmR0ETADeNLMDgaeBW4DewHeBx8ysvkXMXyH4LirDZVLa26aER lnAcvD Lq7 3LgxfDzpHwJ LW7N7ZY/37gG8Cr4bo/aOvzZbgI BHQC3ifYLtgZj2AacCzBN/7EOAFd38WuA2YEtYzPEuZl4fDpwh2tt2Be1oscxJwOHAacLOZHZljvNHk7hq6wAAsBU7PMn0hcFrG 35AI1CeZdmegAN14fsHgYez1HNpxvu7gHvD8XHAshyXfQC4PWPekLDuIVni6gckgV5Z5l0OvNJiWrqcVj7DLcAD4XgPgh3BwL34vu4H7sp43z1cdlBb26TF97MZ2ECw0/gvoAa4Hvhli2WfAy4Lx18EbsqYdyXwbCt1ZNumt2TbXuG0C4Hp4XgMWAmMaaXsZt99jtvivox5E4B3wvGJwJut1PND4JEW014EvhaOvwBcmTHv8NQ2AwaFMQzImP8GcFE /x 7 qAj/q5vIPB42GSwgSCxJYADzCxmZneEzRqbCBIRwH4Z63 UpcyVGeNbCRJea1pbtn LsrPVk3IQsM7d17exTFtalv0r4Pyw eV8YLa7fxjOa/X7ylJuf4KEDYC7bwbWAgfuQWznuntPdx/o7le6 7Ywhi mYgjjOIlgJ5SS9XvNcZu25UngKDM7mODE7UZ378wrdlr7ezgIWLyXZTbbDuF4Oc232Z78zUZeebEDkA77CPiqu09vOcPMvgScQ9DcshSoI2jasIzF8tU96wpgQMb7g9pY9iOgt5n1dPcNLeZtIWgCAsDM mZZv9lncPe3zexDgqaOzGaeVF1Zv68slhMk6VTdtUAf4OMc1m3LRwRH/P 0F teTPvbNGW3bevu283sN8ClwBEETWy5ymVbtOYjgmagbNr7G2y2HYC/A5qAVTT/G5Mc6Yi/a6kws qMoRy4F7jVzAYCmFm9mZ0TLt8D2EFwlNqNoC21UH4DfMXMjjSzbsD/am1Bd19B0G79X2bWy8wqzOyUcPZcYKiZjbDgxPEPc6z/V8C3gVOA32ZMb v7amly BlGhL8ebgNed/elOcbQmkeAz5nZp8Mj GoLTpznksT2ZJuuAvqkToRmeJig2ebz7Fni39ttAfAHoJ ZXW1mVWbWw8yOz4hzkJm1lo8mA98xs4MtuBw2dU6gaQ/qlwxK/F3LM8C2jOGHBCcznwKeN7NPgNeA1D/UwwQ/iz8G3g7nFYS7/xH4T DPBCf5UnW3dhL0SwTttu8Aq4Grw3LeJbhefhrwHrtOQLdnMsEJ3D 5 5qM6W19Xy0/wzSCHdZjBL9gBtP6UWvO3P0jgqP27wMNBEfD15Lb/2PO29Td3yH4HpaETUr9w nTCc6pZDaB5RL33m4L3P0TgqalzxE0y7xHcLIWdu2Y15rZ7CyrP0Cwg3oJ ADYTnCSXvaShSdDRPIqvMpiPlClI7XiM7M/Ab9y9/uKHYsUno74JW8suH6 ysx6AXcCv1fSLz4zG01wQ9mUYscixZG3xG9mD4Q3W8zPmNY7vIHjvfC1V77ql5LwdYJmm8UEV858s7jhiJk9RNBUc3XY/CIRlLemnvDk3GaCa6xTdyveRXDZ3h1m9j2C67avz0sAIiKSVV7b M1sEPCHjMS/CBjn7ivMrB/woqvTLxGRgir0dfwHhJfuQXBmP9tNMwCY2RXAFQC1tbUjjzjiiAKEJyKy75g1a9Yad69vOb1oN3C5u5tZqz833H0SMAlg1KhRPnPmzILFJiKyLwhvZNxNoa/qWRU28RC ri5w/SIikVfoxP8UkHqqzmUE/YaIiEgB5fNyzsnAq8DhZrbMgv697wDOMLP3CPoauSNf9YuISHZ5a N394mtzDotX3WKyN5pbGxk2bJlbN  vdihyF6orq5mwIABVFRUtL8w6p1TRIBly5bRo0cPBg0ahFm2jj6lVLk7a9euZdmyZRx88ME5raMuG0SE7du306dPHyX9LsjM6NOnzx79WlPiFxEAJf0ubE 3nRK/iEjEKPGLSElYuXIlF110EYMHD2bkyJFMmDCBd999t1PrePDBBykrK2PevHnpacOGDWPp0qVtrjdhwgQ2bNjQobobGho4/vjjOfbYY3n55Zc7VFZHKfGLSNG5O eddx7jxo1j8eLFzJo1i9tvv51Vq1Z1el0DBgzg1ltv3aN1nnnmGXr27Nmhel944QWOPvpo3nzzTU4  eRm8xKJRIfK3lNK/CJSdH/ 85 pqKjgG9/4Rnra8OHDOfnkk3F3rr32WoYNG8bRRx/NlCnBYwRefPFFxo0bxwUXXMARRxzBJZdcQi6dTp599tksWLCARYsW7TZv8uTJHH300QwbNozrr9/VcfCgQYNYs2YNW7Zs4bOf/SzDhw9n2LBh6VhmzZrFqaeeysiRI/n0pz/NihUrmpU7Z84crrvuOp588klGjBjBtm3b6N69O9dccw3Dhw/n1Vdf5ZFHHmHMmDGMGDGCr3/96 mdwfPPP8/YsWM57rjj OIXv8jmzZv3/AtuQZdzikgzd75xJ  se6dTyzyi9xFcP6b1Htjnz5/PyJEjs8773e9 x5w5c5g7dy5r1qxh9OjRnHJK8EjmN998kwULFtC/f39OPPFEpk fzkknndRmLGVlZVx33XXcdtttPPTQQ npy5cv5/rrr2fWrFn06tWLM888kyeeeIJzzz03vcyzzz5L//79efrppwHYuHEjjY2NXHXVVTz55JPU19czZcoUbrzxRh544IH0eiNGjODHP/4xM2fO5J577gFgy5YtHH/88fzkJz9h4cKF3HnnnUyfPp2KigquvPJKHn30USZMmMAtt9zCtGnTqK2t5c477 Tuu /m5ptvbvsLb4cSv4iUtFdeeYWJEycSi8U44IADOPXUU5kxYwbxeJwxY8YwYEDwjPoRI0awdOnSdhM/wMUXX8ytt97KBx98kJ42Y8YMxo0bR3190JnlJZdcwksvvdQs8R999NFcc801XH/99Zx99tmcfPLJzJ8/n/nz53PGGWcAQbNNv3792o0hFovxhS98AQiagWbNmsXo0aMB2LZtG/vvvz vvfYab7/9NieeeCIAO3fuZOzYsTl8a21T4heRZto6Ms XoUOHMnXq1D1er6qqKj0ei8VoasrtyZ7l5eVcc8013HnnnXtU32GHHcbs2bN55plnuOmmmzjttNM477zzGDp0KK   uoelVVdXU0sFgOCcxyXXXYZt99 e7Nlfv/733PGGWcwefLkPSq7PWrjF5GiGz9 PDt27GDSpEnpafPmzePll1/m5JNPZsqUKSQSCRoaGnjppZcYM2ZMh u8/PLLmTZtGg0NDQCMGTOGv/zlL6xZs4ZEIsHkyZM59dRTm62zfPlyunXrxqWXXsq1117L7NmzOfzww2loaEgn/sbGRhYsWLBHsZx22mlMnTqV1auDDovXrVvHhx9 yAknnMD06dN5//33gaB5qDOudFLiF5GiMzMef/xxpk2bxuDBgxk6dCg33HADffv25bzzzuOYY45h PDhjB8/nrvuuou ffu2Wd7NN9/MU0891eYylZWVfOtb30on2379 nHHHXfwqU99iuHDhzNy5EjOOeecZuu89dZb6ROwP/rRj7jpppuorKxk6tSpXH/99QwfPpwRI0bw17/ dY8 /1FHHcUtt9zCmWeeyTHHHMMZZ5zBihUrqK v58EHH2TixIkcc8wxjB07lnfe6fj5l7w erGz6EEsIvm1cOFCjjzyyGKHIR2QbRua2Sx3H9VyWR3xi4hEjBK/iEjEKPGLCEBONz9JadrTbafELyJUV1ezdu1aJf8uKNUff3V1dc7r6Dp EWHAgAEsW7YsfWmjdC2pJ3DlSolfRKioqMj56U3S9ampR0QkYpT4RUQiRolfRCRilPhFRCJGiV9EJGKU EVEIkaJX0QkYpT4RUQiRolfRCRilPhFRCJGiV9EJGKU EVEIkaJX0QkYpT4RUQiRolfRCRilPhFRCKmKInfzL5jZgvMbL6ZTTaz3J8ZJiIiHVLwxG9mBwLfAka5 zAgBlxU6DhERKKqWE095UCNmZUD3YDlRYpDRCRyCp743f1j4N BvwErgI3u/nzL5czsCjObaWYz9QBoEZHOU4ymnl7AOcDBQH g1swubbmcu09y91HuPqq vr7QYYqI7LOK0dRzOvCBuze4eyPwO DvixCHiEgkFSPx/w04wcy6mZkBpwELixCHiEgkFaON/3VgKjAbeCuMYVKh4xARiaryYlTq7j8AflCMukVEok537oqIRIwSv4hIxCjxi4hEjBK/iEjEKPGLiESMEr ISMQo8YuIRIwSv4hIxCjxi4hEjBK/iEjEKPGLiESMEr ISMQo8YuIRIwSv4hIxCjxi4hEjBK/iEjEKPGLiESMEr ISMQo8YuIRIwSv4hIxCjxi4hEjBK/iEjEKPGLiESMEr ISMQo8YuIRIwSv4hIxCjxi4hEjBK/iEjEKPGLiESMEr ISMQo8YuIRMw nfjdncZEY7HDEBEpKe0mfjP7tpnFLXC/mc02szMLEVxHXfnClVz 3OXFDkNEpKTkcsT/VXffBJwJ9AK BNyR16g6SW1FLZt2bCp2GCIiJSWXxG/h6wTgl 6 IGPaXjGznmY21czeMbOFZja2I W1Jl4ZZ9NOJX4RkUzlOSwzy8yeBw4GbjCzHkCyg/X FHjW3S8ws0qgWwfLy6quqo5NOzbh7ph1aF8lIrLPyCXx/yMwAlji7lvNrDfwlb2t0MzqgFOAywHcfSewc2/La0u8Mk6TN7GtaRvdKvKybxER6XJyaeoZCyxy9w1mdilwE7CxA3UeDDQA/21mb5rZfWZW23IhM7vCzGaa2cyGhoa9qiheGQdg446OhCsism/JJfH/HNhqZsOBa4DFwMMdqLMcOA74ubsfC2wBvtdyIXef5O6j3H1UfX39XlUUrwoSv9r5RUR2ySXxN7m7A cA97j7z4AeHahzGbDM3V8P308l2BF0urrKOkCJX0QkUy6J/xMzu4HgMs6nzawMqNjbCt19JfCRmR0eTjoNeHtvy2tL ohfl3SKiKTlkvgvBHYQXM /EhgA/FsH670KeNTM5hGcOL6tg VllW7j36k2fhGRlHav6nH3lWb2KDDazM4G3nD3jrTx4 5zgFEdKSMXqcSvI34RkV1y6bLhH4A3gC8C/wC8bmYX5DuwzlBbUUvMYmrjFxHJkMt1/DcCo919NYCZ1QPTCE7KljQz0927IiIt5NLGX5ZK qG1Oa5XEuJVcV3HLyKSIZcj/mfN7Dlgcvj QuCZ/IXUuXTELyLSXC4nd681sy8AJ4aTJrn74/kNq/PEq Js3K4jfhGRlFyO HH3x4DH8hxLXsQr43y06aNihyEiUjJaTfxm9gng2WYB7u7xvEXVieKVcV3HLyKSodXE7 4d6ZahZMQr43yy8xOSnqTMusw5aRGRvNnnM2FdVR1JT7KlcUuxQxERKQn7fOJP372rK3tERIAoJP4q9ckvIpIply4brjKzXoUIJh90xC8i0lwuR/wHADPM7Ddm9hnrYg vrasK  RXR20iIkAOid/dbwIOBe4neE7ue2Z2m5kNznNsnUJH/CIizeXUxh8 gWtlODQBvYCpZnZXHmPrFEr8IiLNtXvnrpl9G/gysAa4D7jW3RvDJ3G9B1yX3xA7pqa8hvKycp3cFREJ5dJlQ2/gfHf/MHOiuyfDB7OUNHXNLCLSXC5NPYe0TPpm9ksAd1 Yl6g6WV1VnU7uioiEckn8QzPfmFkMGJmfcPJDR/wiIru0mvjN7Iawo7ZjzGxTOHwCrAaeLFiEnSBeqYexiIiktJr43f32sKO2f3P3eDj0cPc 7n5DAWPssHiVjvhFRFLa6pb5uHD0txnjae4 O29RdbK6yjolfhGRUFtX9fykjXkOjO/kWPImXhVn887NJJIJYmWxYocjIlJUbfXH/6lCBpJP8co4jrO5cXO6CwcRkajK6dGLZjYMOAqoTk1z94fzFVRnS9 9u2OTEr IRF4ud 7 ABhHkPifAc4CXgG6TOJPd9Smdn4RkZyu478AOA1Y6e5fAYYDXeqwOXXEr2fviojklvi3uXsSaDKzOMF1/AflN6zOldnUIyISdbm08c80s57A/wNmAZuBV/MZVGdLPYVLTT0iIjkkfne/Mhy918yeBeLuPi /YXUudc0sIrJLrlf1HAgMTC1vZqe4 0v5DKwzVZdXUxWrUlOPiAi5XdVzJ3Ah8DaQCCc70GUSP4T99ejkrohITkf85wKHu/uOPMeSV/HKuI74RUTI7aqeJUBFvgPJt7oq9dcjIgK5HfFvBeaY2QtA qjf3b Vt6jyIF4ZZ XWlcUOQ0Sk6HJJ/E FQ6cKH gyE/jY3fP CMd4VZxF6xfluxoRkZKXy WcD5lZDfB37t6ZmfPbwEIg3olltkpP4RIRCbTbxm9mnwPmAM G70eYWYd AZjZAOCzwH0dKWdPxCvjbGncQlOyqVBVioiUpFxO7v4QGANsAHD3OcAhHaz3P4DrgGRrC5jZFWY208xmNjQ0dLC6XXfvfrLzkw6XJSLSleWS BvdveUF8K0m7PaY2dnAanef1dZy7j7J3Ue5 6j6 vq9rS4t3VGbnr0rIhGXS JfYGYXAzEzO9TM/i/w1w7UeSLweTNbCvwaGG9mj3SgvJyoa2YRkUAuif8qYCjBpZy/AjYCV 9the5 g7sPcPdBwEXAn9z90r0tL1fqr0dEJNDWw9argW8AQ4C3gLHu3mXPjKZ76NTduyIScW1dzvkQ0Ai8TPDUrSPpwJF Nu7 IvBiZ5bZGh3xi4gE2kr8R7n70QBmdj/wRmFCyo 6yqCNXyd3RSTq2mrjb0yNdOUmnpSKWAU15TU64heRyGvriH 4maWypAE14XsD3N0LcsdtZ9LduyIibSR d48VMpBCiFepa2YRkVwu59xn6GEsIiIRTPxq6hGRqItU4q rqlNTj4hEXqQSv474RUQimPi3NW2jMdHY/sIiIvuoaCX sNsGneAVkSiLVOJP3b2r5h4RibJIJX511CYiErXEr47aRESimfjVUZuIRFmkEr ewiUiErHE36OyB6DELyLRFqnEX15WTm1FrU7uikikRSrxg 7eFRGJZuLXEb IRFjkEn9dVZ2O EUk0iKX OOVcV3OKSKRFr3EX6U2fhGJtuglfp3cFZGIi1zir6uqY0diB9ubthc7FBGRoohc4ld/PSISddFN/LqkU0QiKrqJX0f8IhJRkUv86qhNRKIucolfXTOLSNRFL/FXqalHRKItcom/e0V3QIlfRKIrcok/VhajR2UPXdUjIpEVucQPuntXRKItsolfJ3dFJKoKnvjN7CAz 7OZvW1mC8zs24WOQR21iUiUFeOIvwm4xt2PAk4A/tnMjipkAHWV6pNfRKKr4Inf3Ve4  xw/BNgIXBgIWOIV kpXCISXUVt4zezQcCxwOtZ5l1hZjPNbGZDQ0On1huvjLNx50bcvVPLFRHpCoqW M2sO/AYcLW773b47e6T3H2Uu4 qr6/v1LoHxgfSlGxiycYlnVquiEhXUJTEb2YVBEn/UXf/XaHrH913NABvrHyj0FWLiBRdMa7qMeB YKG7313o gEGdB9A/9r vLFCiV9EoqcYR/wnAl8CxpvZnHCYUMgAzIzRfUczY9UMkp4sZNUiIkVXjKt6XnF3c/dj3H1EODxT6DiO73c8G3ds5N317xa6ahGRoorknbuwq53/9RW7XVAkIrJPi2zi71vbl4HxgTrBKyKRE9nEDzCm7xhmrZpFU7Kp2KGIiBRMtBN/vzFsadzC22vfLnYoIiIFE nEP/oAXc8vItET6cTfp6YPQ3oO0fX8IhIpkU78EFzW ebqN9mZ2FnsUERECiLyiX9039FsT2xnXsO8YociIlIQkU/8ow4YhWHMWDmj2KGIiBRE5BN/XVUdR/Y5ktdX6kYuEYmGyCd CK7nn9swl21N24odiohI3inxEyT pmQTc1bPKXYoIiJ5p8QPHHfAcZRbua7nF5FIUOIHaitqGbrfUF3PLyKRoMQfGtN3DAvWLmDzzs3FDkVEJK U EPH9zuehCeYvXp2sUMREckrJf7Q8PrhVJRVqH9 EdnnKfGHqsurGbH/CN3IJSL7PCX DGP6juGdde wZtuaYociIpI3SvwZxv/deGJlMb723NdYsXlFscMREckLJf4Mh/U6jF c/gtWbV3Fpc9cyqJ1i4odkohIp1Pib2FMvzE8dNZDYHDZs5fx2orXih2SiEinUuLP4rBeh/HohEfpV9uPb077Jn9Y8odihyQi0mmU FvRt7YvD531EMfufyw3vHwD9711H 5e7LBERDpMib8N8co4955 L2cNOoufzv4pX/7jl3l 6fM0JZuKHZqIyF6zrnAUO2rUKJ85c2bR6k96kimLpvDwgodZtnkZ/Wr7MfGIiZx/6PnUVdUVLS4RkbaY2Sx3H7XbdCX 3CWSCf6y7C88svARZqycQU15DZ8f/HnOHHgmQ3oNoXd172KHKCKSpsTfyRatW8QjCx/h6SVP05hsBKB3dW8O7XkoQ3oNYUjPIQzuOZiB8YH0quqFmRU5YhGJGiX PNm4YyML1i7g/fXv8/6GXUPm07zilXEGxQcxMD6QgfGBHNTjIPar2Y/e1b3pXdObuso6YmWxIn4KEdkXKfEXUNKTLN 8nCUbl/Dhpg/5cNOHLN20lKUbl7Jq66rdli zMnpW9aRPTR/2r9mf/bvtGg7odgD13erZmdjJ u3rWb9jPeu2r2P99vVs2LGBmvIaDqk7hME9BzO452D6VPfRrwsRAVpP/OXFCGZfV2ZlDOgxgAE9Buw2b2vjVpZvXs667etYt30da7ev3TW bS0NWxt4d/27rNm2Bqf1nXJNeQ09q3qyeedmPmn8JD09XhlncM/B9KvtR3V5NZVllVTFqqiMVVIZC8bTQ3kV1bHq9PtYWYwyKwsGyigrC16rYlXUlNdQU15DdXmwvHYuIl2XEn BdavoxpBeQ9pdrinZxJpta1i9dTUNWxuoKq iV1UvelUHQ015DQDuTsO2BhZvWMySjUtYsmEJizcuZm7DXBoTjexI7mBnYic7EjtIerJTPkOZlVEdq6YiVkHMdu0sUuPNXsOdR5mVUV5WTnlZORVlFc1ey8vKSSQTNCYbaUw20pRsCsYTjZSXlVNbUUu3im50K 9GbUUttRW1AGxt2sqWxi3Nhp2JnemdVGpH1XK8Olad3oHVlNek4zWz3XZ6ZVaW3iFmflaj Y7PzDBst89YbsH7zPINC8YpS6/X3o7U3Ul6Ml2GSEeoqSdCmpJN7EjsCIamHenx7Ynt7EzsJOEJkskkSZIkPUkimSDpyfQy25q2NRsaE43Bcp7I poa0tOTiXRSTyf3cDyVLCtiFc12CE3JJrY27krwW5u2srVxK45TWx7sEGoraule0Z1uFd2ojFWyvWn3WLc1bWN70/Y2f0UVW ZOAQcnSPYtYy6zsvQOJXOoKNu1Iy4vKydmsWCn1WIHYwQ7D8eDHQpJcJrVFbNYsx1Vs517WVB/qp7UzhKj2fKpWDN3mM0Gmr9vtuPNMi09nlGHYc120i0PQlqun/m5WpaZWq irIJYWYyYxdLfY8t1sn0nmQcKMYul1ysmNfVIOkGkjpi7qtTByp7 U7k7O5M72d60PRgSwWvCE kj6oQncJxEMnwNd4aZO7VU/akkmXpNepKmZFPzwYPXVIJNehJ3T5fdMsFnHtmnfxlkJFacYGfpzetpTDam60/FmpqXKjcz0bv7bjsCjPQvmVQcqWVT5e3wHSSSCRKeaFZP6vOnPkvq/W4DyWYHF6kh8/vZl6R2ALGy2K5ffjT/lWdYepnMnU3q/d3j7s7abNwRSvzS5eztUZSZpc9n6Ma70pVtZ5D5Pr1TSk1n104781dm6ldMap3U/GzlOJ7eaSaSifSONbWTa7ljbllm1l 33tT8NdnU7FdVemecij 1Q81YJ EJKsoqOv07LkriN7PPAD8FYsB97n5HMeIQkdKTakKR/Cn4t2tmMeBnwFnAUcBEMzuq0HGIiERVMXarY4D33X2Ju 8Efg2cU4Q4REQiqRhNPQcCH2W8XwYc33IhM7sCuCJ8u9nM9vZxWPsBpf4Q3a4QI3SNOLtCjNA14uwKMULXiLNYMQ7MNrFkT 66 yRgUkfLMbOZ2S5nKiVdIUboGnF2hRiha8TZFWKErhFnqcVYjKaej4GDMt4PCKeJiEgBFCPxzwAONbODzawSuAh4qghxiIhEUsGbety9ycz BXiO4HLOB9x9QR6r7HBzUQF0hRiha8TZFWKErhFnV4gRukacJRVjl iyQUREOo/ukhARiRglfhGRqHH3fXYAPgMsAt4HvpenOh4AVgPzM6b1Bv4HeC987RVON A/w3jmAcdlrHNZuPx7wGUZ00cCb4Xr/Ce7muey1tFKjAcBfwbeBhYA3y61OIFq4A1gbhjjj8LpBwOvh VOASrD6VXh /fD YMyyrohnL4I HR7fw t1dHG9xkD3gT UMIxLg23xxxgZqlt73DZnsBU4B1gITC2BGM8PPwOU8Mm4OpSi3OP81ZnFVRqA8E/52LgEKCSIKEclYd6TgGOo3niv4vwnxb4HnBnOD4B GP4x3EC8HrGBl4SvvYKx1N/SG Ey1q47llt1dFKjP1Sf4BAD Bdgu4ySibOcL3u4XgFQZI7AfgNcFE4/V7gm H4lcC94fhFwJRw/KhwW1cRJMvF4d9Cq38PrdXRxvf5r8Cv2JX4SzHGpcB LaaVzPYO5z8EfC0cryTYEZRUjFlyykqCm6JKNs6cPktnJsFSGgiOHp7LeH8DcEOe6hpE88S/COgXjvcDFoXjvwAmtlwOmAj8ImP6L8Jp/YB3Mqanl2utjhzjfRI4o1TjBLoBswnu6F4DlLfcpgRXhY0Nx8vD5azldk4t19rfQ7hO1jpaiW0A8AIwHvhDW sXK8ZwmaXsnvhLZnsDdcAHhEe3pRhjlpjPBKaXepy5DPtyG3 2riEOLFDdB7j7inB8JXBAOzG1NX1Zlult1dEmMxsEHEtwRF1ScZpZzMzmEDSd/Q/B0e8Gd2/KUm46lnD RqDPXsTep406svkP4Dog1XF8W sXK0YAB543s1lh9ydQWtv7YKAB G8ze9PM7jOz2hKLsaWLgMntlFEKcbZrX078JcGD3bWXQh1m1h14DLja3TftTRkd0V4d7p5w9xEER9VjgCPyGc eMrOzgdXuPqvYseTgJHc/jqAX3H82s1MyZ5bA9i4naCL9ubsfC2whaM4opRjTwptNPw/8dm/L6IjOrmNfTvzF7BpilZn1AwhfV7cTU1vTB2SZ3lYdWZlZBUHSf9Tdf1eqcQK4 waCk9FjgZ5mlrrRMLPcdCzh/Dpg7V7EvraNOlo6Efi8mS0l6FV2PMFzJUopRgDc/ePwdTXwOMGOtJS29zJgmbu/Hr6fSrAjKKUYM50FzHb3Ve2UUew4c7IvJ/5idg3xFMEZfMLXJzOmf9kCJwAbw59yzwFnmlkvM tF0Jb4XDhvk5mdYMFjp77coqxsdewmXPd YKG7312KcZpZvZn1DMdrCM5BLCTYAVzQSoypci8A/hQeFT0FXGRmVWZ2MHAowcmzrH8P4Tqt1dGMu9/g7gPcfVC4/p/c/ZJSijH8/mrNrEdqnGA7zaeEtre7rwQ MrPDw0mnEVx1VjIxtjCRXc08bZVR7Dhz01knC0pxIDjD/i5BW/GNeapjMrACaCQ4ivlHgjbZFwguw5oG9A6XNYKH0CwmuHxrVEY5XyW4nOt94CsZ00cR/NMuBu5h16VeWetoJcaTCH4mzmPXZWkTSilO4BiCSyTnheXcHE4/hCApvk/wM7sqnF4dvn8/nH9IRlk3hnEsIrxCoq2/h9bqaGe7j2PXVT0lFWO47Fx2XRp7Y1vbohjbO1x2BDAz3OZPEFztUlIxhsvXEvzqqsuYVnJx7smgLhtERCJmX27qERGRLJT4RUQiRolfRCRilPhFRCJGiV9EJGKU GWfZ2abw9dBZnZxJ5f9/Rbv/9qZ5YvkgxK/RMkgYI8Sf8bdsq1plvjd/e/3MCaRglPilyi5AzjZzOaY2XfCTuH zcxmmNk8M/s6gJmNM7OXzewpgrtJMbMnLOjwbIGFnZ6Z2R1ATVjeo G01K8LC8ueb2ZvmdmFGWW/aGZTzewdM3s0vGMTM7vDzN4OY/n3gn87EhkFf9i6SBF9D/iuu58NECbwje4 2syqgOlm9ny47HHAMHf/IHz/VXdfF3YnMcPMHnP375nZv3jQsVxL5xPcmToc2C9c56Vw3rHAUGA5MB040cwWAucBR7i7p7qvEMkHHfFLlJ1J0K/KHIJuqvsQ9JsD8EZG0gf4lpnNBV4j6GzrUNp2EjDZgx5HVwF/AUZnlL3M3ZME3WcMIuiyeTtwv5mdD2zt4GcTaZUSv0SZAVe5 4hwONjdU0f8W9ILmY0DTid4qMpwgj6FqjtQ746M8QTBA1aaCHrQnAqcDTzbgfJF2qTEL1HyCcGjJ1OeA75pQZfVmNlhYW WLdUB6919q5kdQfCYvJTG1PotvAxcGJ5HqCd4ROcbrQVmwbMS6tz9GeA7BE1EInmhNn6JknlAImyyeZCgL/1BwOzwBGsDcG6W9Z4FvhG2wy8iaO5JmQTMM7PZHnTRnPI4wfME5hL0jHqdu68MdxzZ9ACeNLNqgl8i/7pXn1AkB qdU0QkYtTUIyISMUr8IiIRo8QvIhIxSvwiIhGjxC8iEjFK/CIiEaPELyISMf8fd2V2Foblb5YAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "up_limit=10.\n",
    "figure(figsize=(6, 4))\n",
    "\n",
    "\n",
    "plt.plot(x_2, PGNN_noisefree['fold1']['vd_pen'], c='tab:green', label='Con. Noise free')\n",
    "\n",
    "plt.ylim(0.0, up_limit)\n",
    "plt.xlabel('Iterations')\n",
    "plt.ylabel('Penalty loss')\n",
    "plt.title('Learning curve of Penalty function')\n",
    "plt.legend()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **Testing**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def test(ts_dataset, model, criterion, device, lim=35., preds=None, targets=None):\n",
    "    ts_loader = DataLoader(ts_dataset, config['batch_size'], shuffle = False, \n",
    "    drop_last=False, num_workers=0, pin_memory=True)\n",
    "    sample_num = len(ts_dataset)\n",
    "    hookF = [Hook(layer[1]) for layer in list(model._modules.items())]\n",
    "\n",
    "    if preds is None or targets is None:\n",
    "        model.eval()\n",
    "        preds, targets, jtlayers = [], [], []\n",
    "        total_loss = 0\n",
    "        total_ts_mse = 0\n",
    "        total_ts_pen = 0\n",
    "        cnt = 0\n",
    "        for x, y in ts_loader:\n",
    "            x, y = x.to(device), y.to(device)\n",
    "            with torch.no_grad():\n",
    "                pred = model(x)\n",
    "                preds.append(pred.detach().cpu())\n",
    "                targets.append(y.detach().cpu())\n",
    "\n",
    "                jtlayer = hookF[2].output\n",
    "                jtlayers.append(jtlayer.detach().cpu())\n",
    "                ts_loss, ts_mse, ts_pen = criterion(pred, y, x, jtlayer)\n",
    "\n",
    "            cnt = cnt 1\n",
    "            total_loss = total_loss    ts_mse.detach().cpu().item() * len(x)   ts_pen    # ( *len(x) means error**2)            \n",
    "            total_ts_mse = total_ts_mse   ts_mse.detach().cpu().item() * len(x)\n",
    "            total_ts_pen = total_ts_pen   ts_pen\n",
    "\n",
    "        total_loss = total_loss / len(ts_loader.dataset)            # Calculate validation loss with each epoch (Criterion within CLASS model)\n",
    "        total_ts_mse = total_ts_mse / len(ts_loader.dataset)\n",
    "        total_ts_pen = total_ts_pen / len(ts_loader.dataset)\n",
    "        preds = torch.cat(preds, 0).numpy().reshape((sample_num, 3))\n",
    "        targets = torch.cat(targets, 0).numpy().reshape((sample_num, 3))\n",
    "        jtlayers = torch.cat(jtlayers, 0).numpy().reshape((sample_num, 9))\n",
    "    return total_loss, total_ts_mse, total_ts_pen, preds, targets, jtlayers"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **Testing with trained NN**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Finished reading the Trajectory Dataset (10000 samples found, each dim = 9)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Wilson Chen\\AppData\\Local\\Temp\\ipykernel_18228\\4245920301.py:60: DeprecationWarning: Call to deprecated function remove_sheet (Use wb.remove(worksheet) or del wb[sheetname]).\n",
      "  wb.remove_sheet(wb.active)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9992027833018557 0.999794452318768 0.999811041425236\n",
      "total_loss : 29.728252, mse : 29.712104, pen : 0.016155\n",
      "0.9966855903085988 0.999399599712413 0.9991714756419465\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgMAAAGDCAYAAAC2gxMSAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8qNh9FAAAACXBIWXMAAAsTAAALEwEAmpwYAACD6ElEQVR4nOzdd5hbxfXw8e Rtvfe11733jtgXMCmGWx6DYaEDim/kIRACC2BAElIXgIhIQk19G6qscEUGzDuva/b9t67NO8fkh1j7K2Srnb3fJ5HD7u3zBzJi3R078wZMcaglFJKqd7LZnUASimllLKWJgNKKaVUL6fJgFJKKdXLaTKglFJK9XKaDCillFK9nCYDSimlVC nyYBSyutEJEtEjIgEuH//UEQWdqKdPiJSIyJ2z0fpGyKyT0ROtToOpY6kyYBSXuJ 0693f3gdeqT5sP9ZIrJMRCpFZJ v m0PY8wZxphn2zru6A9OY8wBY0yEMcbh3Qh9Q0TuEZH/Wh2HUpoMKOVdZ7s/vA498jpy8qFv0p1UCzwF/LILbRxTF Pyez39 Sl1NE0GlPIxEQkWkb KSJ778VcRCXbvmykiOSJym4gUAE LiF1E7hCRPSJSLSJrRCTTffxQEVkiImUiskNELjrUjzHmW2PM80B2O2I6dBn/OndM SLyiyP23yMir4vIf0WkCrhKRKJF5D/uY3NF5PeHLt 7Y/6TiJSISDZw1lH9fSYi1xzx 7Uiss39/LaKyHgReR7oA7zrvqryq2PcbkgTkUXu579bRK49KuZXReQ5d7tbRGRiK6 BEZGbRWQXsMu9bZ6IrBeRChH5SkRGH3H8be7nXe1 7U9xb39GRH5/xHEzRSTnGP2dDtwBXOx fhva ndSyls0GVDK934DTAXGAmOAycCdR xPAeKAvsB1wM BS4EzgSjgh0CdiIQDS4AXgSTgEuDvIjK8C7HNAgYBc4Hbjrq3PR94HYgBXgCeAVqAgcA49zmHPuCvBea5t08ELjhehyJyIXAPcKX7 Z0DlBpjfgAc4H9XVx4 xukvAzlAmruPB0Rk9hH7z3EfEwMsAh5r4/kvAKYAw0VkHK4rK9cD8cA/gUXuZG4IcAswyRgTCZwG7Guj7e8wxnwEPAC84n5 YzpyvlKepMmAUt71tvtbZYWIvO3edjlwnzGmyBhTDNwL/OCIc5zA3caYRmNMPa4P2DuNMTuMywZjTCmuD9t9xpinjTEtxph1wBvAhV2I915jTK0xZhPwNK4k5JCvjTFvG2OcuD60zwR 5j6 CPgLroQE4CLgr8aYg8aYMuAPrfR5DfCwMWaV /ntNsbsbytQ99WRE4HbjDENxpj1wL9xJRWHLDfGfOAeY/A8ruSrNX8wxpS5X/frgH8aY1YaYxzuMQ6NuBI5BxCMK2kINMbsM8bsaStmpfyV3hdTyrsWGGOWHrUtDTjyw26/e9shxcaYhiN zwSO9UHTF5giIhVHbAvA9aHXWQePimvUcfb1BQKBfBE5tM12xDFpx2jreI73/NqSBpQZY6qP6ufIWwEFR/xcB4SISIAxpuU4bR79HBeKyI P2BYEpBljPheRn G6ojFCRBYDP /omBCl/IVeGVDK9/JwfdAc0se97ZCjlxI9CAw4RjsHgc NMTFHPCKMMTd2IbbMdsZ1ENe35IQj o4yxoxw788/RlvHc7znd3SfR8sD4kQk8qh cls5py1HP8f7j3p9w4wxLwEYY140xpyE69/SAA 5z6sFwo5oJ6Wd/SllGU0GlPK9l4A7RSRRRBKAu4DWppf9G/idiAwSl9EiEg 8BwwWkR ISKD7MUlEhgGIiE1EQnB9gxcRCRGRoDZi 62IhInICOBq4JVjHWSMyQc Bv4sIlHuvgaIyAz3Ia8CPxGRDBGJBX7dxvP7hYhMcD /gSJyKFkqBPofJ4aDwFfAH9zPbTTwI1p/LTviX8ANIjLFHVe4iJwlIpEiMkREZotr4GcDUI/r9g7AeuBMEYkTkRTgZ630UQhkiYi FytL6R gUr73e2A1sBHYBKx1bzueR3B9uH4MVAH/AULdl8fn4rpPn4frkvhDuO5lA5yM60PqA1zfmOvdbbTmc2A38AnwJ2NMa8dfieuy VagHNfgwlT3vn8Bi4EN7uf35vEaMca8BtyPayBkNfA2rgGU4BprcKd7zMUvjnH6pUAWruf/Fq6xFkfflukUY8xqXAMhH8P1/HYDV7l3BwMPAiW4Xvck4Hb3vudxPe99uF7vYyZUbq 5/1sqIms9EbdSnSHG6FUqpXo7EckC9gKBrdxPV0r1UHplQCmllOrlNBlQSimlejm9TaCUUkr1cnplQCmllOrlNBlQSimlerleW4EwISHBZGVlWR2GUkop5RNr1qwpMcYkHmtfr00GsrKyWL16tdVhKKWUUj4hIsctC663CZRSSqleTpMBpZRSqpfTZEAppZTq5XrtmAGllFLH19zcTE5ODg0NDW0frPxKSEgIGRkZBAYGtvscTQaUUkp9T05ODpGRkWRlZSEiVoej2skYQ2lpKTk5OfTr16/d5 ltAqWUUt/T0NBAfHy8JgLdjIgQHx/f4Ss6mgwopZQ6Jk0EuqfO/LtpMqCUUsovFRYWctlll9G/f38mTJjAtGnTeOutt3zW/759 xg5cuQx9z3zzDPk5eV1uM3169fzwQcfHP79nnvu4U9/ lOnY/QUTQaUUkr5HWMMCxYs4OSTTyY7O5s1a9bw8ssvk5OT853jWlpaLImvtWTA4XAc97yjkwF/ocmAUkopv/Ppp58SFBTEDTfccHhb3759 fGPf8wzzzzDOeecw zZsznllFMoKytjwYIFjB49mqlTp7Jx40bg 9 6R44cyb59 9i3bx/Dhg3j2muvZcSIEcydO5f6 noA1qxZw5gxYxgzZgyPP/74MWN7/fXXWb16NZdffjljx46lvr6erKwsbrvtNsaPH89rr73GzJkzD1e5LSkpISsri6amJu666y5eeeUVxo4dyyuvvALA1q1bmTlzJv379 fRRx/1yuvZFp1NoJRSqlU/ 9nPWL9 vUfbHDt2LH/961 Pu3/Lli2MHz/ uPvXrl3Lxo0biYuL48c//jHjxo3j7bff5tNPP XKK69sM95du3bx0ksv8a9//YuLLrqIN954gyuuuIKrr76axx57jJNPPplf/vKXxzz3ggsu4LHHHuNPf/oTEydOPLw9Pj6etWvXAvCPf/zje cFBQVx3333sXr1ah577DHAlbBs376dZcuWUV1dzZAhQ7jxxhs7NC3QE/TKgFJ LCcnh2 //Ran02l1KEpZ6uabb2bMmDFMmjQJgDlz5hAXFwfA8uXL cEPfgDA7NmzKS0tpaqqqtX2 vXrx9ixYwGYMGEC /bto6KigoqKCk4  WSAw22218UXX9yh4w8566yzCA4OJiEhgaSkJAoLCzvVTlfolQGl/NR7777LLy67jIK6Ok5ZsIDXXnsNm03zd V7rX2D95YRI0bwxhtvHP798ccfp6Sk5PA38fDw8DbbCAgI E4ifeR0u Dg4MM/2 32w7cJjufqq69m3bp1pKWlHfee/5ExHdl3W9P8jo7FinEQ s6ilB8qKy5m/09/wvmX92Hm74dR4ljJk/960uqwlPKZ2bNn09DQwBNPPHF4W11d3TGPnT59Oi 88AIAn332GQkJCURFRZGVlXX4sv3atWvZu3dvq33GxMQQExPD8uXLAQ63CfD0009/Z/BfZGQk1dXVx20rKyuLNWvWAK4xBoe0dZ5VNBlQyg999NOfsefsMBZNs9EcFUrZubE8 8Ujlo2cVsrXRIS3336bzz//nH79 jF58mQWLlzIQw899L1j77nnHtasWcPo0aP59a9/zbPPPgvA eefT1lZGSNGjOCxxx5j8ODBbfb79NNPc/PNNzN27FiMMcc97qqrruKGG244PIDwaL/4xS944oknGDduHCUlJYe3z5o1i61bt35nAKE/kNaerBVE5I/A2UATsAe42hhT4d53O/AjwAH8xBiz2L39dOD/AXbg38aYB9vqZ LEiebQSE l/ImjuZlXThnFgz8M5PLhV3Dr4Ou5/s8zWJPl4Ofxv LKBVdaHaLqBbZt28awYcOsDkN10rH /URkjTFm4rGO98crA0uAkcaY0cBO4HYAERkOXAKMAE4H/i4idhGxA48DZwDDgUvdxyrVLa166im mmQn0ARw09ibCIiJ4WcDfgQCr63RWwVKKc/zu2TAGPOxMebQtdBvgAz3z/OBl40xjcaYvcBuYLL7sdsYk22MaQJedh rVLe0653X HqYcHrW6UQGRQIw/IIfMXkH5KWVU1tfa3GESqmexu SgaP8EPjQ/XM6cPCIfTnubcfbrlS3VBCYR1OgsGDYeYe32SPCmVSXRVOYjWeWPWVhdEqpnsiSZEBElorI5mM85h9xzG AFuCF47fU4X6vE5HVIrK6uLjYU80q5TG5W7dycIAQ0mxjbNLY7 w7bdbVRNYZPt32tiWxKaV6LkvqDBhjTm1tv4hcBcwDTjH/G GYC2QecViGexutbD 63yeBJ8E1gLDDgSvlZWtfepH1/YVRwUMItH23Aln63NMZdd9drO1XhNM4sYm/X9hTSnUXfvdu4p4Z8CvgHGPMkZNKFwGXiEiwiPQDBgHfAquAQSLST0SCcA0yXOTruJXyhJ3bvqY8Ujht7PeHvdgjwskqCqchDLYUbbEgOqVUT V3yQDwGBAJLBGR9SLyDwBjzBbgVWAr8BFwszHG4R5seAuwGNgGvOo Vqlup8hWAMCEjCnH3D88ajQAb337qs9iUsoKx1o 2BvL/T7zzDPYbLbDixvB/xY0as2ZZ55JRUVFl/ouLi5mypQpjBs3ji //LJLbXWV35UjNsYMbGXf/cD9x9j AeB/a0Iq1QENDQ2UJzsJabLTP6b/MY8ZMess0rO/YX31Vz6OTqmeKyMjg/vvv79DRYA8sQzxJ598wqhRo/j3v//9vX0OhwO73d7lPtrLH68MKNUrbfjqK/Iy7WQ1xB53PEC/uXMZlGM4GFSM0 jiRar3mjlzJrfddhuTJ09m8ODBh79ZOxwOfvGLXzBy5EhGjx7N3/72tzbbmjdvHlu2bGHHjh3f2/fSSy8xatQoRo4cyW233XZ4e1ZWFiUlJdTW1nLWWWcxZswYRo4ceTihWLNmDTNmzGDChAmcdtpp5Ofnf6fd9evX86tf/Yp33nnncBXDiIgIbr31VsaMGcPXX3/Nf//7XyZPnszYsWO5/vrrcTgcAHz88cdMmzaN8ePHc GFF1JTU9Pp1/EQv7syoFRvtWXphxwYBBeHjDruMfbwcJIK7TSMNeyr2kf/6GNfQVDKkwoeeIDGbds92mbwsKGk3HFHl9poaWnh22 /5YMPPuDee 9l6dKlPPnkk zbt4/169cTEBBAWVlZm 3YbDZ 9atf8cADDxwuZQyQl5fHbbfdxpo1a4iNjWXu3Lm8/fbbLFiw4PAxH330EWlpabz//vsAVFZW0tzczI9//GPeeecdEhMTeeWVV/jNb37DU0/9b1rw2LFjv7eccW1tLVOmTOHPf/4z27Zt46GHHmLFihUEBgZy00038cILL3DmmWfy 9//nqVLlxIeHs5DDz3EI488wl133dWl11KTAaX8xO4DqzFDhKkj57R6XIozFchjVc63mgyoHktE2tx 3nmuWhyHliAGWLp0KTfccAMBAa6Pt0PLHLflsssu4/777//OYkarVq1i5syZJCYmAnD55ZfzxRdffCcZGDVqFLfeeiu33XYb8 bNY/r06WzevJnNmzczZ47r/2WHw0FqamqbMdjtds4//3zAdQthzZo1h5dsrq vJykpiW   YatW7dy4oknAtDU1MS0adPa9Rxbo8mAUn6ilGJAGN5vcqvHZfYdQ3h9Liu2LOXiEZf4JjjVq3X1G3xnxMfHU15e/p1tZWVl9OvX7/Dvh5b 9cSyvwEBAdx6663HXAipNYMHD2bt2rV88MEH3HnnnZxyyimce 65jBgxgq //rpDbYWEhBweJ2CMYeHChfzhD3/4zjHvvvsuc bM4aWXXupQ223RMQNK QGHw0FNdDNhDZASntLqsRknTadfoWF3 ffvbyrVU0RERJCamsqnn34KuBKBjz76iJNOOqnV8 bMmcM///nPw8lBe24THHLVVVexdOlSDhWlmzx5Mp9//jklJSU4HA5eeuklZsyY8Z1z8vLyCAsL44orruCXv/wla9euZciQIRQXFx9OBpqbm9mypWOT3E455RRef/11ioqKDj P/fv3M3XqVFasWMHu3bsB162FnTt3dqjtY9FkQCk/sHfvXqpS7KTVhBz38ughg2bPJrPQkB9YSYtTlzRWPddzzz3H7373O8aOHcvs2bO5  67GTBgQKvnXHPNNfTp04fRo0czZswYXnzxRQDuuusuFi1qvQRNUFAQP/nJTw5/AKempvLggw8ya9YsxowZw4QJE5g//7s1QDZt2nR4kN 9997LnXfeSVBQEK //jq33XYbY8aMYezYsXz1VcdmAA0fPpzf//73zJ07l9GjRzNnzhzy8/NJTEzkmWee4dJLL2X06NFMmzaN7du7Pp7D75Yw9hVdwlj5k7dfe40Hyu/lxNos/vJ/77V5/H0XDOe1s 28Pf9tBsS0/uaoVGfoEsbdW09YwlipXmfL s pDxZGJR9/JsGRIhuiANhWus2bYSmleglNBpTyAwcKXfcTR/U/duXBoyXEDiCgxbBx3ypvhqWU6iU0GVDKD5SJa8DS0CHtmyKUMHI0mSWwKWetN8NSSvUSmgwoZTGHw0F9ZDMJlRAZm9yuc9KmTCG9xJDbVODl6JRSvYEmA0pZ7MCBAzQk2EmuCWr3OQMmTiS xEl5UAN1zXVtn6CUUq3QZEApi 3cuZOKBBtpxLb7nIiICIJKXDOB9lft91ZoSqleQpMBpSy2edsqGoOF/tH92j74CEH1oQDsrdzbxpFKdU8FBQVccsklDBgwgAkTJnDmmWd6pMDOkXQJYxdNBpSy2P4DGwAY1ndsh86LDElBnIY9hTq9UPU8xhjOPfdcZs6cyZ49e1izZg1/ MMfKCws9Hhfh5Yw7ogPPviAmJiYLvV7aAnjdevWMX369O/sO7RCoa9oMqCUxcpqDwIwsN EDp0XMWAwyRWwPWe954NSymLLli0jMDCQG2644fC2MWPGMH36dIwx/PKXv2TkyJGMGjXq8LLBn332GTNnzuSCCy5g6NChXH755bSnsJ4uYawLFSlludrAaoKbhPSBYzt0XuKYMaRlLyU7SscMKO966NuH2F7m2SWMh8YN5bbJtx13/ bNm5kw4dgJ8ptvvsn69evZsGEDJSUlTJo0iZNPPhmAdevWsWXLFtLS0jjxxBNZsWJFm sZ6BLGmgwoZSmHw0F9tIOkiiACgkM6dG7WpEkkrzZs6l BMabNNQ2U6imWL1/OpZdeit1uJzk5mRkzZrBq1SqioqKYPHkyGRkZgOsDd9  fW0mA6BLGGsyoJSFDh48SG2CnYHVHUsEAPoPGkRIiYNmu53i mKSwpK8EKFStPoN3ltGjBjB66 /3uHzDi1rDB1b2liXMFZKWWbbzq1Uxgrp9oQOnxsSEoK90vVzTnWOhyNTylqzZ8 msbGRJ5988vC2jRs38uWXXzJ9 nReeeUVHA4HxcXFfPHFF0yePLnLfeoSxkopS2zcsgIjwoD4QZ06P6DBdUXhYPVBT4allOVEhLfeeoulS5cyYMAARowYwe23305KSgrnnnvu4SWKZ8 ezcMPP0xKSkqr7ekSxq3TJYyVstCPf7GAz0bt4d9pv2LKnB90 Py/zD2Vpy8r4NohV/Pjabd6IULVW kSxt2bLmGsVDdSUe abtR/4KROnR S0Yf4KjhQ5NlCLEqp3kWTAaUsVB9QS1iDIT5zYKfOjx46lORyw75yrUKolOo8TQaUslBdpIP4Khu2gM5N7EkePZqkCshrKfVsYEqpXkWTAaUsUlFRQV2sjfj64LYPPo5 gwcTVuagKrBJVy9UHtdbx5R1d535d9NkQCmL7N2bTWWMkCgxnW4jMzMTW3EzALk1uR6KTCnX1NXS0lJNCLoZYwylpaWEhHSsdokWHVLKIlu3rKQlQMiIzOx0G4GBgdirXUVKcqpzGBTbuSmKSh0tIyODnJycw3PuVfcREhJyuApje2kyoJRF9mWvg74wsM/ILrUTZCKAZvKrtPCQ8pzAwED69evYstqq 9LbBEpZpKjStcDQiKFTu9ROWFQygS2G3OJsT4SllOqFNBlQyiKVLaXYnIbM/mO61E5Yn77EV0Fu2T7PBKaU6nU0GVDKIjUhDcRWCUEh4V1qJ3rgIBKqDDl6m0Ap1UmaDChlkdooQ1yNvcvtJA4fRkIVFDnKPRCVUqo30mRAKQtUV1dTGSvENod1ua2Mfv0IrXBSEdBIs7PZA9EppXobTQaUssD2jaupDheSAhO73FZmZia2shaMQFFdkQeiU0r1NpoMKGWBbZuWA5CZMKDLbUVERGAqnQDk1 R3uT2lVO jyYBSFtiftw2AYYM6t1rh0exNrmpj TV5HmlPKdW7aDKglAVK3B/ao8bN8Eh7wfYYAHKL93ikPaVU76LJgFIWqJYqgpoM8TGpHmkvJDGNyDpDXuk j7SnlOpdNBlQygI1IU1EVwsi4pH2wrL6klAJeZUHPdKeUqp38dtkQERuFREjIgnu30VEHhWR3SKyUUTGH3HsQhHZ5X4stC5qpdqnJtIQVee5pUHiBrkKD U16KIySqmO88tkQEQygbnAgSM2nwEMcj uA55wHxsH3A1MASYDd4tIrE8DVqoDmmprqYgWoh1dqzx4pLRBg4ipNBRJtcfaVEr1Hn6ZDAB/AX4FHLmQ9nzgOePyDRAjIqnAacASY0yZMaYcWAKc7vOIlWqn7PWrqA4T4oMSPNZmZp8 BFY6aLQ7qGmq8Vi7Sqnewe SARGZD QaYzYctSsdOPKGaI572/G2K WXtm39BoDU2L4eazM9PR1T3gJo4SGlVMd57qZlB4jIUiDlGLt A9yB6xaBN/q9DtctBvr06eONLpRq08G8HZAFA7K6tlrhkUJDQ5Fa18 FdYX0j nvsbaVUj2fJcmAMebUY20XkVFAP2CDe5R1BrBWRCYDuUDmEYdnuLflAjOP2v7Zcfp9EngSYOLEieZYxyjlbUXVuQCMHDnNo 0GtrjWOSisLfRou0qpns vbhMYYzYZY5KMMVnGmCxcl/zHG2MKgEXAle5ZBVOBSmNMPrAYmCsise6Bg3Pd25TySxWmApvDMCBtsEfbDQ50jZstKNnr0XaVUj2fJVcGOukD4ExgN1AHXA1gjCkTkd8Bq9zH3WeMKbMmRKXaVhXUQFQV2G1dX774SGGJqYTXl1FQrrUGlFId49fJgPvqwKGfDXDzcY57CnjKR2Ep1SVV4U4iaz1/US4sM4P46i3kV Z6vG2lVM/mV7cJlOrpnE1NlEdDRGOIx9uO7t fuGpDgc4mUEp1kCYDSvlQ9f69lEdAtER7vO3EgQOJqYZSU XxtpVSPZsmA0r50NaNKzA2ITEizeNtp2dmElzpoCqgiWZns8fbV0r1XJoMKOVDe/ZtBCA9xbMzCQDS0tKQihaMQGl9qcfbV0r1XJoMKOVDuWX7ABgyZJLH246OjsZZ6QSgoLbA4 0rpXouTQaU8qGSphIARg8c6/G2RQR7YxDgqkKolFLtpcmAUj5Ubq8lvNpJUlySV9oPJAKAwrIDbRyplFL/o8mAUj5UFdpCeJV4rf3giHgCWwwFpfu91odSqufRZEApHzHNzZRHGcLrvFfrKyQ1jdhqKNTCQ0qpDtBkQCkfaSwooDQKIhzhXusjsm9f4mqgQBcrUkp1gCYDSvlI/oFtNAcIMYHxXusjbsBA4qoNhc26PIdSqv00GVDKR7bvWgNAcmwfr/WR2rcPEZVOymx1uJbzUEqptmkyoJSP7C/YBUCfPiO81kd6ejoBlQ6a7YaqJi1LrJRqH00GlPKRwup8AIYNHOe1PlJTU6HC4epPaw0opdpJkwGlfKSspYKAZsPQvkO91kdYWBjUuH4u0tULlVLtpMmAUj5SGVBPRJUhJibGq/0ENAcDUFSryYBSqn00GVDKR6rCWgit8X4/wfYoAIrKD3q/M6VUj6DJgFI 4GxqoiICwhoCvd5XUFwi4fWGovIcr/ellOoZNBlQygeaCvIpi4Rwp/cKDh0SkpZGbA0U1 jKhUqp9tFkQCkfyN /DYddiA72XsGhQ6L69CG2xlCgYwaUUu2kyYBSPrBz9zoAkrxYcOiQuIEDiamFkuYKr/ellOoZNBlQygf257sKDmVmeq/g0CGpWVmEVzmpsNdrFUKlVLtoMqCUDxwuODRgjNf7SktLw17loMWmVQiVUu2jyYBSPlDmqMDe4t2CQ4ckJydDlasKYXFdsdf7U0p1f5oMKOUDlQH1hFcZYmNivd5XUFDQ4SqExfWaDCil2qbJgFI UBXaQliNICI 6S gxVWFUK8MKKXaQ5MBpbzM2dhIRSSENgT4rM8giQSgqCLXZ30qpbovTQaU8rLmggLKIiHMEeazPoOiEghpNBSVH/BZn0qp7kuTAaW8rPDADpoDhOgg7xccOiQ4NYXYWiiq0iqESqm2aTKglJft2rMegISYTJ/1GZ6RQUwNFNYW qxPpVT3pcmAUl524HDBIe9PKzwkpl9/YmsMJc3lPutTKdV9aTKglJcVVLkG8Q3uO9JnfSZn9SWi2lAudT7rUynVfWkyoJSXlbVUYHMYhmcN91mfKampBFY5aLQ7qW2u9Vm/SqnuSZMBpbyswl5HWLUhOTHZZ32mpqZCpVYhVEq1jyYDSnlZVWgLodX4rOAQQGRkJM5qdzKgVQiVUm3QZEApL3LW1/u84BC4Eg97YxCgVwaUUm3TZEApL2ouKKA0EsIcoT7vO5BwAIqq8nzet1Kqe9FkQCkvKsvNpjFIiAyM83nfgWGxBLYYisoO rxvpVT3osmAUl60O3sDAAlR6T7vOzg5mZgaKNYrA0qpNmgyoJQX7cvbAUB6 hCf9x2Wnk5cDRTWFvm8b6VU96LJgFJelF/pKjg0KNN3NQYOierbl5gaQ3Fjmc/7Vkp1L36ZDIjIj0Vku4hsEZGHj9h u4jsFpEdInLaEdtPd2/bLSK/tiZqpb6vtLkMcRpG9B3h874T /UjusZQZmp83rdSqnvx7XyndhCRWcB8YIwxplFEktzbhwOXACOANGCpiAx2n/Y4MAfIAVaJyCJjzFbfR6/Ud1UG1BNaY8hIy/B53ylpaQRWOakLaKGhpYGQgBCfx6CU6h788crAjcCDxphGAGPMoRue84GXjTGNxpi9wG5gsvux2xiTbYxpAl52H6uU5SpDmgmtBrvd7vO U1NTkcoWQAsPKaVa54/JwGBguoisFJHPRWSSe3s6cOQcqRz3tuNt/x4RuU5EVovI6uJifXNU3uWsq6MiAkLqfJ8IACQmJuJ0lyQurS 1JAalVPdgyW0CEVkKpBxj129wxRQHTAUmAa KSH9P9GuMeRJ4EmDixInGE20qdTzNhYWURkJGiTWX5 12O1LvSkT0yoBSqjWWJAPGmFOPt09EbgTeNMYY4FsRcQIJQC6QecShGe5ttLJdKctU5u2jPkSIDIi1LIYAp6vyYXF1gWUxKKX8nz/eJngbmAXgHiAYBJQAi4BLRCRYRPoBg4BvgVXAIBHpJyJBuAYZLrIicKWOtHf/FgDiItMsiyEoKBqb01BYnmNZDEop/ d3swmAp4CnRGQz0AQsdF8l2CIirwJbgRbgZmOMA0BEbgEWA3bgKWPMFmtCV p/9uZsg2RISx1kWQwBCYlE11ZQXKlVCJVvOFtaePHp21hfupn5065k ozLrQ5JtYPfJQPuGQFXHGff/cD9x9j AfCBl0NTqkPyKnIgGQZmDLMshtC0VGJrdlFcU2hZDKp3ueveM3lnYD62ZMPH2X/g7upCzp/3c6vDUm3wx9sESvUIpc2uEfxWFBw6JCIzk5haQ1F9iWUxqN7j09f/wbv98 i3Ha4suoSMPCcP5z5NcYVemfJ3mgwo5SXltjpCa5z0zehrWQxx/fsTUwNlzirLYlC9x3Nr/0FgMzyw8Dluve1OJuWPpT7I8Mh//8/q0FQb2kwGRGSaiDwuIhtFpFhEDojIByJys4hE yJIpbqjypBmQqsMwcHBlsWQmplJSJWTSnsjDqfDsjhUz7flqyWsH9jCsAPRjBw6DoCf3/8Ew7Y6WBq0ldpGLYvtz1pNBkTkQ AaXIPzTgdSgeHAnUAI8I6InOPtIJXqbpyNjVSGG8sKDh2SkpKCvaoFI1DWoAsWKe95 Z0/47ALl836yeFt0dHR9K/IoiEY3lzydwujU21p68rAD4wxPzLGLDLG5BljWowxNcaYtcaYPxtjZgJf SBO5WNNVVVU5Gq5hs5qKSykNApCm627KgCuksSHqhBq4SHlLcYYtoXmEFdqOH32Rd/Zd/7CO0gsN7yz4y2LolPt0WoyYIxpc9RRe45R3YejsZGPLriAr2dOZsvZp/DiuPHk7NhhdVjdTnXefmpChXCbtXfSQkNDMbWun0t0EKHykp2rvmR3HxhSnYKIfGffhBNPou9WB7viqqlsqLAmQNWmdg0gFJHzRGSXiFSKSJWIVIuIjkjqYYwxvHXR2bw0bBu3/DSQG34ayOIzG/n0ovlUlukl5o44cMC1aGZcRKrFkYC9OQiA4jq9MqC8490P/onDLswcveB7 0SEPrZBOG3Cx1/ 1/fBqXZp72yCh4FzjDHRxpgoY0ykMSbKm4Ep3/viwft45sQ8NmTZuHH0jdw45kY2Dgrg7YsD O8Vl1odXrey56ArGUhJHmBxJBBojwSgqPxgG0cq1Tk7qrcR2Gw496wfHXP/7DN SES94aPN7/g4MtVe7U0GCo0x27waibJUU0kJrxW yoFE OOMR7hp3E3cNPYm/jjrz xKFw5mHGTzihVWh9lt5Lk/eAekW1dw6JCA6HjC6w3FFToGRHmes6WF3ORG0ooCCA0KPeYxU08/g/67HGwNKcRVUFb5m/YmA6tF5BURudR9y A8ETnPq5Epn3ru9zfx TgbJzaPYc7AOYe3n9r3VObGnMRHU22898CvLIyweyludN2fH95nuMWRQHBKCrE1UFytVQiV521c9iE5SUJ/ hz3mODgYGJLwqgJM wt2u7D6FR7tTcZiALqgLnA2e7HPG8FpXyrpaqKJTFbCKs1/OmqJ7 3//a5v0OMsH9QObn79vk wG6oXGoJqXPSv49HVt/ukoiMDGJrDEU6ZkB5wbKvXseIcMKIua0eNyDRVXtgyfKXfRGW6qB2JQPGmKuP8fiht4PriUpLS/n1VVfzwLRpvPe73/nFJbMX//oLtvazcaJjAuFB4d/bnxCawJyQSXw92sabD91lQYTdT2VIEyFVhoiICKtDITori haKGkutzoU1QPtqdyBOA1nzr2y1eNOOONSYmoM3 xf7qPIVEe0dzZBhoi8JSJF7scbIpLh7eB6mvr6em6ePZuTt32FY3glu7e9xHsXzLc0ITDGsLTha0LrDHdd/f Oe9wt8 7GYRd21K/yiwTGn5mmJlfBoVr/qPadnJVFRLWh3Fav/3bK44rCa0goE6JCWx9TPm76dDL3OdgVXKx/h36ove9WTwOLgDT34133NtUBf33oIebE1nP3NUG8erKd/5xm54mp2Sz/w92WxbRp8RtsGGQYVphITETMcY/rE92XAeURbB9jZ 3nn/ksvu6ouaiY0kgIabK24NAhqampBFQ5aLY5qW6utjoc1YPUlJaSlwKptZFtHhsUFERsUSCVEYaCmnwfRKc6or3JQKIx5ml3BcIWY8wzQKIX4 pxqquryX7mXzx/XhAJEUksuWAJf5z6R/Li4LHmN2gsKLAkrueXPYrTJlw595dtHnv 0AspjBPee 2v3g sG6vNO0BVuBAubb9B kJKSgpUtgBQUqeFh5TnfP3ey1RGCP2jB7fr MTgLABWb1jsxahUZ7Q3GSgVkStExO5 XAGUejOwnub1114jcXokpdHCH075MynhKZw 5HRmVk9ma5aNxx 9zucxtdRUszqtjIx9hlNOOLPN48 bfT2BzYZs2eWD6Lqv3BzXaOmYsCSLI3GJi4vDUeUEtCSx8qx12z8DYNr4M9p1/OgJp2F3GL7a LEXo1Kd0d5k4IfARUABkA9cAFzlpZh6pBVP/ZsvpgUwOXAQY5PGHt7 hxseJ Ggg3dSs6kr8O088Hdf QslMcLIgNHtOj48KJzBpVHsG2Jj58YNXo6u 8o VHAo0fqZBOCqAGdrDAC0JLHyrP2NBxCn4eTJZ7Xr ClnzSejyLCjNtvLkamOam8ykGGMOccYk2iMSTLGLIBWJpWq7ygtLcUekENluHD55Ou/sy80JJRxTZMpixL /eIdPo1r8Z73CWw2XH/5Pe0 Z3bWqZRFCe 99pj3AuvmDpbuA6B/hvU1Bg4JNGEAFFXpvVrlOcXhdSSWCRHB7Zs1k56RQVyek5yoOh1E6Gfamwz8rZ3b1DF8snQp9ZPCiWiyM73f7O/tv/X6 0k62MJ7rMXZ1OSTmJqqK9mUUUvmXhsDs9p3vw/gorN gt1h2FS1xovRdW9FDa5L8f5QcOiQgIhYApsNRWUHrA5F9RBOh4PiBENCdUiHzouuCqM BPaX6dUBf9JqMiAi00TkViBRRH5 xOMewNqF2ruRrxe9zfZBdmYEjyTQHvi9/enp6aTujyU/Hj5f8pRPYlr05l pChdGBLXvFsEhMREJ9C0MIC 1kcbGRi9F172VUU1QvWFwB5IsbwtMTCK2BkqqrBmoqnqerV99TkmMkBbYscW4UsNct89W6bgBv9LWlYEgIAIIACKPeFThGjeg2qEoZxX1wcKM4ccfpHf6zOsJqze8sMk3q3p9tPdDQhoMl13Y8RLDI wDyEu1seTdV7wQWfdXGdxEaJWTqCj/WcsrLC2d2FooqiuyOhTVQ6z85j0AhvUd36Hzho ZTUCLYfX2L7wRluqkVpMBY8znxph7genGmHuPeDxijNEh5e2Qm5uLvS IE04YefxBNuedeyGp6xpZm1hBRXGOV2NqqK5gU2oNGbsMI4d37MoAwFknXg7Al9  6unQuj3T3OxXBYcOic7qS0yNoahBJwEpz9iTvxmAk05a0KHzRp0yh7QSQ3btPs8HpTqtve9Y/xGRPSLysojcLCKjvBpVD7Jy5UoqBgfRvzqM6JDo4x4XFhZGYssgmgOEdz/07nCMD9//O3UhwgD7iE6dP3XafCJrDftsev/5aC3FxZRFQlhTkNWhfEd8VhZRNYYyaqwORfUQxaaYwGbD0L4d 0LRf8AAogoc5IfWeiky1RntXZtgBjAM16DBGOB9ESnzYlw9xrovP NAmo0JEUPbPHb ghuJK3PyQf6nXo3pw53vE9pgOP/cn3bqfLvNTv CEA70cVLfUO/h6Lq3urxcKsMhQo6f FkhNT2doCondfYWGh061kN1XXl0MwllNmzSsatgIkJkVQiVEYayOr1S5S/auzbBScCtwG As4D3gJu9GFePcSD7Gxx2YVzWtDaPPXXOHGI2NLI1qZ7CPO/chWmsr2VjUiWpOxxMnXJCp9sZGTqUmnDhvXef8VxwPcDBA1swIsSGJ1sdynekpqZiO1SFUGsNqC5yNjZSlADxdWGdOj8xMA2ALTt00SJ/0d6U7jNgAfAkMNMYc5Mx5iVvBdVTOJ1O6gJcb7zjR5/W5vEhISEkNA/EaRMWffKEV2L69ON/UxsqZLb0Q0Q63c7pM64A4KsN73kqtB5h9wHXfdSUJP8oOHRIUlISLVWuZKBYlzJWXbR99XLKI4XUgM4lvQP6TgBg1UbvXgVV7dfeZCABuA YBnwkIktF5HfeC6tn2LdvH2QEEl0npMZnteucuWdfRUqhk8VFX3olpvc2v0Vwo GM065v  BWjJl2GvEVhv3i3cGO3U1OyX4ABvTp3HgMbwkKCkLqXLOB9cqA6qpV6z8BYGDKyE6dP qk04iuNWwv3urJsFQXtHfMQAWQDezFVY54AHCy98LqGTZv3kx5ZgAD6qLb/S38jDPPJHxDPTvj68nL3eHReJqbGlgXV0r6jhbmzGlfLfHjERH6FAWTl qgpaXFQxF2f8XugkOjsjo S8Pb7A7XKorFtTq9UHXNnsItAEwad2qnzh8 fhxJBU5ybDpmwF 0d8xANvBnIA54AhjiHlSoWrF53UoKE22MiBjY7nOioqIIr  DEeH9z570aDzLv3iR6jBIqkgmKKjro90HBfWnNlxY8ulrHoiuZyinmqAG/yo4dEhQYDTiNBSVHbQ6FNXNFTYWEthsGDd2eqfODw8PJ7LYUBDdjMPp8HB0qjPae5tgoDHmTGPMA8aY5cYY39TM7eb2Za8CYETG2A6dN uMy123Cgo9e6vgnbWvENRkmH7S5R5pb8bEcwH4YtWbHmmvJ6gKbiKs2hAcHGx1KN8TEJ9IdB0UVfh2QSzV85QF1xFXDgH2gE63EdkQQXMAHCzb68HIVGe1OxkQkU9EZDOAiIwWkTu9GFePUNGYB8DQgW3PJDjSOfPnE7aujh0evFXQ0tzEqog8MnY0M2  Z4pHnjj3QmKrDHuatcY4gGlpoSLC/woOHRKckuIqSay3CVQXlcc4ianufCIAkBjuWutu21bvjI9SHdPed61/AbcDzQDGmI3AJd4Kqidobm6mMaKBoGZD1sCOletMSkoioNI1Sve9Zf/0SDzffPEiVeEQnh9JXFycR9q0BwaSmWvnYFKjrkDGEQWHmv3vqgBAZJ9MVxXCep1NoDqvrqKMkhiIc3StlsbAgVMBWL/9Kw9EpbqqvclAmDHm26O26aixVuzatYuW1ADSKgOxd JS2ozTLiWlwMHHxZ6Zh7tozcvue3zneqS9Q/q0pFATLqzbtsKj7XZH9XkHqYiASD8rOHRIXP/ xNRCqaPK6lBUN7b6q49w2IXU8IwutTPy5FOIqTbsKd/tochUV7Q3GSgRkQGAARCRC3DNKlDHsWvXLiqTA jTEtup88897zxC1taxI66enP1buhRLS10t34TlkrSjiQULLupSW0cbP3gWAIs/9c0CS/5s766NroJDkR1bxc1XkjMyCKlyUmFv0EFbqtM2bXN9kx UNa5L7QwZPpzYYgf59goPRKW6qr3JwM3AP4GhIpIL/Ay40VtB9QTbt62jMlIYHN254jOZmZm0lLgu53d1VsE3Hz9LeSSY7AAGDRrUpbaONmf 1UTUGbaVbfRou93Rnv2bAEhNbv/sEV9KTU0loMqBEShvLLc6HNVNHSh1fZOfMrntQmqtCQkJIaJcKIpq0duMfqC9dQayjTGnAonAUGPMScaYfV6NrJvbl7MOgGFpnZ9vPnPuxa5bBWVdu6f20eY3sbcYRgzq2v 8xxKTnExGruFgVLXH2 5u8spcBYcG9fWvgkOHpKamYrQkseqiYmcpYfWGAf27/nce1RhBQzAUlOmiZ1ZrNRkQkSuPfADnA ce8bs6jvI61/StQVkdGzx4pHPPPZegNbXsjGvgYPb6TrXRXFLCiqgCYnc2ccl5l3Y6ltYkV0VREgtFFXleab 7KGl0FVAZ1dc/F/WMiIjAUe36BqYliVVnlYc1EFsONlvXZ80ku2cUbNmmaxRYra1/zUnHefwOeMq7oXVv9YE1BLYYMgeM7XQbgwcPpjYvEoD3P/t3p9r49v2nKIkW6nYYJk2a1OlYWjMwwXX148Mlz3il/e6iQmoIaDQMzPTP2wQigjQHAnplQHVeWYwhptYzM2YG9ZsIwHpNBizXajJgjPnxoQfwE2AlMBP4Buj8V94errGxkcZoQ2KlEBAe0aW2ZpxyPil5DpZUfNOp8z/a S42h2HakLO6tDBRa2adcjkBLYZVu7/wSvvdRVVIM2HVhoCArs2/9qYgcf09FpVrFULVccUHsymLFhLo3MDoo42cMpuwBsOeMp1RYLU2r/OISICIXANsA04FLjDGXOyuNaCOYe/evTQm2Emq63r2fN5552HW1rAzvpG9az/v0LkNe/bwdWI5ETsbufwCz1QdPJZR004iI9 QbS/wWh/ zlFTQ2UEhLgXA/JXAdHxhNcbisp1gSnVcd9 /SEA6TH9PNLekFGjSCx2kidlHmlPdV5bYwZuBrYCE4DTjTFXGWM8u3rO9/scKyLfiMh6EVktIpPd20VEHhWR3SKyUUTGH3HOQhHZ5X4s9GZ87bFj5w4q42ykSdeL 4wePZrynHAA3vusYwWI1nzwNIWxQu1OmDhxYpdjOR6bzUZiYSB5CS3UN9d7rR9/1pKfT2kkhLWEWh1Kq4KSk4mpheKaQqtDUd3Qtuw1AAwfPMUj7YWGhhJZBsURWuHeam1dGfgbEAWcBCxyfwhvFJFNIuKtKwMPA/caY8YCd7l/BzgDGOR XIdrwSREJA64G5gCTAbuFhHPXMPqpC0719IcKPTzQPYsIly04CriclpY1rQZZ11du84zTieL936MOAynjDjHa7cIDkkP6IPDLqxc/b5X /FXtTkHKY EyIAYq0NpVXhGBjE1RgcQqk7JrdwHwNQTzvRYmxG1IVSHQ0WdXh2wUlvJQD9cH7LzgLOPeBz63RsMrgQEIBo4NER9PvCccfkGiBGRVOA0YIkxpswYUw4sAU73UmztknPQNd98ULpnpphdccUVVK2qYleqYcu7z7frnLo1a/gmo47gXQ3cuND7JSEmjHG95MtWveX1vvxR9u71OG1CYlTXqrJ5W0xWFrG1UNysdQZUx5VIBdFVhuSENI 1GR/gKr2 Y/dKj7WpOq6tZOCAMWb/8R7gunzv4Zh BvxRRA4Cf8K1JgJAOnDkqKcc97bjbf8eEbnOfethdXGx974ZVdS78pd mZ5Z0z49PZ0EhmFzGF7a8AzG6WzznK8  A/58UJwcRwDB3p/dPtJ8y4iucTJttpdXu/LH2XnbgUgM22IxZG0LqlvX8KrDeVSp4VeVIeVRzQRW HZt/y qSMBWL9ZFyyyUlvJwDIR bGI9Dlyo4gEichsEXkW6PA9ehFZKiKbj/GYj6uy4f8ZYzKB/wP 09H2j8cY86QxZqIxZmJiYqKnmv2eukPTCgeO9Vib1152HcFr61iaVU3Bx  1emxzXh7v1KwgoNHJlSdd47EYWpOQkEBiDhyIrcdp2k5Wepq8KlddiWEZ/llw6JCU1FQCqxw02ZzUNNdYHY7qRozTSWksxNR7dlzMyAmzCGgx7MjvWtl11TVtJQOnAw7gJRHJE5GtIpIN7AIuBf5qjHmmo50aY041xow8xuMdXMnFm 5DX8M1DgAgF8g8opkM97bjbbdMQ0QLiRVCYHSMx9o899xzKVljqA8WXl76l NeHTBOJ2vu/TnLhxla1jTyg4t/4LEY2pJQH0ttKOzat85nffqL4hbX/c5xA7tWr93bUlNTQasQqk7I3rqGmjAhMSDBo 0OmzSZpDJDbnPvnY3kD9qqM9BgjPm7MeZEoC9wCjDeGNPXGHOtMcYb7/p5wAz3z7NxJR4Ai4Ar3bMKpgKVxph8YDEwV0Ri3QMH57q3WaKyspLGGCG zrNzzQMDA7np/JsJ2dHAawOKyXvjpWMeV/z8c/w1bRPOFsO8PhcSGuq70e2D01xFjT5d8YrP vQXlYENBNc6SU3wz0WKDklISMBZ5VqkSJMB1RGrVy8FoE/iYI 2m5CQQFSJoTi0d85E8hftridpjGk2xuQbYyq8GA/AtcCfRWQD8ACumQMAHwDZwG7gX8BN7rjKcFVEXOV 3OfeZon9 /dTG2snviXM421fe 217Hu3guow4Z8rHqGl5Ltv5s2FRfxl3V/ZkmWj8OUSfn7dzz0eQ2tOnHsxUbWGNblHr3bdsxmHg4rwFsKrxOuzNrrKbrdjGly1EHRGgeqI3Qc2ADBi IkebzuiKoDSKCdNDp1iaJWuF5f2MGPMcmPMBGPMGGPMFGPMGvd2Y4y52RgzwBgzyhiz ohznjLGDHQ/nrYueti5dwf1oUJSkGcvpQFER0dz28LbcHxVydtjmlj6uxsxLS2H9z/9n5/w3jgHjZ Uc9OMm/DmuIhjGT1xIikHHGSH9K4pQi0lpZREC2GNQVaH0i4BTleiWlylq5Cr9iuozUWMYcrUuR5vO4ZYnDZhb85mj7et2sfvkoHubudeV/bcJ7avV9q/5ZZbaFoeRFClk/uGbuPT316Lo66OFz76I4 lbiZll5P6z4Vf/OIXXum/NQEBAcSUBFMcYyiusHTYhk815 VSEgWRJqrtg/1AQGgMgS1GSxKrDikLqCamEqLCoz3edlrMAAA2bO5YlVXlOe1OBkSkr4ic6v45VEQivRdW95WfvxOAARnDvNJ SEgILzz1Alv/uJcWp42fDV3N9Oem8mDhcwzZL3zz5x0898xzhIeHe6X/tqSGuP6n/urbty3p3woHd22iMUhICE xOpR2CUpKIqYGivTKgOqAiogWYqq88/1x6NATANi8d41X2ldta9e/rIhcC7wOHKqHmwG87aWYurXyWleNgcz04V7rY8qUKTzxwBOs/dVWYj5pYtDBQGa828KXf9jOw4/8jVmzZnmt77aMn3I2gc2G5VuXWBaDr 3YsxaAtBTPDqzyltC0NFdJ4toiq0NR3YSzudk1rbDJ82OhAEacOJP4SsOBmgNeaV 1rb1p3s3AiUAVgDFmF5DkraC6s1qqsTsMqX29lwwALFy4kI/f/RjnngTefHQn3x6I5Jl33uXGG71fbbA10049jYw8J9scved/6pzivQAMzfJMkSlvi zTh9gaQ3Fj7xrboTpv26ovqQ0VkoK8c/Wrb79 xJY4KQyo8kr7qm3tnf/WaIxpOjRSWkQCcJUNVkdpDGkivspOYLznBxAebfbs2cyePdvr/XREamoqMTmGjVObqWuuIyzQO98k/ElJk2tU/tj Y60NpJ0S vUj qBhq7Pa6lBUN7F2/acQDv1SvXP70263E1lu42CmA2OM38/K6Ynae2XgcxG5AwgVkTm4igG9672wuqfGxkYaowxxtfZe/ccc15CAww4bdn5hdSg UWGvxd5k6J/a3 pQ2iU1PZ2gKic1Ac06lUu1yz53ue0xY2a0cWTnRTWF0xgEhVV5bR sPK69ycCvgWJgE3A9rjn/d3orqO4qJyeH lg7cU0hVodiqcF9XMubfrm29bLJPYFpaaEyrIWwKtdSzt1BSkoK4i48VFpfanE0qjsobMxHnIYJXkwGkkJdS8ps2bbca32o42vvu1co8JQx5kJjzAXAU 5t6gjZ 7OpiRQS7TFWh2Kp8acvIL3IyZqS9VaH4nXNBYWURgth9YFWh9JuqampON0liYvrtfCQalt5UC0xVRAa5L23/f59XaW8129d4bU 1PG1Nxn4hO9  IcCSz0fTve288AWjAhpEcdcNLHXGDdpEkn7HewNr8ThdFgdjlc15 RQHAXhpvvMtA0JCcHUun7WZEC1R2Wkg5hqu1f7GDX1VEIbDXtKdnq1H3Vs7U0GQowxh5c4c//c80eGdVDuQdd9taxuMsXMW4KDg4ksCaI GHYVbbU6HK8q3rWVqnAhLizZ6lA6xNbiupVVXFNocSTK3zVWVlISK8S2eDfhHTx2LEklhnyja2ZYob3JQK2IjD/0i4hMAHRViaOUV7qq7vXJHGpxJNZLjnC9Bl ufMPiSLxr56EaA0kDLY6kYwIDIhFjKCrLsToU5ec2ffUJtaFCSliaV/sJDQ0lqtRQHN7o1X7UsbU3Gfgp8JqIfCkiy4FXgFu8F1b3VO2et52a1ruvDACMO/0iksoNX 7t2eVFD5a6agwMSx9hcSQdExCXQFQtFFf2nrLRqnM2bvoSgAEZ3q jEVUXQmUk1DTVtH2w8qg2kwERsQPTgaHAjcANwLBDCwip/6mXWoKbDDFp/awOxXInnXIKSXua2R5WQouzpe0TuqmiZtclzXEDxlkcSccEp6S4qhDWaBVC1boDRTsAmDjhVK/3FW9zLa62c/86r/elvqvNZMAY4wAudS9hvNn9aPZBbN1OY0gzMTVgD9WJFrGxsYTku8YNbNrVc0cHlwXWYm82DM3oXreGIjMzia0xFDXoAELVupKWYsRpGDVwgtf76pvsKmq0Zt2nXu9LfVd7bxOsEJHHRGS6iIw/9PBqZN2M0 mkMRyi67074rY7SU9wfVv cs1bFkfiHc7aWkojnYRWGOz27vXvHtOvHzE1UNqi5V9V6ypCGoiuEgLt3p8 O3LcTGxOw/acDV7vS31Xe5OBscAI4D7gz 7Hn7wUU7dUXFxMY5SN6KZgq0PxG5NOO5f0QidfFa2yOhSvaMrNpShGCKsLsjqUDkvJzCS02kmlrQGncVodjvJTxhgqYpzE1vimjsaIk2aQVA65jbqipq 1a20CY4x1y B1Ezk5OVRHCbFVEVaH4jemz5jBSx82sXNKFQ0tDYQE9KzKjE0HD1IYA2n5nl/f3dtSU1OxVzlw2AIobygnPjTe6pCUHyrft4/iWGF4iW/ xqNjY4ktdVKcUOuT/tT/tHcJ42gReUREVrsffxaR7vcO6EV7DuykJUBICPH AkXdRVxcHMGFITQHwJqtPa9GVc7uzTQECwmR3a/IVEpKCsZdhbCkXud1q2PbsGIJdSFCRmRfn/UZXRVIaZSzRw889kftvU3wFFANXOR VAFPeyuo7mj//i0AZMT1sTgS/9InbRI2p HL9T1vXaudB1z3Nft7aSU3b4qNjcVR5bo9oMmAOp6tu74BYPjAyT7rM9YZgyNAOOCetqt8o73JwABjzN3GmGz3416geyzR5iOFRdkA9E0bYnEk/mX6govpk2f4pmy91aF43MHKAwCM6uP9 deeJiJIs s cHGdzihQx5ZT6fpAnjb5dJ/1mR7t mhZveZjn/Wp2p8M1IvISYd EZET0QqE31HpLuuanjbI4kj8y8kzZhC9u4m90XVUN1ZbHY5HFZtyACYNnmRxJJ0TYMIBKKrQKoTq2EqlgsAmQ5/4LJ/1OWLoCQBs2vWtz/pU7U8GbgAeF5F9IrIPeAzXUsbKrdZZCUBKulYfPFJISAgxtQk4bbBy04dWh MxxhjKQhoJqXKSmphqdTidEhgVT2iDobhckwF1bOWRzcRX2LCJ75bnHjPzNKJrDAer9/usT9VGMiAifQCMMRuMMWOA0cBoY8w4Y8xGXwTYXTTYG4ioN4QlpFgdit8ZNep0AlsMn6x92 pQPMZRWkpJtCG02ndvkp4WmJRIbC0U1xRYHYryQy319ZTGQ1y9b4uoJfbpQ1Kpoche6dN e7u23snePvSDiLxhjKkyxmiVkmNoDG0hukYQW/f9cPCWUy69nP4HnKxq3G51KB7jmlYohDd138U7wzIyiKmBol4 ZqC5uZmXXnqJq666igsuuID777 f/Hyd57776y8piYakIN9/wYmpsFEapYVufamtTy454mcdMHgcDQ0NNEYK0Q3tKtvQ6wwcOJDE3YbCqGayK7KtDscjavZnUxoFsUGJVofSaTFZWcTUGkqayq0OxTKbN29m/PjxXHbZZXz00Uds3ryZO  8k8GDB/PSSy9ZHZ6l1q7 GCPC4MwxPu87pjGCujChtK7U5333Vm0lA Y4P6sj5OXlUR9lI8ahaxIcz DA4QC8u/K/FkfiGdt3rsaIkJowwOpQOi2pb18iqw1l1FkdiiW /vprTjjhBIqLi3njjTfIy8tj /bt7Ny5k3HjxnHZZZfxr3/9y owLZOd55ouPXXyGT7vOzXMVbtj9cZlPu 7t2orGRgjIlUiUg2Mdv9cJSLVIqK3C9wO5hykOkKIC9A6TMcz/YKrGZBvWLz7I6tD8Yjsgp0ADM0caXEknZeWnk5AlYNGu4Pa5t5V8W3r1q2cdtppJCUlsWrVKs477zxs7lt8gwYNYsmSJZxxxhnceOONfPpp71w0p8hRhBjD6P6 X4ZmcF9Xn s2f HzvnurVpMBY4zdGBNljIk0xgS4fz70e5SvgvR3u7M3Y2xCUoQOHjyeCWedxaDtDg5GVJNf0/3vxxY0uAbdjbfgjdJT0tPTodIB9K5aA5WVlZx77rmEhoaybNkyMjMzMQ4HLcXFOGtdSVFwcDCvvPIKAwcOZOHChVRW9r7BbBVhDURXCsF236 3Mm7aXAKbDdnFO3zed2 lo908IDdnGwB9EnVYxfGICFn1rpKmr67t/rcKSgJqsDcbRvUfZXUonZaYmIipciUDvakK4S233MKePXt49dVXSWppIe/Xt7Nz6jR2TT ZHRMnsfeii6n 5BMiIyN57rnnyM/P59e//rXVYfuUayaBEFdvzXoiGWPHkVoGRUbHDPiKJgMeUFp EIC 6Vp9sDUTFyxk H7D29tfx5juOwTFNDVREtlCeLkhKLD7rVh4iM1mQ5pcg157SzLwwQcf8N///pff3HEHI7P3svec VQtXkzk3Dkk3/VbEm65GWdVFTk330L b3/LpPHjuemmm3jyySfZsmWL1eH7zK4vPyc3AdICrLnaabPbiS2DsvBGS/r3Bw6Hw6f9aTLgAdUNrjfS1DQtONSaiRddxLiNDkqC69hQ3H3XK2/Oy6MgTgiv7b6JwCGBuKsQVuVZHIn3VVdXc8MNNzBq HCudTgoeughwk8 mYFLPibt/vuJu wyEm  mf7vLiL 2mupeO118m77Nb/97W JjIzktttus/op My61R/jsAtD08daFkN0fSgV0Yb65t5Z7Hbq1KlceOGFPutPkwEPGDh6AOI0JKVrKeLW2O12Mu3DCW4yPPfVP60Op9PqD ynMAaiTIzVoXRZSFg8AS2GIvfVrZ7sD3/4AzkHD/L85MlUv/Em8TdcT8ZjfyMg4bsrjUpgIEm3/pzEn/2MqvffR954g9tuu43333 ftWvXWhS9b2XnbwbgpGnnWBZDcmgGRoTVaz xLAYr5eXlERkZ6bP NBnwgLKWcmLrbASGhVsdit b9vNfceIWw7KyFVQ0VFgdTqfs3rGKlgAhOdp3y7p6S0hKCjG1UFzZ/Qd1tubAgQP85S9/4V8zZhL09Tck/uynJP3sZ4i4SqlUNlbyn03/4eZPbua6j6/j0bWP0nzFOUSdfTbFj/6Na2bMICoqigcffNDiZ IbhRRjcxhGZ461LIYh7pUSV69abFkMVnE4HCSfINj6VvisT00GPOAkBnJxcT rw gW k ZwtBNdlrshtd3vGZ1OJ2y/aCrEveAjO47ePCQiMxMYmqhsIeXJL7zzjuZFx7OCQUFxFx8MfHX/29plQ/3fsjpb5zOX9f lbyaPCqbKnlq81PMe2seX185loCkJKoeeIBbbriB119/nZ07d1r4TLzPGENZdBNxFTYC7YGWxXHSnAsRY9hbuM2yGKxSUlJC1LRoWmy5PutTkwEPWHDzX7jpnnesDqPb6DNtPiP3OXluzb9pcbZYHU6H7a92LV08cWD3XK3wSDH9 hFTYyhq6LmjttevX89Xr7zCXckphE2dSspv7zx8ReCfG/7Jr774FQNjBvLGOW/w1vy3eGXeK3xw3gdMSJ7AvWv/wKf/dxJNu/dwbZ8 BAQE8Pe//93iZ RdVXv3kp8oJDZZO3s8ve9AEsqgwNY7BrceKS8vj5pIIYYIn/WpyYDyuVPuuJ3J3zoot9fxyYHudz wSCoJbDRMHDrR6lC6LHnQIGKqDWXUWB2K1/zh3nv5S0YmQVFRpD38EBLgmkHx1OaneGz9Y5wz4ByeOv0pBsf bwBwWkQaT5z6BKdlncZjFW zcd4Qmp7/L5ctWMCzzz5LXV3Prdq4fvEiimOErBjrB0QnltopjOt Xxi6at B3TQGCfEh8T7rU5MB5XMhYWEkxIwnpczw9 WPdKtphsbppCSimfAyQ1hY912k6JD0zExCK53UBrTQ0NJgdTget2nTJrJWfMXAwEDSH/wDgUlJAHx 8HP suYvnJF1BvedcB Btu9fDrfb7Dxw0gMMixvGo2MLKXJUcEtmHyoqKnj55Zd9/VR8Zsu25QBMHjfH4kggqTmOshihIH v1aH41MG9rgGcqTHpPutTkwFliTn3P8BpK1rIduSxPHe51eG0W0thIYWxEF7X/acVgqsKoa3CNZ 5qK7I4mg87z933MEPYmMJXTCfiBkzADhYfZDbv7ydYXHDuO/E 7Db7Mc9P8gexMMnP0wTLbxwWSqRX37JhOHDeeKJJ3z1FHzuYP1 AE4cMdPaQIABia41TT799BWLI/GtomJX8tM3daDP tRkQFkiJSuL0KoMEioNjy3/Y7e5OlC7P5uiGIjpAdMKAcLCwpAq12tfWFdocTSetXXTJk7ctJmmkBAy77gDAKdxcteKuzAYHpn5CCEBbVfYy4rO4ppR17A8rpiNsdX8dspUVq9ezcaNG739FHzOGENxTB3htZASbn159WlTXFMbN 392uJIfKui2jWgN1WTAdUbzHzgQeasaGFr416 LfjW6nDaZdu2lThtQkpMz5k9Ym9yXeXoacnAkl/8kpEhISTdcTv2KNdguNd2vMbqwtX8ctIvyYjMaHdbV428ivSIdP57dgRDdu4gJCCA559/3luhW6Zu/35yUoTkmojDgyytNOGEU4itNBxs9t2oen9Q0 xaCyOlpycDInKhiGwREaeITDxq3 0isltEdojIaUdsP929bbeI/PqI7f1EZKV7 ysi0jOu3/YCIyZOJKAkmdhqw98 f8jqcNpl2771AAzu130XKDpakLgKmxRU5Fgciefk7d7NxOxscmNjSb/oIgAKagt4ZM0jTE2dyrkDz 1Qe8H2YG4aexPZkfWsiinj1pOm88ILL/i8ZKy3rV/0Bnnx0D/KP0qrBwQEEF8EBdE9bzxLaxrs9QS0GOKSs3zWp1VXBjYD5wHfWZ9SRIYDlwAjgNOBv4uIXUTswOPAGcBw4FL3sQAPAX8xxgwEyoEf eYpKE Y88DDzP6qmQ2Nu1hVsMrqcNp00D2tcMqwaRZH4jlBMYmENRgKSvdbHYrHrPjlL4mz2 lz128Pf8N9ZM0jOIyDu6fd3alvvWf2O5OMiAzenBXM/JAg8vPz eST7jcbpjXrtnyCEWHmFOsqDx4trj6KojioKuo9Vwcag1uIqhNs9uOPZ/E0S5IBY8w2Y8yx1qacD7xsjGk0xuwFdgOT3Y/dxphsY0wT8DIwX1z/R88GXnef/yywwOtPQHnMuClTkLwkomsMf1v2gNXhtKnIXklQvWH0wNFWh IxwWlpxFdDQWXPeLOtPnCAgVu3sT4yksFnnAHA qL1fLj3QxaOWNih2wNHCrAFcO3oa9md0EK2OcDY Hiee 45T4ZuuVyH62/g5GGzLI7kfzJjh2JE PKzV60OxSeMMTSFGaLqffvx7G9jBtKBI4uk57i3HW97PFBhjGk5arvqRk6//0FOXtHMuqbdrClYbXU4x2WMoTiymehSQ2CgdZXZPC0yK4u4akNebc8oSbzy9jsIADJ 9UvANWjw4VUPkxSaxI9Gdu3C4bz 84gLiuXDSXZ NWYsb775JtXV1R6I2nqOujoKE1uIqrYRGxJrdTiHTZs6H4BV2z zNhAfKS0tpSFSiGr27R1vryUDIrJURDYf4zHfW322I6brRGS1iKwuLi62Kgx1lCknnkhLbiLRtYZHl/zO6nCOy1FeTl4CRNVas8a7t8QPHkxcNRS3VFgdSpc1l5URv2YNXwfYOemCCwBXueFNJZv46YSfEhbYtdoQQfYgLhh6IWsHQEpLKaahgbfeessToVtu7/sfsL2PkOVItTqU75hx0mlEVzrJdvScMS2tyc/PpzbSRrTx7Vo3XksGjDGnGmNGHuPRWt3eXCDziN8z3NuOt70UiBGRgKO2Hy mJ40xE40xExMTEzvztJSXLLj/QaYsb2KtM5u1uf45s6Bo92Yqw4V48V1VMF9I69eP8EonlfYGmp3NVofTJet/fz8hQPTVVyMitDhb Pv6vzM4djDz s/zSB8XDb4Im9j5eEQTl/TvzwsvvOCRdq327bLXqA4TJg 1vtjQkUJCQkjIFw7GNXabKchdcfDAXmrDhIRg316d8bfbBIuAS0QkWET6AYOAb4FVwCD3zIEgXIMMFxnXX8Yy4AL3 QsBXSSgGzrhpJNoyEskqtbw6JL7rA7nmNZuXAZARsJQiyPxrPT0dGyVLRiB0vruu0aBs74e  LFfNbQwDk33wzAu3ve5UD1AW4eezM28czbXXJ4MnP6nsqysXYu6pPG0qVLKSjo/gs9Zde6hnHNn3K xZF8X1JTPCUxQu6uDVaH4nUHsl3VB5Oj0nzar1VTC88VkRxgGvC iCwGMMZsAV4FtgIfATcbYxzuMQG3AIuBbcCr7mMBbgN LiK7cY0h I9vn43ylEt/9yDjvmxkjexn3cGVVofzPdtyXUVmRg490eJIPCspKQlT7poiV1DbfT/Ucl9 mVCHg/LpJxEaGkqzo5l/bvwnI JHMCvTswPiLhhyIbXBhpzYUhJsNl55pXtXyGsuKeFgSjNhtTb6 uHS3INTJwDw6Zfd 3Vuj4LCPQBkpgzwab9WzSZ4yxiTYYwJNsYkG2NOO2Lf/caYAcaYIcaYD4/Y/oExZrB73/1HbM82xkw2xgw0xlxojGn09fNRnnHS9OnU5icTWWd49ON7rA7new425BHYZDhx7HSrQ/Eom82G1LumMHXXwkPGGHL//R92NDZwzq9 BcBbu98ityaXm8fe7PECOpNSJpEWksyyUXD90KH897//9Wj7vrb9hefZlCX0l0y/KDZ0tJkzL8LuMKzL9f/px11VUekayJue4ruCQ B/twlUL3f1fQ8w/It6VgfksC57hdXhfEdhaC0xJU5SU/xrgJUnBDhdA sKq7rnjIL6deuJLi1lZVwcI0eNotnZzH82/YfRiaM5Kf0kj/dnExsLhp7Ppn42pqWEsnr1anbu3Onxfnxl7TfvUB0mnDHhIqtDOaaxYyeSlO9kX0DPH/hd01wOQEpvuDKg1PGcfPLJVOenEFln JsfjR0wzc0UxjiJrAzwy29OXRUUFkdgsyG/dJ/VoXTK9v/3V6odDkbceCMAi/ctJq82j2tHXeu1f68FAxYgwLd9axkdGtptBxI6qqrYHVmMOA3zx1g22atVAQEBxBQHcjDRQVN1pdXheFWD1CHGkJimyYDq5W64 /f0/7yOVSF5rPOTucVle3dQGi3EOmOsDsUrglJS3YWHDrZ9sJ9xVFcT O0qPm5o4PzLL8cYw9Obn2ZA9ABOzjjZa/2mRqQyNWkyn42xccPwobzwwgvdcrR7zssvs26gkFwbQ3RwtNXhHFdqcBaNQcKarxZZHYpXNQQ1E1EvBAX6dgqzJgPK78ycOZPqghQi6gyP 0lVwvXrPwUgLca32bqvRPbrR1w1FFR3vwGExYveJcAYGk88kdDQUFbkrWBn U6uGnmVx2YQHM95Qy kJAqi0xwcyM7m22/9c1rs8RhjWPfus xPFuYM9Z8SxMcyddICAD5d03MnjLmqDzqJqvP9R7MmA8rviAg/ e3v6LeinpVh WTnWL9U7IZs15v8iKGev//sD1yFhwyFTd1vauH 555lX1MTc29y3SJ4avNTJIclc1a/s7ze96w s4iwhfLlkBZOjYnpdrcK6tevZ3V6BeI0/PCEH1odTqvOnHMesWVOtrRkWx2K11RWVtIYaSOqyfcVTjUZUH5p9uzZ1OyLIqDF8NTiB60Oh y6/QQ2G06adKrVoXhFxqBBRFU5KbPV4TROq8Npt a8PKL2H BLm3DiSSexqXgTqwpW8YPhPyDQ7v031GB7MGcMmMe3Q21cMXQAr7zyCi0tLW2f6Cf2/PFhPhshpDWmkRCWYHU4rYqOjiYxR8hOasLR1DMnjeXl5VEXaSOarlXK7AxNBpRfEhEuuuFWBm1s5iOziYqaEkvjyQuuIr7QSZ MPpbG4S19 vbFXuHAYTOUN5RbHU675bz0EgBRZ5 NzWbj W3PExEYwQWDL2jjTM9ZMGgBTQFQ1qeO5tJSli5d6rO u6L2m5V81biByijhppNusTqcdklzplMbKqxb/YHVoXhFzr69VIdDXKDv14bQZED5rYsuuoiKrxtpDISX3vuDZXE4m5vJi3cQXWbvkTMJAFJSUqDCVXioO9UaKHrjTdbV13HejTdSUl/Ckv1LWDBwAeGBvqvrPiphFFmh6Xw UrgwKblb3CpwNjRw4Le/4fWpQkh1MPNGeKZUs7dNHHUmAB9/83obR3ZPB/ZsxmEXUqJ8P31ZkwHlt4KCgpg 74f02 fg1bJPaXFac/n14M411IYKMS1xlvTvC3a7HVudu/BQbfdIBppycogqK2N7bCyDBw/m9Z2v0 Js4ZKhl/g0DhFhwbAL2Z4pnDoshbfeeova2lqfxtARxhgK7vsdS PzKEiycd3o670 0NJTzjz7MuJLnWyq3W51KF5RkLcbgD4 rjEAmgwoP3f9DTdgW1ZNSXgLH3/1vCUxrNnguuybEd z1iQ4WpCJAKCgpnsUHtrzX9c38IwLLqDZ2cxrO17jxLQT6Rvl 3K68/rPw4aweUADqc3NLFrkv9Pfyp55lj1L3uTpk4WwsjCumX6N1SG1W2JiIgk5NvYkNdJY778JV2dVVLrW2UvzcfVB0GRA bmkpCSSMk8mrsrwygZrkoH1 10lUMeM9a/V3DwtNCIBu8OQX7LX6lDapeT999ne0MC8H/6QTw98SlF9EZcOvdSSWJLDk5maOInPR9m4IiPDb8sTl7/8Mjl/eogHzg2g0Q6PnvVot7v11T9wMPUhwrIlz1gdisfVNLnG6ySn6pUBpb5n4Q03kfptPWsjijlQ6PuSr/ta8ogtczJ1Us9aoOhooanpxFVDbvl q0NpU0tJCXHFxeyMiyUtLY2Xtr9EekS6V0oPt9e5wy6kNAr6D4tg6eLFFBf7V ncshdfJOe e7n/klBy0uGsgLOYMniK1WF12BmnXUdAi GjjW9ZHYrH1eO62pGc1N/nfWsyoPzeCSecQNWOIMRpePHjP/m8/7yIBqIKna5Bdj1YVL/ JFZCTjeoQrjnpZewAfFnnUV2RTZrCtdw0ZCLsNvslsU0q88sIiSEr4e1MDk4hFdffdWyWI7kbGig4P4H2PPw7/jtlSFs69PMgOwBPPyjh60OrVNOnnEqmfucbA7vHmNbOqIpqJngJggPjvB535oMKL8nIsy55FoG73TwXu1Kmp3NPuu7prSAolhDVE1ot7uc2lEJw4eRUGUoaPb/wkN5b7/DgaYm5l57LW/tfosACeCcAdZW0HPVHDiLlUNsXDKwj6WzCowx1G/aTNGf/8yeM85k1dLn dnVAexKaGLw7sG8efeblsXWVXa7ndSyGArjYNvGzz3atnFaV2PD6XTSFGaIrrPmfUaTAdUt/ODKK3GsqKYy1MmSb1/yWb9r13yIESE5qGfWFzhSxuDBRFU4KbfX zTh6ihnQwNxubnsiIwgNSOVd/e8y8kZJ5MQan3RnAWDz6MpEJzDbGxauZLsbN9WyzPGUPH222SfcSb7LryQfS8 zV/HlHPHD yUOVu4LOAyXr/vdWy27v3WP3PqFQC88P5futxW04ED5N12GzsmTWb7qNEc OEPadzr 3EzxcXFNEXZiGoK8nnfoMmA6iYSEhKITZpKfKVvBxJ s/UTAIYPm WzPq3Sp08f7GUtGIGiuiKrwzmu7EWLCAKiZ81mec5yShtKOXfQuVaHBbhrDgSn8cVIOCMykhdffNFnfTtra8m5 Rbyf307ZSHC7 bFcPW1Tj4f00x6VQYfXvIhv7nqNz3iCtf5F19DxgEHK4P3dGlxqIo33yL7rHlULv6Y5okTiLjsEhq2bGX/D66kOS/PgxG3LS8vj7oIsaT6IGgyoLqRy2 8kX6rGlgbUsCBct9k7ttq9hBT6WTatFN80p VoqOjMRWuy6R5Nb59I yIHa 8QoPTyck33sBbu98iITTB0oGDRxIRFoy4iO2ZwvyRmbzwX9 sZNhSXMz H1xJ1Wef8f9OS2Dh3H1sGlVDWlM6/zrpX3z0s4/ISs3yehy EhQUxICiBAriYeU3nVu4qPhvj5F/xx1sxMnpiTmcm7GUM1Je5oc/MGyIrSb/t3f5dBXK/P37qYoQ4i2oPgiaDKhuZNasWVRvcCJOwytfPO6TPvdH1hB9sIUhQ4b4pD8riQj2Btclyvxa/601ELp9OzsCA4jsE88XOV9w9oCzCbAFWB3WYYdqDuwc7iRk/37WrVvn1f6a8/PZd8UVlO7fxdXzW1gxvoIkZxL/PvnfLPnJEqYNmubV/q1y7pn/R0CL4ZmPH nwuaVPPU3J44/zbEA191ziJOqKFAb2GUhaYRpFNeX87lwHXxZ8Re2Kr7wQ bHl7dpCQ7A11QdBkwHVjdhsNoaecSGjdjtZVPSp1 9rl TupjgGwsqCCAz0/SpiVgi2xwCQV7rP0jiOp3DjRlIdThyjx/DenvdwGAcLBi6wOqzvSA5PZlryFD4fbePyhDj 85//eK2vppxc9v/gSopK8rh6Xj31g4O4OvNqlt28jCn9ut 0wY6YffoCBm2DVWlllJbltvu88tdfp jhh3kmto4PfxJPXP84/jjjj3xwxQd8fMfH3D3gbupzGnj0HDtbnvyTz64OFObvASAz2Zpl0jUZUN3K5VddReiKGiqCmlm28yOv9rVu/WIA4gMzvdqPPwlOTCW61pBT4p/LxK576mkAhl56Ke9lv8foxNH0j/b9nOy2LBhyPqWRkDEiinefe47ycs8v/tS0fz/7r7iCsqJ8rj jHskK5Y/T/8jPZ/ 8R4wLaIuIMDNlHk1Bwp fbN9CSxUffkTenb/lP/0a fC6aDITMnnprJc4Pev0w6/Z5RdezkWhF9GAk8eydlL7zUpvPo3/xVblujWXmpTlk/6OpsmA6laGDRtGdU0C8VWGV1d57xsXwKpdXwIwYmTPHy9wSFjfviRUQm7FAatDOab6r1aQ73QSeeIQdpTv4Mx Z1od0jHN6jOLCHs4X4wSzg4O5sknn/Ro 81FRey/6mrKi4u5enoNDArlwRkPcvrA0z3aj7 79qe/Y8AuJ0tjd1Hg/mZ9POXLlnHw5z/n36NbWHxJOONSxvHCWS QGfX9ZP/e/7uXwJWBrB9g441FD3kr/O oaSwDIFmvDCjVPtMXXs3Ytc1869hDbk37Lw921Nb6bBJKnJx44myv9eFvYgYNJrHSkNfgf7MJGmtr6VNRSVFaGosPLMYmNk7LOs3qsI4p2B7MWQPnsXJ4ABf0SeLff/sbzc2eua3lqKnl4PU3UFdYyNWDSrFNjuSn43/Kmf39MzHypsDAQOanXkpTINz ryuOe1zxF1 w76ab ec0B0vPCmF25myenPsk0cHRxzw ICCAP175FyIONPGvjF2Ubd/oradwWIPUAZCUmOX1vo5FkwHV7Vxy6aVULq8EDK tecYrfThamtkTW0/kvmZGjhzplT78UdqwocRWGkqo9ulI6vb49r//JcxmI2HuqXy490Mmp0z2i9oCx3PZ0MtosjlZPg5Orqv3SBEi09JC7v/9jPrt27kxsAD7gkROzzqdH438kQci7p6uuuk3jN0QzOq NTzy1 8vurTnxRfJvvl6/jzP8PmMIM4fdD5/nvlngu3BrbY7a YsojYmUhkh/PP9e7wU/f80BzVjc0BMSIzX zoWTQZUt5OQkIBj0HjG7jG8nb3IK0sbb9/4GTWhEFoe1msGDwL069 fgPIWmuxOyhrKrA7nO/a  x5OY4g4dxoHqw/67S2CQ/rH9Gd6 nQ nhrMValJPHjvvTQ1NXW6PWMMBffeS 2Xy/ltVQFNt/QjIyqDu6fd3SvGCByPiPDgbYvou9fJM9HfcM8dZ1C1ZTPFy1ew9Jxz OaF3/Orq21sHBnIT8f/lLun3d3u2Sc/  FvSN/QwBtROynK2eW159DU1ERzOETVi2XLSWsyoLql T/6EZnf1FNqq OLnC883v7KTR8CkBo/2uNt 7OMjAxMmQPwr mFxhiCdu4kPziYz6tXEWgL5JS /j WY GIhVQEtbBxhDCzqpqnnnqq022V/vOfVLz2Ov8oLaHw58Nwhjj504w/ERHk zr2/iYlLZ2/XvE6ffYKbwzJYf4nl/B/S67j/lOyefAiO8Hp6fx77r 5ZtQ1HUqcTjvtNGq gsYg4eVPul7t8HgKCgpoibIRbVH1QdBkQHVT8 bNY8/WemKrDa u7vwb7PF8W7iGqGrDpElzPd62P7Pb7djrXd a/CkZ2L5pE0OAluHDWLxvMdPTpxMVFGV1WG2anDKZYXHDeHd2BNckJfK3e  loqKiw 1ULlpE8V//H 9WVbH28qFUJVTxi4m/YHj8cM8H3U0NHDiMt 9axxWcQ1RDDAWJkWSmjOb3J/6e9y54j8mpkzvcps1mY8EF19N/Twtv1n2N03hn7YLcnBzqI23EWFR9EDQZUN1UcHAwSWfO46QNTr6u2kBBbYHH2jbGsD2sjNi9TUyZOtVj7XYXwbgGVfnTjIKvn3ueEJuNhnPGUlxfzBn9z7A6pHYREa4fcz25ofWsHBPAD202br/99g61UfvVV Tefgff1tfz5vhUasbXcEqfU7h06KVeirr7CggI4LaF9/POr5az9JZveO7iF5k/cD5B9s5/477sssuwf11NaVgLq3cs82C0/5O3cyeVEUJCcLxX2m8PTQZUt3X5NddgW1GFAd7c/prH2j2wey2lkSAHnAwYYM00HyuFxaUS0mjILWp9qpYvlS//EiewIb2G0IBQZmTMsDqkdpudOZuR8SN587QI5sREsf FF1i6dGm7zm3YupX9N93MnoYG/l90KDELE0gIS DeE 7t1eMEfCktLY0wRz8CWwwfrnvFK30U7thGdRikxWZ4pf320GRAdVuTJ09mlzOS0XudvLH1FRxOh0faXb72bQASggf2yjfcyH79SKqE/WX kQw0NDSQUFBASWQEnxWuYHr6dEIDQq0Oq91EhJ M/wlFtho npfKvWnp/PLyyzlwoPUrLw07drBn4VUU1NTwO4Hpf5xLfn0 D05/8LhT4pR3zLz0SobvcbKsao1XZtmU5GdjREhL6ufxtttLkwHVbYkI4xdeyYTVzRQ5K1mRt8Ij7X6Ru5zoasOEqWd5pL3uJn7YMJLLDQdq/WOxohWffcbooGAOTh9ISX0Jp/Tx/4GDR5uWNo0ZGTN4ZWQVdekR/DEyiotOPZXc3GPXyahbs4ZdF19CcXk5v7PbuO7FX/NpwafcMPoGJiRP8HH06rzzzydiWwOlIU3srfT8ImlV1a7xOcmJmgwo1SlXXHklB9ZUEl1jeG1z15eLbWlpZkN4CfE7G5kxY2bXA yGMkaPJqHcUCRVXhsw1REbXn DEJuNvVNiCbAFMD1jutUhdcpvpvwGsdn4782DSQkP4yED10yZyttvv33426azoYFt995H9uVXkFtVxRMZ6dz13j94fMfjnJh2IteOvtbiZ9E7xcXFEVodB8CXm9/3ePv1jkoAUlOsuy2pyYDq1tLT0ykbPoJZGw1fFHxFYW1hl9pbv 5DakOgeVczI0aM8FCU3Uv/gQMJKmmh2eakuK7Y6nCoXbkSB/CVbQ9TUqcQGRRpdUidkhqRys8n/Jxvqjfy5V8uITEtjUciIqj96U95cOAgnh8/gVVjx8FLL7G0tpbNVy3kd288yT1r7yE9Ip2HTn7Ir1Zn7G2GnHAmyeWGz3d87PG2m4MaAUiOsGbFQtBkQPUAZ1x3Hekr6zAYXtnRtQE n296FzGGqJCh2Gy983 PhIQEKHWNvzhYfdDSWIqLi8moqGDbgChy6vI4tc plsbTVRcPuZgzss7gHzkvs/9ftxJ/688Z0r8/pwcH07  nj0R4aw48wwuW7Oa835yBTd8cgOC8Njsx3ScgMVmX3EFA7MdbDT7PVrorKGhAWeEweaEuJA4j7XbUb3z3U71KPPnz2d1USOTdjh5eeuL1DXXdbqtz2rWk5nj5OQ553gwwu5FRLA3uEq1Wp0MfLp4MWNDQ9l0UjKCMDNzpqXxdJWIcM8J9zA2cSy3f/NbPp0eyYRPP2Xc1i2cuG0rl61ezTWPPMKm k1c/sHlOIyDf8/9N1nRWVaH3usNHjyY8BxoDDTsKNvhsXZzcnJoiQ4guiEAu83usXY7SpMB1e2FhoYScdppzPvGSbWjljd2vdGpdvbmbmFfdAOha s4/fTetfrb0YID4rE5DQeKvVeCtT22vP0OITYbm/o0Mi5pnF vRdBeYYFhPHHqE0xLm8b9K 9n4UcLeWf3O6wpXMOiPYu45uNr NlnPyMzMpOXz3qZgbEDrQ5ZuaWHDwXg682LPdbmgf37aYiyEeu0doaMJgOqR7j4uuvI21nNkFzh S3P0ezs Apx769wVTKsORhM//79PR1itxKS0YeESjhQZF0yYIyhae1a8mNgd0ses/v0nNUjwwLDePyUx/nt1N SX5vPnSvu5KqPruI3y3/Dvsp9/GLiL3jhrBdItfAesvq MXPPI6HSsHzLEo 1mb9zJxVRQmKgdbcIAHQ0iuoRpk6dypOxsSz4qoWH0gt4P/t9Fgxc0O7zjTF8XLKCvhVOBkzzz2VxfSly8CCSK9azv3K/ZTHs2LGDIc3NrBgfD9R3yymFrbGJjYuGXMQFgy9gd8VuSutLSQhNoH90f0svF6vjm3LuuWTd8wd2DsjDGOOROiRlO3dSNhSmRKd7IMLO0ysDqkcQEU74yY J29pIv7Ignlj/BE2O9q8Qtyl3DXvDa4lZWctZ8 Z5MdLuIWX0aJIqDLktJZbFsPSjjxgbGsrGUSEMjRtKRqR11dm8ySY2BscOZlraNAbFDtJEwI9FREQQUxRIVZiTnOocj7RZkbOP hAhNa6PR9rrLE0GVI9x RVX8HZDA5cvrievNq9DMwte/vIxgpsMeRuamDVrlhej7B76DxlCeKmTmoBmqpuqLYlh53vv0xBpY0dYRY 6RaC6tz4RrnEDn337tkfaq6pxFRxKtbDgEGgyoHqQsLAw4i 8gP7ZDsZUxvDkxiepaKho87zS lIW169lzBYHA ecRWBgoPeD9XMDBgyAYte4C099A qI5uZmZOsWvh0kGEyPu0Wguq8T5l5KWIPhi00feKS9JmcVAClJ1o5TsiQZEJELRWSLiDhFZOIR2 eIyBoR2eT 7 wj9k1wb98tIo K 2aNiMSJyBIR2eX b6wVz0n5h2t/ lNeq6zgstdKqG6q5o r/9jmOU vfIxmcdKyuIwLL7rIB1H6v5CQEGw1rsvVVkwv/OabbxhrD2Dl6BAyIzMZFDPI5zEodSwT5s2j30Ene6Tr5bqdTidOd8GhFIsHi1p1ZWAzcB7wxVHbS4CzjTGjgIXA80fsewK4Fhjkfhya /Vr4BNjzCDgE/fvqpfq378/lbNmkVzo5LycdBbtWcRnBz877vE51Tm8vPctTthm2FwXwuzZejn6kBBcefX 0t0 7/uTjz5icEwo29JaOLXPqb1ywSjlnwICAkgsDaE41tmuK4 tyc/Pxx7tSrqTwpI8EF3nWZIMGGO2GWO V7XBGLPOGHMo3doChIpIsIikAlHGmG Mq4j3c8AC93HzgWfdPz97xHbVS/3fPffw3/JyFryQzZCwftz 5e3HXFzE4XRw75e/RVocJC6qZME112C36 CtQ8JT xBbbdhTsNXnfWcv/pitg y0iNHxAsrv9It0jRv46LOurYeyd 9enDEBhDXbLV J05/HDJwPrDXGNALpwJE3LnPc2wCSjTH57p8LgGTfhaj80dChQ6mcNZOqJgc/e9NBoC2QaxZfw87ynYePcRonD616iG KV3PlJw7eOFjOj370I8ti9kdRw4aRXmrY7eMrA Xl5UTs38fKIUJicDyjE0f7tH l2jL71B9gdxiWr /aokX7t26lNkqIJ8JDkXWe15IBEVkqIpuP8ZjfjnNHAA8B13ekT/dVg MuNi0i14nIahFZXVxs/QIsynvuevhh/lRWRvy6bP5cOw D4ZL3LuH33/ye57c z9UfXc1L21/irNVQ82UVJ152GWlpaVaH7VfSJ04krdRwoKXQK2u4H8 yZcsYGxHK oF2ZvU9BZv483cW1RuNPvU0MgoNe03XBtcWb9lCWaSQYvEtAvBiMmCMOdUYM/IYj3daO09EMoC3gCuNMXvcm3OBIycZZ7i3ARS6byPg/m9RKzE9aYyZaIyZmJiY2NmnprqBrKwsxv/kx3xcXU34n5/l6cRbOaPfGby9 20eXvUwudU53Lw2kQuWwf8rLeGOO 6wOmS/M3DECMKLHNTbWyhtKPVZv58uXox9RDiNAYZT uosAuV/bDYbKSXB5Cc6aGxu7HQ71dnZlERBapT1NTT8KuUWkRjgfeDXxpgVh7a7bwNUichU9yyCK4FDScUiXIMNcf 31WRD9R6/uu02Xk5MYE9TI3U/uZ1fHBzJl cvY8mU5/nXm4mc/HEB/7d3Lz 69Vb69bN2jq8/6tu3L6bQNb3wWGMuvOXAJ5 wfoiNSAllUvIkn/WrVEcMiBxKc6Dw4QdPd7qNpsJcKiKE9ATry59bNbXwXBHJAaYB74vIoVUfbgEGAneJyHr349D1k5uAfwO7gT3Ah 7tDwJzRGQXcKr7d6UICQnh ddf5 bSUtY3NlJw733sGz F8gWXUr99B78pK6Nq4ADuvPNOq0P1SwEBAdhqXKsXZlfsaeNoz8jOzia9qoLVg4ST06cTaNeaD8o/zZ17NQAr1izqdBvNjkoA0hOs/zJiydoExpi3cN0KOHr774HfH ec1cDIY2wvBfRaojqmgQMH8vrixZx  umMRbhs3DhqbcJDy5cTkprKskWLCA21dhSvPwsMSyS4qYw9 VtgqPf7W7JkCSnDIqkJFU4deIb3O1SqkyacMIe4bw0HTG7bBx9DfX09tggnAGkR1o9X8qvbBEp5w TJk9m4cSOZZ8/j9o0beGDjRs5ZuJCVK1fSt29fq8PzayEDBpJWCtnFO9s 2AM /egjKkeHEmzsnJB2gk/6VKqz0spCyE1x0FRX1 Fzd 3aRWC868qXJgNK UhGRgbPPvssxcXFFBYW8o9//AMdRNq2hDFjSC8zZNcc8HpfDoeDoq  Ys1gG1PCRhAWGOb1PpXqikExw6mMED557T8dPnfHtm044wIQY33BIdBkQCnVikETJhBf7KTYXkN9S71X 1q9ejV90uyURQlzhrc5A1kpy51 qmvs sr173X43INr1lAWIyRKFIE268fGaDKglDquESNHYitwLQWdXZnt1b4 /vhjAsdHYHfCrEGnebUvpTxh6qhZhDQYcjqxTkHF1q2URAmpYSleiKzjNBlQSh1XTEwMUuJaF2BnmXfHDXy6 CMOjAhiTEsq0cHRXu1LKU wiY3U8lDyUg11HSxk13LgAMXRkB7rH OWNBlQSrUqMDCR4CbDjvyNXuujqqqKhqLt5McLs1Ome60fpTxtSMwIchLhmxfaP27A6XQSXF5GWaSQFqfJgFKqGwgdPITMEtjuxWTgs88 I2l8JABzpl7htX6U8rTTZ12OEWHNpo/bfU5ubi4x0XYcdv YSQCaDCil2pA0eTJ9igw76/Z7bY2Cjz/ mLoxYQwpDyUt0foCLEq117T JyJOKAw4biX879myZQshya6CXn0i 3grtA7RZEAp1aqhkyYRl99Cta2R4nrvLPD1 TcfU5Bm5 SgEV5pXylvCQsMI6k6hJx0Q8WOHe06Z8vKlTQlumr aTKglOoWhg0bhiPHtRjLkctAe8q ffsITqsBYM7Qsz3evlLeNjRmBLtThfX/fb5dxxd9u4r8WCFIAkkOT/ZydO2jyYBSqlURERE4KlzzoHcUb/N4 0uWLCF4UhT9CgxDpuiUQtX9zJ1yHo1Bws5tn7Xr JY9u8mPg8zwdL9Zots/olBK TVbYgYJlYZtOWs83vZ7y9 jqk8gM6vSsIWHe7x9pbxtSsYUAEpCyjAOR6vH1tbWEltZSX6cjT6x/jM RpMBpVSbYseNo3  YXPZVo 263A42Fa3DoC5mad6tG2lfCU5PJmohiD2ZgiFX3/d6rEbNmxgcEgwRbFC3yj/mFYImgwopdph8MyZ9M1zkmvKqWio8Fi7q1evJmJMKINyDQNOON1j7Srla0Ojh7MjQ9j ysutHvf1l18SFx9Ms92QGZnpo japsmAUqpN4ydORPa6BhFuKtnksXZf/vhlmjODODE7iJAROpNAdV zR51OeaSQv21lq8ft/fRTShPsAGRFZfkgsvbRZEAp1aa0tDTqCw3iNGwsWOexdj/N RQxhlNipyB2u8faVcrXJqROAKA8qobGiopjHmOMwbl1KwfcC6YOjB3oo japsmAUqpdnKl9ySiBDQda/ bTXrm5ubT0rWfoQciacopH2lTKKoNiBhFsAtmTaWPl448f85itW7cy0OFkb1oQ8SHxxIXE TjK49NkQCnVLjGTJjEoz7C5aqdHKhE   Gz2NKDOXGrk/ATTvBAhEpZx26zMy51AhszoeK99495zPvvvcfEsDDy 4T61VUB0GRAKdVOw6dPJ lgC9XSwN6qvV1u7/3972NvMcyo70Ngmn/UZ1eqK05MP5GCRBvRTWVU5OZ b/ Gd94hPiCAnIhGBsUMsiDC49NkQCnVLlOmTMGxow6AVfmrutRWTV0N5cnlTNgNqVNmeCI8pSw3NW0qANv72Vj2 99/Z19ubi6RO3dRFAsNNDMoVpMBpVQ3lJycTGljGPGVhq/3ftaltv7x8T wR9qZtclJxGwdL6B6hsGxg4kNjuXrfjZCPl1GU1PT4X3PPPMMcyMiyB6XCsDQuKFWhXlMmgwopdotcPw4Ru43rC5ag9M4O93Ooj2LCKlxMq4kgrDx4zwYoVLWsYmNKalT2DU0lL4ivPqb3wBQXV3N 489xrCQEHKm9iXEHsLg2MEWR/tdmgwopdpt Jw59N3bQiX17Crf1ak2cipzKIstY8Y2GzHTZyKBgR6OUinrTE2dSlVAA5uT7US88QavvPQSP/rRjzhfbJjgYHbE1DE8fjgBtgCrQ/0OTQaUUu120vTp1G6tBWB57vJOtfGHJX9AgHO bSFy9iwPRqeU9U5KPwmA/dfPYkhQMDv 7 fYlyzhzKgoon74A7ZX7mJ04miLo/w TQaUUu02aNAg9tcK/fINn z8oMPnVzRUsLxqOekbmkisCyD8pJO8EKVS1kkOT2Zc0ji icgj uqruDAmhruSUwibNIn9CybR7GxmUsokq8P8Hk0GlFLtJiIETJ3KpJ1ONtXspLiuuEPnv7D1BZx2J9dsDCb8hGnYIyO9FKlS1pnTdw47y3fSdMPFZL3 Opn/epI T/2Hr4pWEmQLYmLyRKtD/B5NBpRSHXLyOecQubEegE8PfNru8 qa63hu83PIhlpG5LcQfdZZ3gpRKUvN6TsHm9h4Z887hI4cQcT06RAQwOc5nzM eTxhgWFWh/g9mgwopTrktNNOY0N2FRnFhne2vd7u897c9SZ1po5Z34IEBxNxii5ZrHqmlPAUZmXO4vWdr9PQ0gDAhuIN7K/az5n9zrQ4umPTZEAp1SFJSUkUZvXllA1ONlVtZ0fZjjbPaXY089Smp6jfWceVVZFEzJyJPSLcB9EqZY0fDP8BFY0VPLPlGQCe2vwUEYERzM2aa21gx6HJgFKqwyadfTbRq oIdMDL21tfvx3g/b3vU9xQTMbSasKamog d4H3g1TKQhOSJ3Ba1mn8c8M/ufWzW1l2cBk/GvUjwgP9MwnWZEAp1WGXXHIJH RXMGuDk7d3v0VuzffrsB/iNE6e2vwUFMHN1QkEpqe77qEq1cPdPe1upqRNYcn JZwz4ByuGnGV1SEdl39VPVBKdQvDhg0jL6svN69oYtlYw6NrH Whkx865rHLDi5jb VeWhblM9KZSMzFFyN2u48jVsr3IoMi cep/8DhdGC3 fffvF4ZUEp1yvmXX84H 0s5 xsHH z9gC9yvvjeMcYY/rXxXwTUBPCTvGhsERHEXnyRBdEqZR1/TwRAkwGlVCctXLiQl2trmP VoW9jJPd fS8l9SUANDmaqGio4MXtL7KldAtNb QyMzCIuIULsUdHWxy5UupoeptAKdUpycnJnH7JJbyzeDE3vyT89oeNXLfkOlLDU/k672uanc0ABOYH8qsDUdgSYohbeKXFUSuljkWvDCilOu2Xv/wlfysqIqVE NWaNCoaytlRtoNLhl7CbZNuY759PoP/tJcxAYEk//KX2KOirA5ZKXUMemVAKdVpw4YN49LrruO F1/kgY9a OfGVBxlpYQMWUtjSh7R77/H5QkphE2frtMJlfJjemVAKdUl999/Pxvj4ni4ro7G5BSi5p1FfV4eNR99RInDiVx4AZl//Qti07cbpfyVXhlQSnVJTEwMixYt4rTTTuO5V19h6NCh7Nixg7jYWN5ZtIih06ZZHaJSqg2aDCilumzEiBGsXbuWRx99lI0bN3LOOefwk5/8hNTUVKtDU0q1gxhjrI7BEhMnTjSrV6 2OgyllFLKJ0RkjTHmmOsnW3ITT0QuFJEtIuIUke8FJiJ9RKRGRH5xxLbTRWSHiOwWkV8fsb2fiKx0b39FRIJ89TyUUkqpnsCqET2bgfOA75csc3kE PDQLyJiBx4HzgCGA5eKyHD37oeAvxhjBgLlwI 8FbRSSinVE1mSDBhjthljjrnuqYgsAPYCW47YPBnYbYzJNsY0AS8D80VEgNnAoUXVnwUWeCtupZRSqifyq7k IhIB3Abce9SudODgEb/nuLfFAxXGmJajth v/etEZLWIrC4uLvZc4EoppVQ35rVkQESWisjmYzzmt3LaPbgu dd4IyZjzJPGmInGmImJiYne6EIppZTqdrw2tdAYc2onTpsCXCAiDwMxgFNEGoA1QOYRx2UAuUApECMiAe6rA4e2K6WUUqqd/KrOgDFm qGfReQeoMYY85iIBACDRKQfrg/7S4DLjDFGRJYBF AaR7AQeMf3kSullFLdl1VTC88VkRxgGvC iCxu7Xj3t/5bgMXANuBVY8yhAYa3AT8Xkd24xhD8x3uRK6WUUj2PFh1SSimlegG/KzqklFJKKf hyYBSSinVy2kyoJRSSvVymgwopZRSvVyvHUAoIsXAfg82mQCUeLC93khfQ8/Q17Hr9DXsOn0Nu87Tr2FfY8wxK 712mTA00Rk9fFGaar20dfQM/R17Dp9DbtOX8Ou8 VrqLcJlFJKqV5OkwGllFKql9NkwHOetDqAHkBfQ8/Q17Hr9DXsOn0Nu85nr6GOGVBKKaV6Ob0yoJRSSvVymgx0kIicLiI7RGS3iPz6GPuDReQV9/6VIpJlQZh rR2v4VUiUiwi692Pa6yI05 JyFMiUiQim4 zX0TkUfdrvFFExvs6Rn/XjtdwpohUHvF3eJevY/R3IpIpIstEZKuIbBGRnx7jGP1bbEU7X0Ov/y361RLG/k5E7MDjwBwgB1glIouMMVuPOOxHQLkxZqCIXAI8BFzs 2j9UztfQ4BXjDG3 DzA7uMZ4DHguePsPwMY5H5MAZ5w/1f9zzO0/hoCfGmMmeebcLqlFuBWY8xaEYkE1ojIkqP f9a/xda15zUEL/8t6pWBjpkM7DbGZBtjmoCXgflHHTMfeNb98 vAKSIiPozR37XnNVRtMMZ8AZS1csh84Dnj8g0QIyKpvomue2jHa6jaYIzJN8asdf9cjWuJ fSjDtO/xVa08zX0Ok0GOiYdOHjE7zl8/x/t8DHGmBagEoj3SXTdQ3teQ4Dz3ZcUXxeRTN E1qO093VWrZsmIhtE5EMRGWF1MP7MfUt0HLDyqF36t9hOrbyG4OW/RU0GlD96F8gyxowGlvC/Ky1K dJaXOVbxwB/A962Nhz/JSIRwBvAz4wxVVbH0x218Rp6/W9Rk4GOyQWO/Jaa4d52zGNEJACIBkp9El330OZraIwpNcY0un/9NzDBR7H1JO35W1WtMMZUGWNq3D9/AASKSILFYfkdEQnE9SH2gjHmzWMcon LbWjrNfTF36ImAx2zChgkIv1EJAi4BFh01DGLgIXuny8APjVazOFIbb6GR91PPAfXPTTVMYuAK90juacClcaYfKuD6k5EJOXQeB8RmYzr/VIT yO4X5//ANuMMY8c5zD9W2xFe15DX/wt6myCDjDGtIjILcBiwA48ZYzZIiL3AauNMYtw/aM LyK7cQ1OusS6iP1PO1/Dn4jIObhG2ZYBV1kWsJ8SkZeAmUCCiOQAdwOBAMaYfwAfAGfy/9u7f1Ar6ziO4  POjgIDiJiUAiJDuZQ0mCXwmo0HcKhxaDNlnDKqMXFwWgyaBQClTYdDJUG/7Rm3bpqiC0NQTQ5iFI3/DY8v BI59pFz7le/L1f03Oe34/n Z2Hh3M PL9zfl/4BbgLvPdkRrp8LeIa7gPeT/I3cA94x2D/HzPAfmAuyWzb9zHwHHgvLtJiruHU70VXIJQkqXNOE0iS1DnDgCRJnTMMSJLUOcOAJEmdMwxIktQ5w4AkSZ0zDEh6qCTrRkqn/p7kt7Z9J8kXUzrnwSTvPqT9rbY2haQJcJ0BSYuW5DBwp6o m I5VjGsxf5SK/Y1rk9an5mqujutsUi98MmApEeSZFeSs237cJIvk3yb5Nckbyf5NMlckvNt7XWS7EhyOcnVJBcWKGX7BvD9v0EgyQdJbrQqll8BtNXXLgFTq 8u9cQwIGlSnmf4It8LnAAuVtV2huVTd7dA8Dmwr6p2AMeBI2OOMwNcHXn9EfBiq2J5YGT/d8CrE38XUoesTSBpUs5V1XySOYa6E fb/jlgE7AVeAH4ptVcWQmMK1izkQeLU/0EnExyhgdLt/4BPDO54Uv9MgxImpQ/AarqfpL5kUIq9xk awJcr6qd/3Oce8Dqkde7gdeAPcAnSba3KYTVra kx Q0gaSlchNYn2QnDDXck2wb0 9nYHPrswJ4tqouAoeAtcCa1m8LcG3qo5Y6YBiQtCSq6i GUqxHk/wIzAKvjOl6juFJAAxTCSfa1MMPwLGqut3aXge nuaYpV7410JJy06S08CHVXVrgfYNwKmqenNpRyY9nQwDkpadJFuBDVV1ZYH2l4H5qppd0oFJTynDgCRJnfM3A5Ikdc4wIElS5wwDkiR1zjAgSVLnDAOSJHXuHwbimq3tJ0XDAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "total_loss : 118.760689, mse : 118.739850, pen : 0.020940\n"
     ]
    }
   ],
   "source": [
    "ts_path = os.path.abspath(\"./Resources/TestingData_PeriodicTraj_Single.xlsx\")\n",
    "ts_dataset = TrajDataset(ts_path)\n",
    "\n",
    "test_config = {'save_path' : 'models/model_DDNN.pt'}\n",
    "\n",
    "model = DDNN(config['layers']).to(device)\n",
    "ckpt = torch.load(test_config['save_path'], map_location='cpu')  # Load trained model\n",
    "model.load_state_dict(ckpt)\n",
    "\n",
    "criterion = CustomLoss(mode='mse')\n",
    "\n",
    "\n",
    "from sklearn.metrics import r2_score\n",
    "\n",
    "\n",
    "ts_loss, ts_mse, ts_pen, pred, target, jtlayers = test(ts_dataset, model, criterion, device)\n",
    "\n",
    "ts_accuracy_1 = r2_score(pred[:, 0], target[:, 0])\n",
    "ts_accuracy_2 = r2_score(pred[:, 1], target[:, 1])\n",
    "ts_accuracy_3 = r2_score(pred[:, 2], target[:, 2])\n",
    "\n",
    "print(ts_accuracy_1, ts_accuracy_2, ts_accuracy_3)\n",
    "\n",
    "\n",
    "index=0\n",
    "\n",
    "figure(figsize=(8, 6))\n",
    "total_steps = len(pred)\n",
    "x_1 = np.linspace(0,2.499,total_steps)\n",
    "\n",
    "plt.plot(x_1, target[:, index], c='black', label='Ground-truth')\n",
    "plt.plot(x_1, pred[:, index], c='tab:red', label='Unc. Noise free')\n",
    "\n",
    "act_pos=jtlayers[:, 0:3]\n",
    "act_vel=jtlayers[:, 3:6]\n",
    "act_accel=jtlayers[:, 6:9]\n",
    "\n",
    "\n",
    "print('total_loss : {:4f}, mse : {:4f}, pen : {:4f}'.format(ts_loss.item(), ts_mse, ts_pen.item()))\n",
    "\n",
    "\n",
    "test_config = {'save_path' : 'models/model_PGNN.pt'}\n",
    "\n",
    "\n",
    "model = DDNN(config['layers']).to(device)\n",
    "ckpt = torch.load(test_config['save_path'], map_location='cpu')  # Load your best model\n",
    "model.load_state_dict(ckpt)\n",
    "criterion = CustomLoss(mode='mse')\n",
    "\n",
    "\n",
    "def test(ts_dataset, model, criterion, device, lim=35., preds=None, targets=None):\n",
    "    ts_loader = DataLoader(ts_dataset, config['batch_size'], shuffle = False, \n",
    "    drop_last=False, num_workers=0, pin_memory=True)\n",
    "    sample_num = len(ts_dataset)\n",
    "    hookF = [Hook(layer[1]) for layer in list(model._modules.items())]\n",
    "\n",
    "    if preds is None or targets is None:\n",
    "        model.eval()\n",
    "        preds, targets, jtlayers = [], [], []\n",
    "        total_loss = 0\n",
    "        total_ts_mse = 0\n",
    "        total_ts_pen = 0\n",
    "        cnt = 0\n",
    "        for x, y in ts_loader:\n",
    "            x, y = x.to(device), y.to(device)\n",
    "            with torch.no_grad():\n",
    "                pred = model(x)\n",
    "                preds.append(pred.detach().cpu())\n",
    "                targets.append(y.detach().cpu())\n",
    "\n",
    "                jtlayer = hookF[2].output\n",
    "                jtlayers.append(jtlayer.detach().cpu())\n",
    "                ts_loss, ts_mse, ts_pen = criterion(pred, y, x, jtlayer)\n",
    "\n",
    "            cnt = cnt 1\n",
    "            total_loss = total_loss    ts_mse.detach().cpu().item() * len(x)   ts_pen    # ( *len(x) means error**2)            \n",
    "            total_ts_mse = total_ts_mse   ts_mse.detach().cpu().item() * len(x)\n",
    "            total_ts_pen = total_ts_pen   ts_pen\n",
    "\n",
    "        total_loss = total_loss / len(ts_loader.dataset)            # Calculate validation loss with each epoch (Criterion within CLASS model)\n",
    "        total_ts_mse = total_ts_mse / len(ts_loader.dataset)\n",
    "        total_ts_pen = total_ts_pen / len(ts_loader.dataset)\n",
    "        preds = torch.cat(preds, 0).numpy().reshape((sample_num, 3))\n",
    "        targets = torch.cat(targets, 0).numpy().reshape((sample_num, 3))\n",
    "        jtlayers = torch.cat(jtlayers, 0).numpy().reshape((sample_num, 9))\n",
    "    return total_loss, total_ts_mse, total_ts_pen, preds, targets, jtlayers\n",
    "\n",
    "\n",
    "from sklearn.metrics import r2_score\n",
    "\n",
    "\n",
    "ts_loss, ts_mse, ts_pen, pred, target, jtlayers = test(ts_dataset, model, criterion, device)\n",
    "\n",
    "\n",
    "ts_accuracy_1 = r2_score(pred[:, 0], target[:, 0])\n",
    "ts_accuracy_2 = r2_score(pred[:, 1], target[:, 1])\n",
    "ts_accuracy_3 = r2_score(pred[:, 2], target[:, 2])\n",
    "\n",
    "print(ts_accuracy_1, ts_accuracy_2, ts_accuracy_3)\n",
    "\n",
    "print('total_loss : {:4f}, mse : {:4f}, pen : {:4f}'.format(ts_loss.item(), ts_mse, ts_pen.item()))\n",
    "\n",
    "plt.plot(x_1, pred[:, index], c='tab:green', label='Con. Noise free')\n",
    "plt.xlabel('Time (s)')\n",
    "plt.ylabel('Force (Newton)')\n",
    "plt.title('Force{} prediction result'.format(index 1))\n",
    "plt.legend()\n",
    "plt.savefig(\"High resoltion.png\",dpi=300)\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **Write to Excel**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Finished reading the Trajectory Dataset (10000 samples found, each dim = 9)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Wilson Chen\\AppData\\Local\\Temp\\ipykernel_36272\\2336131126.py:60: DeprecationWarning: Call to deprecated function remove_sheet (Use wb.remove(worksheet) or del wb[sheetname]).\n",
      "  wb.remove_sheet(wb.active)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9969158279814434 0.99682119039475 0.9967562912384207\n",
      "total_loss : 256.697388, mse : 250.655481, pen : 6.041882\n",
      "0.9969158279814434 0.99682119039475 0.9967562912384207\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgMAAAGDCAYAAAC2gxMSAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8qNh9FAAAACXBIWXMAAAsTAAALEwEAmpwYAAB7pUlEQVR4nO3dd3hcxdXA4d/ZVW WJRetJPeObbljm2rAGGMDpvdiQg09gYT6JSQhlARSKEkg9F5DNwYMpofi3ovcZfXet873x64cYSxpJW2TdN7n2Qfp7r0zZ9dCOjt35owYY1BKKaVUz2UJdwBKKaWUCi9NBpRSSqkeTpMBpZRSqofTZEAppZTq4TQZUEoppXo4TQaUUkqpHk6TAaVU0InIYBExIhLl /4DEbmoA 0MFJFaEbEGPsrQEJGdIjI73HEo1ZwmA0oFie XfoPvj1fTIzOE/R8lIktFpEpEdoaqX38YY443xjzT1nn7/ E0xuw2xiQZY9zBjTA0ROROEXk 3HEopcmAUsF1ou PV9Mjvz0XN32S7qA64EngV51o44A6GVfE6 6vT6n9aTKgVIiJSKyI/E1E8n2Pv4lIrO 5WSKSJyI3i0gh8JSIWEXkNhHZJiI1IrJcRAb4zh8tIh LSLmIbBaRM5v6McZ8b4x5DtjuR0xNw/iX 2IqEJGbmj1/p4i8LiLPi0g1sFBEeonIE75z94rIXU3D976Y7xeRUhHZDszfr7/PROTSZt9fJiIbfa9vg4hMFpHngIHAu75RlV8f4HZDpoi843v9uSJy2X4xvyoiz/raXS8iU1t5D4yIXC0iW4GtvmMniMgqEakUkW9EJKfZ Tf7XneN770/xnf8aRG5q9l5s0Qk7wD9zQVuA87yvb7Vbf07KRUsmgwoFXq3AzOAicAE4GDgjmbPZwBpwCDgcuCXwDnAPCAF BlQLyKJwMfAi0A/4GzgHyJyUCdiOwoYAcwBbt7v3vYC4HUgFXgBeBpwAcOBSb5rmv7AXwac4Ds FTi9pQ5F5AzgTuBC3 s7CSgzxlwA7OZ/oyt/OsDlLwN5QKavj7tF5Ohmz5/kOycVeAd4uI3XfzIwHThIRCbhHVm5AkgHHgXe8SVzo4BrgGnGmGTgOGBnG23/iDFmMXA38Irv9U1oz/VKBZImA0oF11u T5WVIvKW79h5wO NMcXGmBLgd8AFza7xAL81xtiNMQ14/8DeYYzZbLxWG2PK8P6x3WmMecoY4zLGrATeAM7oRLy/M8bUGWPWAk/hTUKa/NcY85YxxoP3j/Y84Abf cXAX/EmJABnAn8zxuwxxpQD97TS56XAn4wxP/heX64xZldbgfpGRw4FbjbGNBpjVgGP400qmnxljFnkm2PwHN7kqzX3GGPKfe/75cCjxpjvjDFu3xwHO95Ezg3E4k0aoo0xO40x29qKWalIpffFlAquk40xS/Y7lgk0/2O3y3esSYkxprHZ9wOAA/2hGQRMF5HKZsei8P7R66g9 8U1voXnBgHRQIGINB2zNDsn8wBttaSl19eWTKDcGFOzXz/NbwUUNvu6HogTkShjjKuFNvd/jReJyLXNjsUAmcaYz0XkBrwjGmNF5EPgl 2dE6JUpNCRAaVCLx/vH5omA33Hmuy/legeYNgB2tkDfG6MSW32SDLG/LwTsQ3wM649eD8l92nWd4oxZqzv YIDtNWSll7f/n3uLx9IE5Hk/frZ28o1bdn/Nf5xv/c3wRjzEoAx5kVjzGF4/y0NcJ/vujogoVk7GX72p1TYaDKgVOi9BNwhIn1FpA/wG6C15WWPA38QkRHilSMi6cB7wEgRuUBEon2PaSIyBkBELCISh/cTvIhInIjEtBHb/4lIgoiMBS4GXjnQScaYAuAj4AERSfH1NUxEjvSd8ipwnYhki0hv4JY2Xt9NIjLF9/qGi0hTslQEDG0hhj3AN8A9vteWA1xC6 9le/wbuFJEpvviShSR SKSLCKjRORo8U78bAQa8N7eAVgFzBORNBHJAG5opY8iYLCI6O9iFVb6A6hU6N0FLAPWAGuBFb5jLfkL3j uHwHVwBNAvG94fA7e /T5eIfE78N7LxvgCLx/pBbh/cTc4GujNZ8DucAnwP3GmNbOvxDvsPkGoALv5EKb77l/Ax8Cq32v7z8tNWKMeQ34I96JkDXAW3gnUIJ3rsEdvjkXNx3g8nOAwXhf/5t451rsf1umQ4wxy/BOhHwY7 vLBRb6no4F7gVK8b7v/YBbfc89h/d178T7fh8wofJ5zfffMhFZEYi4leoIMUZHqZTq6URkMLADiG7lfrpSqpvSkQGllFKqh9NkQCmllOrh9DaBUkop1cPpyIBSSinVw2kyoJRSSvVwPbYCYZ8 fczgwYPDHYZSSikVEsuXLy81xvQ90HM9NhkYPHgwy5YtC3cYSimlVEiISItlwfU2gVJKKdXDaTKglFJK9XCaDCillFI9XI dM6CUUqplTqeTvLw8Ghsb2z5ZRZS4uDiys7OJjo72 xpNBpRSSv1EXl4eycnJDB48GBEJdzjKT8YYysrKyMvLY8iQIX5fp7cJlFJK/URjYyPp6emaCHQxIkJ6enq7R3Q0GVBKKXVAmgh0TR35d9NkQCmlVEQqKiri3HPPZejQoUyZMoWZM2fy5ptvhqz/nTt3Mm7cuAM 9/TTT5Ofn9/uNletWsWiRYv2fX/nnXdy//33dzjGQNFkQCmlVMQxxnDyySdzxBFHsH37dpYvX87LL79MXl7ej85zuVxhia 1ZMDtdrd43f7JQKTQZEAppVTE fTTT4mJieHKK6/cd2zQoEFce 21PP3005x00kkcffTRHHPMMZSXl3PyySeTk5PDjBkzWLNmDfDTT93jxo1j586d7Ny5kzFjxnDZZZcxduxY5syZQ0NDAwDLly9nwoQJTJgwgUceeeSAsb3  ussW7aM8847j4kTJ9LQ0MDgwYO5 eabmTx5Mq 99hqzZs3aV W2tLSUwYMH43A4 M1vfsMrr7zCxIkTeeWVVwDYsGEDs2bNYujQoTz44INBeT/boqsJlFJKteqGG25g1apVAW1z4sSJ/O1vf2vx fXr1zN58uQWn1 xYgVr1qwhLS2Na6 9lkmTJvHWW2/x6aefcuGFF7YZ79atW3nppZf497//zZlnnskbb7zB eefz8UXX8zDDz/MEUccwa9 9asDXnv66afz8MMPc//99zN16tR9x9PT01mxYgUA//rXv35yXUxMDL///e9ZtmwZDz/8MOBNWDZt2sTSpUupqalh1KhR/PznP2/XssBA0JEBpSKYq6KChlWrMB5PuENRKqyuvvpqJkyYwLRp0wA49thjSUtLA Crr77iggsuAODoo4 mrKyM6urqVtsbMmQIEydOBGDKlCns3LmTyspKKisrOeKIIwD2temvs846q13nN5k/fz6xsbH06dOHfv36UVRU1KF2OkNHBpSKUI49e9hyyqlYa2uJm30Mgx96SGd3q7Bo7RN8sIwdO5Y33nhj3/ePPPIIpaWl z6JJyYmttlGVFQUnmaJdPPldrGxsfu tlqt 24TtOTiiy9m5cqVZGZmtnjPv3lMzftua5nf/rGEYx6EjgwoFaG ueQS6quqeLWyksYln1D24UfhDkmpkDn66KNpbGzkn//8575j9fX1Bzz38MMP54UXXgDgs88 o0 fPqSkpDB48OB9w/YrVqxgx44drfaZmppKamoqX331FcC NgGeeuqpH03 S05OpqampsW2Bg8ezPLlywHvHIMmbV0XLpoMKBWBcr/6iv679/B9Rn/GP/Qg U4nq  9J9xhKRUyIsJbb73F559/zpAhQzj44IO56KKLuO   35y7p133sny5cvJycnhlltu4ZlnngHgtNNOo7y8nLFjx/Lwww8zcuTINvt96qmnuPrqq5k4cSLGmBbPW7hwIVdeeeW CYT7u mmm/jnP//JpEmTKC0t3Xf8qKOOYsOGDT aQBgJpLUX251NnTrVNM30VCrSPDtvHpO3baf3a6 SlZPDo0ccwaFFxWR/9CGpgwaFOzzVA2zcuJExY8aEOwzVQQf69xOR5caYqQc6P JGBkTkzyKySUTWiMibIpLa7LlbRSRXRDaLyHHNjs/1HcsVkVvCErhSAdLY2Ej8ps0UJCeTlZMDwIQrrsAqwlf/ EeYo1NKdUcRlwwAHwPjjDE5wBbgVgAROQg4GxgLzAX ISJWEbECjwDHAwcB5/jOVapL uSttzgoOprkww/fd zgs86ixhhKlnwSxsiUUt1VxCUDxpiPjDFNUym/BbJ9Xy8AXjbG2I0xO4Bc4GDfI9cYs90Y4wBe9p2rVJe0 sWXsIhw0Lnn7DtmiYqizJZBVmUldXV1YYxOKdUdRVwysJ fAR/4vs4C9jR7Ls93rKXjPyEil4vIMhFZVlJSEoRwleocYwyVq1djgKT9aqL3PvhgBkRH80UEljJVSnVtYUkGRGSJiKw7wGNBs3NuB1zACy231D7GmMeMMVONMVP79u0bqGaVCpitW7eS6XLSmJqKJT7 R8 Nmj8fgLVvvRWGyJRS3VlYig4ZY2a39ryILAROAI4x/1vusBcY0Oy0bN8xWjmuVJfy eefMyImlvhRo37yXMrEiRQC9WvWhj4wpVS3FnG3CURkLvBr4CRjTPMKE 8AZ4tIrIgMAUYA3wM/ACNEZIiIxOCdZPhOqONWKhC  ewzBsXEkH6AmuzWXr2oSUwkvaKixeIrSnUXB9o OBjb/T799NNYLJZ9mxvB/zY0as28efOorKzsVN8lJSVMnz6dSZMm8eWXX3aqrc6KuGQAeBhIBj4WkVUi8i8AY8x64FVgA7AYuNoY4/ZNNrwG BDYCLzqO1epLifv22 xihA3qoXiKCNGMDY2Fq2RoVTgZGdn88c//rFd1yxatIjU1NRO9fvJJ58wfvx4Vq5cyeHNVg9B69sgB0PEJQPGmOHGmAHGmIm x5XNnvujMWaYMWaUMeaDZscXGWNG p5r37 oUhGioqKCBF lstgRIw54TsZhh2KLjmbFp5 GMjSlIs6sWbO4 eabOfjggxk5cuS T9Zut5ubbrqJcePGkZOTw0MPPdRmWyeccALr169n8 bNP3nupZdeYvz48YwbN46bb7553/HBgwdTWlpKXV0d8 fPZ8KECYwbN25fVcHly5dz5JFHMmXKFI477jgKCgp 1O6qVav49a9/zdtvv72vimFSUhI33ngjEyZM4L///S/PP/88Bx98MBMnTuSKK67YlyB89NFHzJw5k8mTJ3PGGWdQW1vb4fexiW5UpFSEWLlyJSNjYzFRUcS0UGWwz4wZ1D38CCX//W Io1M9WeHdd2PfuCmgbcaOGU3Gbbd1qg2Xy8X333/PokWL N3vfseSJUt47LHH2LlzJ6tWrSIqKory8vI227FYLPz617/m7rvv3lfKGCA/P5 bb76Z5cuX07t3b bMmcNbb73FySefvO cxYsXk5mZyfvvvw9AVVUVTqeTa6 9lrfffpu ffvyyiuvcPvtt/Pkk0/uu27ixIk/2c64rq6O6dOn88ADD7Bx40buu 8 vv76a6Kjo7nqqqt44YUXmDdvHnfddRdLliwhMTGR  67j7/85S/85je/6dR7qcmAUhFixYoVjIiNJWrIECTqwP9rxh10EB7A2sb9TKW6upZ26Gx /NRTTwX twUxwJIlS7jyyiuJ8v0/1LTNcVvOPfdc/vjHP/5oM6MffviBWbNm0bT67LzzzuOLL774UTIwfvx4brzxRm6  WZOOOEEDj/8cNatW8e6des49thjAe9ohc1mazMGq9XKaaedBnhvISxfvnzfls0NDQ3069ePb7/9lg0bNnDooYcC4HA4mDlzpl vsTWaDCgVIVasWMGV8Qkkjhnd4jmWhASqe/ViQH4 dXV1fm3jqlRndfYTfEekp6dTUVHxo2Pl5eUMGTJk3/dNW/8GYtvfqKgobrzxxgNuhNSakSNHsmLFChYtWsQdd9zBMcccwymnnMLYsWP5bztH8OLi4rBarYC35shFF13EPff8eIOyd999l2OPPZaXXnqpXW23JeLmDCjVU21ZsZK FgtxbeysZsaMYVJ8HOtWrQpNYEqFQVJSEjabjU9982PKy8tZvHgxhx12WKvXHXvssTz66KP7kgN/bhM0WbhwIUuWLKGpKN3BBx/M559/TmlpKW63m5deeokjjzzyR9fk5 eTkJDA eefz69 9StWrFjBqFGjKCkp2ZcMOJ1O1q9v37z2Y445htdff53i4uJ9r2PXrl3MmDGDr7/ mtzcXMB7a2HLli3tavtANBlQKgLU1dXBnt0AxLaRDPQ7 igSLVa2L1kSitCUCptnn32WP/zhD0ycOJGjjz6a3/72twwbNqzVay699FIGDhxITk4OEyZM4MUXXwTgN7/5De 80/qq85iYGK677rp9f4BtNhv33nsvRx11FBMmTGDKlCksWPDjavdr167dN8nvd7/7HXfccQcxMTG8/vrr3HzzzUyYMIGJEyfyzTfftOu1H3TQQdx1113MmTOHnJwcjj32WAoKCujbty9PP/0055xzDjk5OcycOZNNmzo/n0O3MFYqAixfvpz7Z8/mN/0zGL70U6Jbub/oKC5m2xFH8vXwYVz63nshjFL1JLqFcdfW5bcwVqon2rhxo3clQUICURkZrZ4b068f RYLvXbvafU8pZTylyYDSkWAjRs3MiI2jvhRo1qcRd1ccd  DLXbMZ2cNKWUUqDJgFIRYeOGDYyOi2u58uB XKNHkyBC6Q8/BDkypVRPoMmAUhGgeOMmEkXanDzYpPfh3hnVeYs/DGZYqofrqXPKurqO/LtpMqBUmDmdTqIL8gHaXFbYZPjUqex0OGhYsSKYoakeLC4ujrKyMk0IuhhjDGVlZcTFxbXrOi06pFSY5ebmMtRXLa2lPQn2N2zYMBbZ7Ry Ny YoakeLDs7m7y8vH1r7lXXERcXR3Z2druu0WRAqTDbtGkTI2Ji8fTujbVXL7 uiY6OpjgpkaT6BtyVlVg7uXuaUvuLjo7 UbU/1b3pbQKlwmzTpk0MiYkh3s9RgSZOX bfuOmnO60ppVR7aDKgVJjlbt3K0LhYEoa3XlltfwljxwLQsHFDMMJSSvUgmgwoFWbFW7eSJBZiBrdvSDZ77FiKXU4qVqwMUmRKqZ5CkwGlwsy127snQcyQwe26bujQoWxutNO4cWMQolJK9SSaDCgVRo2NjSRVVQMQ087JWkOHDmWbw4GlsBDj8QQjPKVUD6HJgFJhtHPnTgbHROOxWlvdnOhABg4cyA6nA4vLhTO/IEgRKqV6Ak0GlAqjbdu2MTg6BpORgVit7bo2Ojqaut69AXDs2B6M8JRSPYQmA0qF0bZt2xgcE0P80KEdut46cCAA9m3bAhmWUqqH0WRAqTDavnUrA2NiSBo9qkPX9x8xgiqPB8f2HQGOTCnVk2gyoFQYVW7eTJQIsUM6NjIwdOhQchsbacjNDXBkSqmeRJMBpcLItXsPADGDB3fo qFDh7Ld4aBRkwGlVCdoMqBUmHg8HuLKy4D21xhoMnToUHY6HEh1Ne6qqgBGp5TqSTQZUCpM8vPzybZYcMbFEeVbFdBeQ4cOZbfTAYBjj 5gqJTqGE0GlAqTHTt2MDgmBk9GRofbSEtLozw6GgBnniYDSqmO0WRAqTDZs2cPQ2JiiO3ENrEigtWXTDj3ajKglOoYTQaUCpP87dvpFxVNr4PGdKqdPoMGUQs49uwJTGBKqR5HkwGlwqR 61YAkkZ2rMZAk4EDB5LncuHM2xuIsJRSPZAmA0qFidv3Sb6jKwmaDBw4kJ319dh9ux8qpVR7aTKgVJjElJTgAWJ8JYU7auDAgexxOnHl52Pc7sAEp5TqUTQZUCpMUurqqImNxRIX16l2Bg4cyF6nE1wuXMXFAYpOKdWTaDKgVBjY7XZsHkNDWsfqCzS3LxkAnPn5nW5PKdXzaDKgVBjk5eUxOCYak5nZ6baysrIocrsAcBYWdro9pVTPo8mAUmGQv349iRZrh7cubi4mJgbS0wFwaTKglOoATQaUCoOydesA6H3QQQFpr8/AgTSI4CwsCkh7SqmeRZMBpcKgbvt2ADImTAhIewMHDqTE48FVWBCQ9pRSPYsmA0qFgWvvXjzG0Gv48IC0N3DgQPIaG3VkQCnVIZoMKBUG0aVllIkgMTEBaW/gwIHstdux79UqhEqp9ovYZEBEbhQRIyJ9fN LiDwoIrkiskZEJjc79yIR2ep7XBS qJXyT2JdHVWdrC/Q3MCBAylyOfFUVGAcjoC1q5TqGSIyGRCRAcAcoHl91eOBEb7H5cA/feemAb8FpgMHA78Vkc4v3lYqiPq4XDSmpgasvQEDBlDodCHG4CwuCVi7SqmeISKTAeCvwK8B0 zYAuBZ4/UtkCoiNuA44GNjTLkxpgL4GJgb8oiV8lN9eTlpFgtk9A9Ym1lZWRS6vIWHdBKhUqq9Ii4ZEJEFwF5jzOr9nsoCmu/Rmuc71tJxpSLS3uXLAYgbNDhgbfbt25dS39c6iVAp1V5R4ehURJYAGQd46nbgNry3CILR7 V4bzEwsJObwyjVUaXr15MK9Bo1MmBtWiwWpE8fQEcGlFLtF5ZkwBgz 0DHRWQ8MARYLSIA2cAKETkY2AsMaHZ6tu/YXmDWfsc/a6Hfx4DHAKZOnWoOdI5SwVazNZdUoH Aagw0ScvKor6iEmeBViFUSrVPRN0mMMasNcb0M8YMNsYMxjvkP9kYUwi8A1zoW1UwA6gyxhQAHwJzRKS3b LgHN8xpSKSMy PKrebAWPGBLTdrKwsSo1Hdy5USrVbRCUDbVgEbAdygX8DVwEYY8qBPwA/ B6/9x1TKiJZS0oo9HhISEgIaLvZ2dkUNtpxlehqAqVU 4TlNoG/fKMDTV8b4OoWznsSeDJEYSnVKQk1NRTExga83aysLArtdhxFOoFQKdU XWlkQKkuz3g8pDqdNKSkBLztrKwsil0u3CUleHNnpZTyjyYDSoWQq6SUaIB fQPedlZWFiVuF7hceKqqAt6 Uqr70mRAqRCq9e1WGJOdHfC2s7KyKHG5AHTegFKqXTQZUCqEitetAyAlQLsVNpeZmbkvGXDqigKlVDtoMqBUCFVu3QJA37FjA952fHw8jsQkQEcGlFLtE9GrCZTqbhp378HpcpE9dGhQ2o/N6A9OlyYDSql20ZEBpULIFBWx1 kkOwhzBgD6ZGdTD7h050KlVDtoMqBUCMVUVlJqEeLj44PSflZWFqVut44MKKXaRZMBpULEeDwkNTZSl5gYtD6ys7MptNtxFmvhIaWU/zQZUCpE3GVlRBuDKy0taH14Cw85setmRUqpdtBkQKkQce7dC0BUxoF27w6MploDnrIyrUKolPKbJgNKhUjtjh0AJAweErQ vMmAG3E48NTWBq0fpVT3osmAUiFStnEjAL1HjQxaH023CQDdylgp5TdNBpQKkdrtOyh3ucgaNixofaSnp1Np8f5vrSsKlFL 0mRAqRBxFORT6HIxYMCAoPUhItC7N6AjA0op/2kyoFSISHkFhS4nWVlZQe0npn9/QEcGlFL 02RAqRCJramhJiqauLi4oPbTOzOTBrQKoVLKf5oMKBUCHrudeJcLe3JS0Puy2WyUuly4SvQ2gVLKP5oMKBUCTffvJT096H3ZbDYKHQ4cRVqFUCnlH00GlAoBV6G3ImBMhi3ofWVmZlLicmkVQqWU3zQZUCoEanbtBiBp8KCg92Wz2Sj2VSFUSil/aDKgVAiU524FIH1k8AoONWmaMyB2O 7auqD3p5Tq jQZUCoEanfuos7jJjOIBYeaNI0MgNYaUEr5R5MBpULAXlBAoTO4BYea9O3blzLjAbTWgFLKP5oMKBUKpaUUu1xBLzgEYLFY8KSmAjoyoJTyjyYDSoVAdG0tVVFRxMbGhqa/fv0AHRlQSvlHkwGlgsy43STa7diTgl9wqEmqzUYjOjKglPKPJgNKBZmrrAwrQFrvkPWZmZVFmdutIwNKKb9oMqBUkLmKvJ/Oo0NQcKiJzWajwG7HqVUIlVJ 0GRAqSCr2bkDgMSBwV9J0MRms/mqEBaErE lVNelyYBSQVa2xVtwKG3EiJD12VRrwK1VCJVSftBkQKkgq965A6cxZISg mATm81GiduFNDZqFUKlVJs0GVAqyOz5 ZS4XAwYFPx9CZo03SYAdCtjpVSbNBlQKsg8paUUupwhKTjUpH///pS43AC4inVFgVKqdZoMKBVk0dU1VFujiImJCV2f0dG4UpIBLTyklGqbJgNKBZExhsTGRhoTE0Pe974qhFp4SCnVBk0GlAoiT20tscbg6Z0a8r572WzY0ZEBpVTbNBlQKohchYUARPfPCHnftsxMyjweHRlQSrVJkwGlgqhqR gLDjXZV4VQkwGlVBs0GVAqiEo3bwYgddiwkPdts9kodjqx 0YnlFKqJRGZDIjItSKySUTWi8ifmh2/VURyRWSziBzX7Phc37FcEbklPFEr9VNNIwP9x4wJed9NtQbcpaUh71sp1bVEhTuA/YnIUcACYIIxxi4i/XzHDwLOBsYCmcASEWkq6fYIcCyQB/wgIu8YYzaEPnqlfqxx717KXC6yhwwJed ZmZkUu11IQwOeujosYVjRoJTqGiJxZODnwL3GGDuAMabphucC4GVjjN0YswPIBQ72PXKNMduNMQ7gZd 5SoWdp6SEYpcrpAWHmvy4CqGuKFBKtSwSk4GRwOEi8p2IfC4i03zHs4A9zc7L8x1r6bhSYWetqqbKaiU6OjrkfWdkZOxLBnQSoVKqNWG5TSAiS4ADrbW6HW9MacAMYBrwqogMDVC/lwOXAwwcODAQTSrVqsSGBuoTEsLSd1xcHHZf3zoyoJRqTViSAWPM7JaeE5GfA/8xxhjgexHxAH2AvUDz9VnZvmO0cnz/fh8DHgOYOnWq6fALUMoPHoeDJI8HT3p62GKw9u0LHqP7EyilWhWJtwneAo4C8E0QjAFKgXeAs0UkVkSGACOA74EfgBEiMkREYvBOMnwnHIEr1ZyzqAgAa/9 YYuhl82GAx0ZUEq1LuJWEwBPAk KyDrAAVzkGyVYLyKvAhsAF3C1McYNICLXAB8CVuBJY8z68ISu1P9UbdsOQEJ2dthiyLDZKMvLo4/OGVAh4nE4qHjueew7tpN66mkkTJ4U7pCUHyIuGfCtCDi/hef CPzxAMcXAYuCHJpS7VKyaRMWoNfQgEx56RCbzUah3cEwTQZUCBhj2H39DTQsXYpdhMr/vMmgJx4ncebMcIem2hCJtwmU6haqtntHBsJRcKiJzWajyOHAXqRVCFXwVb3/Pg1Ll/JASQmnV1awvaGB7b/4JZ7GxnCHptqgyYBSQVKfl0e9x0PWyJFtnxwkNpuNErcLt04gVEFmPB52/PFuNjQ2MuX3v2N1bi7PJ8QTVVlJyYsvhTs81YY2kwERmSkij4jIGhEpEZHdIrJIRK4WkV6hCFKprshdXEyRy0VmZmbYYrDZbBS7XOCrQqhUsNR 9RVxFRV8kpzMFT//OYmJiVz58MOsbKhnz6OP4p36pSJVq8mAiHwAXIp3ct5cwAYcBNwBxAFvi8hJwQ6yuzEuF 7KynCHoYLMWlVJhcUSloJDTTIyMrzJALqiQAVX7qOPUeZyceiNv8Ri8f5pmXXUUSxPSyOxqor6FSvCHKFqTVsjAxcYYy4xxrxjjMk3xriMMbXGmBXGmAeMMbOAb0IQZ7dR9 13bJ11FFtmzGTPlT/HXauf1rqr PoGGuLjwxqDzWajVJMBFWQeux3LypV8Zm/klDPP/NFzE37 c o9Hrb869EwRaf80WoyYIxpc7szf85RXs78fPKuuQZrSgrpl19O7ZdfsvfGX0bc8FlZWRn33nsvN9xwA1988UW4w mSjNtNituNq1d476SlpKRQZfUuGtKSxCpYar75hhiPB9ekySTsV3Hz5LPP5vP6evj2W4wvMVWRx68JhCJyqohsFZEqEakWkRoRqQ52cN1Nyd8fxDgcDHjsUfr98hf0v UW6j7/gup3IqdG0ubNm5k0aRKP/fa3LHnySY488kjuueeecIfV5bhKy4gCrP3CV3AIQESw9u3jjUknEaog2fHKq9S63Rx01pk/eS4lJYXi4cOIczqp /77MESn/OHvaoI/AScZY3oZY1KMMcnGmJRgBtbdOIuKqHr3XXqfcw4xviI0vc87l7jx4yl56GGM0xnmCKG2tpbzFizg91HRvD9kKK9lZrFoylT 9H//x6JFWsahPSq25QIQlx3 PbNSMjK8VQh1ZEAFiWPZMr5vaGD28ccf8PnR55xDvcfDzpdeDnFkyl/ JgNFxpiNQY2km6t6623weOh97jn7jonFQp rr8KZl0fV   HMTqv395 O79oaODghAT63XQj/W 7lSFOJ0 PGMlVV1xBo64V9lvxRu//LuEsONQkw2ajHJ0zoILDWVRMcm0tJX37kpaWdsBzZs fz5d1tbi  grjdoc4QuUPf5OBZSLyioic47tlcKqInBrUyLqZqnffIX7qFGIGDfrR8aQjjyRm2DAqw5wx5 bm4nn5FSbHJ5D9p/tIv/RS0i68ENvdf2SkMcyur fRR3UCkL8qcrcB0HfUqDBH4qtC6HDoyIAKispvvgYgacb0Fs/Jzs5mXVISsQ0NNK5dG6rQVDv4mwykAPXAHOBE3 OEYAXV3Tj37sWRu43k2T/drFFESD3jdBpWr6Zx85YwROf1j9/8hoWpqUQfO5uUefP2He81fz5JxxzD1f3689gDD DWrN4vDXl7cBpD1kEHhTsUbzLQ2IjDt3GSUoG0Z/Fiat1uRs d2 p5KYcfDkD1V1 HIizVTn4lA8aYiw/w Fmwg suar/8EoCkI44AYPv27SxduhS73Q5ArwULkOhoqv7zn7DEV1xczMAvvoSoKAbd8X8/eb7v9dcRZwxH1zfwwQcfhCHCrsdV5Cs4lBX OQNNhYd0ZEAFg3P1alY2NDDj0ENbPe/Q449nU2MjRUuWhCgy1R7 ribIFpE3RaTY93hDRMK3FVsXU/vlV0RnZREzZAgPP/wwI0aM4Oijj2b8 PHs2LGDqN69STziCKo//BDj8YQ8vufuu49jExOJOflkog w3W7cyJEkHnUU56Sl8fi//hXy LoiS2UFFRYhKir8e4HZbDaK3d4qhFrXQgWSu7aOpMoq8pIS6dOnT6vnHnbYYXxbX49l61bdqyAC Xub4CngHSDT93jXd0y1wXg81C9bRsLMGXzzzTdcf/31zJs3j5dffpnS0lLmz59PQ0MDKXPn4iospGHV6pDG53a7kbfewm2xMOKXv2jxvLRzz6G3xYLz66 p1OqJbYqvq6c LrwFh5pkZGRQsq/wkI4OqMBp3LgBCxAzuu3NuPr378/O5GSsbnfIf8 ptvmbDPQ1xjzlq0DoMsY8DfQNYlzdhmPnLjxVVcRPnMgtt9xC//79eemllzjrrLN4 eWX2bhxI3fffTdJRx2FxMRQvTi0w/BfvPUWs6OiqTr4YKLS01s8L/GQQ/CkpXFyUhJvv/12CCPseowx9HK5wl5wqInNZmuWDOiKAhU4hb77/xmHHuLX UlTJgPQsGZN0GJSHeNvMlAmIueLiNX3OB8oC2Zg3UXDqlUAbIuK4quvvuLWW28lKSkJgDlz5nDmmWfy17/ lbL6OhKPOJyaDz8KaUXC3H/8Eysw4Y7bWz1PrFb6nnEGMxMSWfSyrhVujbuiklgRLP0iI1/u27cvpb7bT1p4SAVS6fffU h0Mvaww/w6f Lhh7PT4aD8u2 DHJlqL3 TgZ8BZwKFQAFwOrAwSDF1Kw2rVmFJTubpjz8mNjaWCy 88EfP//73v6ehoYG//e1vJB8zG1dREfaNoSnpUFddzYjdu9jVuzcpI0a0eX7K3OOwimD5YRn19fUhiLBrKt2yGYD4rMiYVmOxWBDf m dRKgCalsu6 2NjB8/3q/TZ86cyZqGBhpXr4m4Muw9nb/JQLYx5iRjTF9jTD9jzMnAwCDG1W00rF5N3PjxvPzKKyxYsIBe w0djxo1ihNPPJHHH3 cmOnTQYSazz4LSWyf/ Uv9LdG0fvss/w6P3b0aFx90jkqPo7PQhRjV1S0YQMQGQWHmqRkZOAQ0dsEKmDctbUkVddQmJC4b7SzLRMmTGCj20V0bS2ugoIgR6jaw99k4CE/j6lmPA4H9txcKtLTKCsr45RTTjngeVdffTUlJSX859NPiMsZT 1nn4ckPuf771Pq8XDwVVf5db6IkH7CicxMSOTTd98NcnRd176CQ2NGhzmS/9lXhVBHBlSANG7wTh70tCPpjY6OpmHAAEDnDUSaVpMBEZkpIjcCfUXkl80edwLWkETYhTm2bQO3m5XlFYgIsw9QdAjgmGOOYfjw4Tz55JMkz5pF45o1uEqDuxlkbV4ew6pr2D54MNaYGL vS51zLFEilH7ySRCj69rq9 zBbQzZ4/wbOg0Fm81GkVOrEKrAqVztXRGQPnVKu65LnzwZlzE0huh2qPJPWyMDMUASEAUkN3tU4503oFrRuNl77/j99euYMmVKi twLRYL5513Hp999hn1vop1tZ8Hd vglQ8/glWE7HPObtd18Tk5OGNivInE9u1Biq5rcxUVUeZ20z/TFu5Q9rHZbOQ3NODU2wQqQIp  IEKl4sxM2e267qcKVPY7rBTsXJVcAJTHdJqMmCM dwY8zvgcGPM75o9/mKM2RqiGLss 5atEBPDe8uXc8wxx7R67nnnnYcxhtd  IGo/v2p/Ty4twocn3zCVqeTw849t13XSVQU0VOmcFhiIks//TRI0XVt3oJDFiwWf /CBZ/NZqPY6cKpJYlVgNi3bmWrw0HOhAntum7ixIlsttuxb9oUpMhUR/j72 oJEdkmIi LyNUiEjnjnxHMvnkzrowM7C4XM2bMaPXcESNGMG3aNF548UWSjjySuq  wjgcQYmrcfduMmtq2D1wALGxse2 PuP448mIjmb94g DEF3XF19XT118ZBQcarKv1oBWIVQBYIwhvqSUXR4PWe0suZ2Tk8MWu4Po6mpcFRVBilC1l797ExwJjME7aTAVeF9EyoMYV7dg37KForg4AKZPb3lHrybnnnsuK1eupHrkCDz19dQvXx6UuNY/ hgAmWee2aHrkw73rik2K1cELKbuJJIKDjXJyMigxK1VCFVguAoLiXW5qE3rjYi069qEhARqfTU4dHQgcvi7N8FhwI3A7cB84D3g6iDG1eW5KipwlZSwvq6WAQMGYLO1ff/41FO9u0K/t2sXEhMTtFUFDR9/zJrGRmafd16Hro 22ajr3Zsxdgc7d 4MbHBdnKumhkQRrH0jo BQk6bNikALD6nOs2/13iWW/bZk91eyry5B46bNAYtJdY6/twk A04GHgNmGWOuMsa8FKygugPHNu/ysq927vRrVABg4MCBTJkyhdfefZeE6dODMm gMTeX9OpqtmVmkpKS0uF24qZNZXJCPF9ovYEfKfTVGIjLjoyCQ00yMjKaJQM6MqA6p863EiBl7NgOXT9y2jSKXU6qV seBZHC32SgD/B7YCawWESWiMgfghdW1 fwfWL ds8epkzxf nNqaeeynfffYd74gQcO3fuaydQtj/zLG5jsJ12WqfayT7uOBItVjYvWhSgyLqHwrXrgMgqOAQQGxuLKzkZ0P0JVOeVrVxFkdPJYD8rD 5v/PjxbG60U7tuXYAjUx3l75yBSmA7sANvOeJhwBHBC6vrc zahbFayXc6/S7VCewrTLS0uhogoKMDxhjqFi/mh/p6jmvnksL9JR18MAAOXR70I1VbtwDQt4OfmIIppX9/bxVCHRlQnWTfuoWtDjsjR47s0PXjxo0j12FHCgowbneAo1Md4e cge3AA0Aa8E9glG9SoWqBY dO6nv1woP3B99fY8aMYdSoUbz06afEDBsW0GSgcf0Gkmpq2JCeRmZmZqfaiurbl9pevRja0ECJftLcp2H3bpzGkJ0TeQtuMmw2KjQZUJ1kPB6iC4vItTs6nAxkZWWRZ7Fgdbtx7tkT4AhVR/h7m2C4MWaeMeZuY8xXxpjgrHnrRhw7d1JstZCUlMTAge3bxuHUU09l6dKlRE0/mLoflgVsKVjBa6/hNIY J54YkPasE3KYEh/Pt998E5D2ugNTVESR20V6hE0ghKZJhE69TaA6xVVQgNXtpjQu9id7rfhLRLAM9E4 tPvmV6nw8jsZEJFPRGQdgIjkiMgdQYyrSzNuN45du8mtr2fcuHHtXnpzyimn4Ha7WQ7gdFL3387/sTUeD9WL3uerujqOP OMTrcHMGDuXJKtVrZ8qPUGmkRXVlFhjWr3v3ko2Gw28usbcOrIgOoEe9M8pk6OLvYe7x0xtefmdjIiFQj JgP/Bm4FnADGmDVA5246d2POgkKMw8HKwqJ23SJoMnXqVAYMGMBzy5ZhSU4OyK2ChpUria2p5fuYmHbNYWhN6qHeegMNy5YFpL3uILmhgbrExHCHcUA2m40ih0NHBlSnNE1qTurgLYImIyZMoMDppGrd gBEpTrL32QgwRjz/X7HXIEOprtw7NoJwLqK8g4lAyLCqaeeygcffUTs9OnUff5Fp/f LnvrbRo8HtLmHhewT63R/ftRFR9PamERLpf OBiHg1SPB3daWrhDOSCbzUaJ24Wpr8dTp1UIVcfUbN5CncdNVicnyTZNIqzVDYsigr/JQKmIDAMMgIicjndVgTqApsx5l8PJQb6Nh9rr1FNPxW63syUlGVdJCY2 9esdYVwuKhct4vPaWub5ChsFimv4MHJiYlivS4RoyNuLRYQoPwpMhcOPag3o6IDqoOqNG9npcDJqdOe26B47dizb7A6suqIgIvibDFwNPAqMFpG9wA3Az4MVVFfn3L0Hd1QUJW4XI0aM6FAbhx56KP369ePlLVtApFO3Cuq /Y6oujqWulwccURgV4T2P II kZFseLDjwLabldUuNa7P3vikMHhDaQFNpuNUk0GVCe59 xhp6PjKwma9OvXj6KYaO Kgvz8AEWnOsrfOgPbjTGzgb7AaGPMYcaYnUGNrAtz5u lJi6OmJgYBgwY0KE2rFYrJ598Mq99 CGx48Z1Khmoev99ao0hedaRxMTEdLidA8mePRuA4i Du VyV1C20VtnvfeoUWGO5MD2bVYEOolQdYjH4SCmspLdTidDO1lYS0Sw MoZ6yTC8Gs1GRCRC5s/gNOAU5p9rw7AuTefImDo0KFYrdYOt3PqqadSW1tLvs1G45q1Hfo053E4qFy8mI rqznBV9AokOJGjsQugnWr/s9cu307bmPI7MA8kVBITk6m1pcM6siA6gjn7t0IUNurV4d2PN1fqu//Fbv /gi7tkYGprXw APwZHBD67qce/eyq76e4cOHd6qdo446itTUVN4sLABjqO7AlsG1n3yCpaGBjxsbOOGEEzoVz4GI1UpVRgZD7HbKysoC3n5X4szPp9jlYkCElSJuIiIk9u PS0STAdUhTfOhrAM7NuK5vxETJ1LkdFKxZk1A2lMd12oyYIy5tukBXAd8B8wCvgUmBz 8rsddW4e7qopNZWWdTgZiYmI48cQTeWrJEmJGjqT6/ffb3UbF629Q5PGQftRRJPtq0wdawuRJjIqN5fsvevatAktZKUXGBO19DoSMjAwqLRZNBlSH2HfsBCC1k5MHm4wZM4ZtDgd1W7YEpD3VcW3OGRCRKBG5FNgIzAZON8ac5as1oPbjzN8LwM6G g5PHmzutNNOo6KigpJRo2hYtQpHXp7/sRQUUPfNN7xRXs4ZZ53V6VhaMnjuXKwibO/AyEV3El9bR00Ahk6DyWazUep2aTKgOqRy4wZKXS6GBOhW2JgxY9jhcGApLOz08mnVOW3NGbga2ABMAeYaYxYaY4K6AbWITBSRb0VklYgsE5GDfcdFRB4UkVwRWSMik5tdc5GIbPU9LgpmfG1x7vUmA/lOZ6dHBgDmzJlDcnIyL V563dXv //LoGVb76JGMMieyPz58/vdCwtSZs HQ/QsGpl0PqIdMblopfTSWOvjm8LHQo2m42ChkZcxZoMqPar27I1ICsJmvTr14/iKCvRDgfu8vKAtKk6pq2RgYeAFOAw4B3fH E1IrJWRII1MvAn4HfGmInAb3zfAxwPjPA9Lse7YRIikgb8FpgOHAz8VkR6Bym2NjUtkQlUMhAfH8 ZZ57JE   S8zEiVT 5w2Mx9PmdcbppOKVV1nmcDB57lySkpI6HUtLrCkplCcmklZUjMeP2LojV3ExVsDSv3 4Q2mVzWYjv7FRNytSHZOfH9BkQETA5i1r7NixIyBtqo5pKxkYgveP7AnAic0eTd8Hg8GbgAD0ApoWoC4AnjVe3wKpImIDjgM NsaUG2MqgI BuUGKrU3Ovfm4LBaqLJZ2b1DUkoULF1JXV8fGQQNx7tpN3ddt71VQ/eFHuIuKeKK4iIsvvjggcbTGM2IEY6Oj2dSJ4khdWZVvaVR8gP7Ng6Wp1oCnpgZPY2O4w1FdiLu6mpj6evYaD9nZ2QFrN2mUN7FompyowqOtZGC3MWZXSw/wDt8HOKYbgD LyB7gfrx7IgBkAc33uszzHWvpeFg49 6lMjqagQMHEhUVFZA2Dz30UIYNG8aD33 PNT2dihdfbPV8YwzlTz9NcXQ0ub16MWfOnIDE0Zr Rx5BitXKqg5McuwOitZ6KzD2HhWYiVXB8qMqhKWlYY5GdSWOXbu8/ 3bF4vF33p1bbPl5ODweKjetClgbar2a tfdKmIXCsiP/q4IyIxInK0iDwDtPsevYgsEZF1B3gswFvZ8BfGmAHAL4An2tt K/1e7puHsKwkSBOonPn5FLrdDPIV0wgEEWHhwoV8/NlnmGNnU/vZZzRubnnqRu3Sz2hct45/5OVx0cKFAUtKWjP4OO9gTOmXXwW9r0hUtXUrALacwGwCFSzNCw/prQLVHk2f3GMHDwlou2PGjmW300nl p45qhgp2koG5gJu4CURyReRDSKyHdgKnAP8zRjzdHs7NcbMNsaMO8DjbbzJxX98p76Gdx4AwF6g eLWbN xlo4fqN/HjDFTjTFT wZpv3nn3r3srK8PaDIAcOmllxITE8NjBQVYkpIo dvfD3iecbko/ssDVCUl8lZ1FZdccklA42hJ7JDB1FgsRPfQvckb9 ymxOViUADmiQRT02ZFgE4iVO3SsG0bbmPoM65zGxTtb/To0exwOHDu3h3QdlX7tFVnoNEY8w9jzKHAIOAYYLIxZpAx5jJjTDCmj cDR/q Phpv4gHwDnChb1XBDKDKGFMAfAjMEZHevomDc3zHQs5TX4 7vJytVVUBmy/QJCMjg3PPPZd/Pf88CeeeQ 3SpdR89tlPzit7/HEcudu4e88eFpx2WqdLhvpLRKjOtDHEbqe6ujokfUYSU1xMgctFRkZGuENpVZ8 fajwfa3LC1V7VK7fwF6nk FjxgS03SFDhrDH7Sa2vFw3LAojv2/8GGOcxpgCY0xlEOMBuAx4QERWA3fjXTkAsAjYDuQC/wau8sVVjrci4g  x 99x0LOWeDdyDHf6Qz4yADAL3/5S rr63mitJTYkSMpuP0OHHv N12i9suvKHn4EfKGDOHd4mJuvPHGgMfQmsQpUxgUE8PyTz4Jab RILaqmsqYmIDeSw0Gi8VCTJ8 uLUKoWqnxu3b2RXAlQRNrFYr9vQ0rMbohkVhFHG/uYwxXxljphhjJhhjphtjlvuOG2PM1caYYcaY8caYZc2uedIYM9z3eCpcse9bVugKTjIwfvx4zjjjDP70178SdfPN4HKx89xzKXvqaYofeIA9V11F1JAhLPz2v8yfP5/p06cHPIbWDPPVMtjxweKQ9htuxuMhxeHAnhLZNQaa9M/IoFqrEKp2MMZgLS5mpzPwyQBA9ODBgC4vDKeISwa6MmdhIQCFQRoZALj33ntxu9388q9/YeCzzxLdP4Pi  6j7PEnSDl2NvenJFNUU8Of/vSnthsLsD4zZuAAHGtWh7zvcHKVlBIN0K9fuEPxi81mo8zj0QmEym u4hKinE5KoqJJT08PePtpY70VDet0w6Kw8TsZEJFBIjLb93W8iERuAfYwcRUWYYBStzug63CbGzp0KL///e954403eOzDxQx 7VWGf7aUEd98zeKxY3nsxRe5/fbbOeigg4LSf2ssMTEUJyeTXlLao0qL1u30fpqJ6 B21aFms9kotNt1ZED5zbFrJwAm0xaU9odOnECV202pblgUNn4lAyJyGfA68KjvUDbwVpBi6rKcRYXURUfTJyMjINt7tuRXv/oVJ510Etdffz0333wzK3bt4s4HHuDyyy/n2GOP5f/ 7/ C1nebRo9iVFQUuT2o FDR2rUApI6I7JUETWw2G3vr6zUZUH5rWlaYMCw4P Ojx4xhl8NBfe7Wtk9WQeHvyMDVwKFANYAxZivQNcZEQ8hVUEgpBO0WQROLxcKrr77KJZdcwp///GdmzJjBH//4R8455xzeeuutkNQVaIntqKOJFmHdW2 FLYZQq/DtuJYxPrJrDDSx2WwUu5y4KyowDke4w1FdQN3WXOweDxkBXlbYZNSoUd4NiwoKg9K apu/yYDdGLPvt4aIROEtG6yacRYVku wBz0ZAIiNjeXxxx9n69atvPnmm2zatInnn3 ehISEoPfdmpEneatUl3/Tdsnk7qJh124qXC4GjhoV7lD8kpGR8b/CQ2VlYY5GdQWVGzewx lkZIC2Lt5fQkIClQkJJNTX42loCEofqnX JgOfi8htQLyIHIu3GNC7wQura3IVFrGzuiYkyUCT4cOHc/LJJzMqQv4QxfTpQ1FUFHE9qM64p6iQfJeLrKywVcFuF61CqNrLuWtXQDcoOqBM34ZFWnwoLPxNBm4BSoC1wBV41/zfEayguiJ3bS2e2lr22u0BLzjU1dRkZzPU4aSutjbcoYRETGUVFVFRREdHhzsUv/woGdB5A6oNxu0muqycXU5HQHZibUnThkX2bduD1odqmb/JQDzwpDHmDGPM6cCTvmPKx9W0rDBINQa6kpSDp5FqtbJ60aJwhxJ0xhhSGhtpSO46i2u8twm8ld40GVBtcRYUYPV4qElKCuptyIxJkwAoWb0qaH2olvmbDHzCj//4xwNLAh9O1 UsLAKgyOUK2rLCrmLkggUA7P7wozBHEnzu8nJiANO3T7hD8VtMTAyS2gsPmgyotjl2encrtGQF9/faqJwc8p1OKtatD2o/6sD8TQbijDH7xnx9X4d3plqEcRV6SxEXOrvOveNgyZg8mSpjcK1bF 5Qgq5xt7ccdEwXSwD7ZmRQa7Xi1DkDqg0OXx2NpmH8YGnasMitcwbCwt9koE5EJjd9IyJTAJ3y2YzTV3Co0iL06dN1PiUGg4hQ1Ls3/crLu33xoaL13oSnV5DWXweLzWajAqMjA6pNVRs3Uedxkz1uXFD76du3L4UWC3E94PdGJPI3GbgeeE1EvhSRr4BXgGuCF1bX4/IVHOqXmYmIhDucsIvJGc8Aq5VN330X7lCCqnzTJgD6jg19xcfOsNlsFDmcmgyoNlVv3sQuh5ORIVixZO TTqzbrT XYdBmMiAiVuBwYDTwc BKYEzTBkLKy1lQSBmQ6Vse09MNW3AyABteeTW8gQRZ3c6dVLvdDApD efOsNls5NXX4SrWX7qqde68vex0OEKyfDlmyBAAHDt2Br0v9WNtJgPGGDdwjm8L43W hzMEsXUprqJC8h2OHj9foMnwOcdSZwwN33fvkQF3QQH5TicDusi BE28 xM4cJeXY3zLDJXan3E4iK2sZK/HHZJVUn1yJgBQtlb3KAg1f28TfC0iD4vI4SIyuekR1Mi6GGdhEbvrajUZ8LFER7M3NZX xSXd v5fdEUl5VYr8fFda6VtRkYGJW4XeDy4ysrDHY6KUI68PCxAY1oaVqs16P0NnTaVBo HklWrgt6X jF/k4GJwFjg98ADvsf9QYqpy3HX1uGpqWFPQ4MmA81ET5pEttXKpq  DncoQWGMIamhgYaUrlNjoInNZqNUCw pNjQtK4wZPDgk/R00bhy7HA4acreFpD/1P34lA8aYow7wODrYwXUVriJfwSGnU cMNDPi1FMB2PTaa2GOJDg8VVXEGwN9 4Y7lHb7cRVCXV6oDqxhu/ePcu xwdmgaH/Z2dns8XiIKioKSX/qf/zdwriXiPxFRJb5Hg ISK9gB9dVOH3VB4u6UH36UBh2zNFUGQ NP/wQ7lCCom67t2xqbBesOOnduVBHBlTrytauo9LtZkiIduQUEep7p5Lc0IBHd9QMKX9vEzwJ1ABn h7VwFPBCqqrcfmqDxZqMvAjFquV/LQ0bKWluN3ucIcTcAWrVgPQO0I2iWqP5ORkGmJjAU0GVMvqc7eyK0QrCZpYBgzEgndzJBU6/iYDw4wxvzXGbPc9fgcMDWZgXYnTV32wWG8T/ETC9On0t1pZvXhxuEMJuLKNGwCwTZwY3kA6qE9GBnVRVl1eqFpkyS9gu8Me0mQg1Vezo7wHVDCNJP4mAw0icljTNyJyKFqBcB9XYRH1MTEkpqYGdSOPrmjCBRcCsOnll8McSeDV79hJhcvF0BDdTw00bxVC0ZEBdUDuqiriGhooio4hNTU1ZP1mH3wwAAXLloWsz0hU/tzzVL33fsj68zcZuBJ4RER2ishO4GG8WxkrIH7iBL5PTtZbBAdgmzKZQhGifEPq3YkpKmSv201GRka4Q kQm81GiVOrEKoDa9pK2J1pC2m/YyZPpsjppHrT5pD2G2kqnn e2k8/CVl/rSYDIjIQwBiz2hgzAcgBcowxk4wxWhXCJ/W003jKqQWHWlI5fDgj7XbK9u4NdygBFVdVRVVsbJctP22z2djb0KDJgDogh28lQfyIESHtd9CgQex2u6Gb/b5oL1dJCVEhXKnU1sjAW01fiMgbxphqY0x1cEPqmvLz8zUZaEHWSScRa7Hw3RNPhDuUgDFuN6kOJ/YQDp8Gms1mI7 hHldpCcbjCXc4YVVSUsIbb7zBM888w9q1a7t1oSx/Va3fgN3jISMnJ6T9WiwWqpKSSKqu7rH/Du7aOjz19UT16xeyPttKBpp/5NEJgy1wu90UFhbq5MEWTD7/fOo9HiqXLAl3KAHjLCwkCrCEeAg1kDIyMihxucHlxl1REe5wwsJut/PrX/ aAQMGcPrpp7Nw4UJycnKYPXs2O3fuDHd4YVW1fh07HA5GjRkT8r6NzUaix4O7vGdWx3T5thYPZTIQ1cbzpoWvVTNFRUW43W4dGWhBVHwcu9PTyS4oxGG3E Nb0taVlW3cCEDSkK6bI/ 48FAJUenpYY4otMrLy5k3bx7fffcdd558Mguys4lzONnc0MAv3nuXqVOn8umnn5IT4k/GkcK1axfbHQ5OC/LWxQeSOGok7NlDxdq19J01K T9h9u ZKBv5IwMTBCRahGpAXJ8X1eLSI2I6O0Cn72 e1uaDLSs19y59LVa e/j3eNWQeFq75SZPuO65koC8CUD7p5ZeKimpoa5c eSt2YNP8w9njM3biLu62 Iys1l5Pff8/6AgcxPSmLu3Ln7/v/uSTwNDcRVVbMHw8CBA0Pef4ZvRUHeN9 EvO9I0FQVNGJuExhjrMaYFGNMsjEmyvd10/cpoQoy0mky0LYZ111LvcdDUTcpTVy1ZTNuYxgwaVK4Q mwH1Uh7EG1BowxXHzxxVSuXccH48aRVFRExm9/w8hv/8vwT5Yw7MPFJOTkcEtcPEe7XFx00UV4eticCseOHQjgyrBhsfi76CxwRsyYQZXbTdXatSHvOxKE4zZB6P Vu6H8/HwAnTPQioTevcntk87g/Hwc9fXhDqfTHHvyKHS5GBLimdaBlJ6eTtNMgZ60P8FDDz3Ef99 h5fHjCE2NpbBL79E73POwRITA0DMoEEMfOJxEo84nNvS 1DxzTc8 uijYY46tJqWFSaOGhmW/ocNG8Y2pxOza3dY g83V3ExloQErEmJIetTk4EA2Lt3L1arlX4hzOK6otSTTiLZYuGHf/wj3KF0mqW0hGIgMTF0/7MGmsVioXe/fjRae04VwtzcXH5zyy08O2YMiVFRDHz8CeJG/vQPnsTEkPWXvxIzIJu/DxnCH//v/yjvQZPZKtauxW0MmVOnhqV/q9VKWWICKZWVPXJFgbO4OKTLCkGTgYDYu3cvNpstJPt9d2WHX301RW43Va 9Hu5QOi2pto7aLpwINLHZbFRaLD1izoAxhiuuuIJfp/ch0 Eg6/4/E9fKJ19rUiKZ991HutvD2RYrd911VwijDa KNWvY7XQwdsKEsMXgzMwkwePB1QN3MHQVl7Czupp77rknZH1qMhAACxcu5N577w13GBEvITmZbUOHMKSqiooNG8IdTod5Ghvp5fHg7tMn3KF0ms1mo9Tl6hHJwHPPPYfnu 85JTGR9MsuJemII9q8JmHSJHqdcgoL09N579//pqiH/GHy7NjBFrudcWFYSdAk0beksXJ1z6tv5youZktpKetCuD DJgMBMGvWLM4777xwh9EljL3 elzGsOKerps81fq2Lo4bFPpZ1oFms9nIb2zs9slAfX09f7jtNv4wYAAxI0bQ97rr/L623y9/gTU6moXJyTzwwANBjDIyeOrqiK qYo/FSv/ /cMWx4BDDgFg99dfhS2GcDDG4CopYXdtbUjnoWkyoEJq vHH84NAyrJleOz2cIfTIXu /x6A3gcdFOZIOs9ms5FXV4eruLhb35v9 9//zhlOJ2lA5h9 j0RH 31tVN  pJ1zNgtSevH2o49SWVkZtDgjgX3rVgTvMH04S22PP QQil1Oqtf2rN0LPTU1mMZGCuyN2GyhK2qmyYAKKRHBM c4koxh82P/Dnc4HVK6xrvcqWl3ta4sIyODYpcT43TiqaoKdzhBUVJSwqt//jPn9U4j7ZxziO/AltNpP7sEiY7m7Ph4nnnmmcAHGUHqfQW1UsI4XwBgwIAB7HR7kD17whpHqDUtKyx2uTQZUN3bSXfczmaHnfKnnuqSNfEbtm2jwuViRAf qESazMzMH1Uh7I7uuecerkpMwpKYSN/rru1QG9H9 5F60omcktqbpx95pFvXHSj67ntq3G5GHXZoWOMQEap79ya1thbjdoc1llDSZED1GP369SN37DjS6usp uCDcIfTblJYyF5j6NWrV7hD6bSsrKz/FR7qhslAcXExK556msMSE l31VVYO7GxVO9zzyUWGF9aykcffRSwGCNN7fr1bLHbmTxlSrhDwTpkCLGAfdeucIcSMk5fMlDicumcAdX9zfndnRQ4nWy9v tNyEqqqaE6KSncYQREdnY2pb5koOmXUHfy17/8hWt79YJ /eh9fucm caPHUvcxIlckN6Hf3WDWhkHYowhtqCAbW43o0aNCnc4pE2ZDMDuzz4PcySh01Tzo0RHBlRPkDN5Mt kpdGnoICaVavDHY7fPA0NpHk8uPuFtiBIsPTt25cK3ySx7jYyUF5ezqYnn SguDgyf3UTlgBskJV23nlkW62ULl3aLZcZOvfmE Ny0di/P1FRbe1jF3zDjz4apzEU/fe/4Q4lZFzFxTiiorAkJJCcnByyfjUZUGEz5dZbqHG7WfW734U7FL9Vbt4MQOzgIWGOJDAsFgtpmZnYu2HhoYcefJCLEhMxWVmkzJsXkDaT5xyLSUjgxKRkXnzxxYC0GUka1nknx8aPjYyVMmMnTGCL3Y5ry ZwhxIyruJiqqzWkJe3D0syICJniMh6EfGIyNT9nrtVRHJFZLOIHNfs FzfsVwRuaXZ8SEi8p3v CsiEhPK16I6bu4pp/BFXBy9N2ygoYvcE8z77jsA0seHrxhLoGVnZ1Mp0q2Sgfr6elY9 iijYuPIuuF6JEDVQS2xsfQ 6USO69WLV59 OiBtRpL8z7/AaQwDDj883KEAEB8fT35cHMklpd166WtzruJiyownpLcIIHwjA uAU4Evmh8UkYOAs4GxwFzgHyJiFREr8AhwPHAQcI7vXID7gL8aY4YDFcAloXkJqrNEhNE33YjLGH64 Za2L4gAZb6KYAO6wbLCJk2TCLtTMvDcs89yfmwc7n79SDn  IC2nXrKKcQCmbt2sWrVqoC2HW5Vy5ezqbGRGRGSDAA4BmST6PHg8m0I1925iospaLT3jGTAGLPRGHOgcZ8FwMvGGLsxZgeQCxzse QaY7YbYxzAy8AC8VbEOBpoKnb/DHBy0F ACph5F1zAJ1FRpK1aRcOOHeEOp02NO3ZQ4nIxfPz4cIcSMNnZ2eytr9u3pKmr83g8fPm3v3NQXBxZ11 HBPjed1xODtaBAzk1NZWnnnoqoG2Hk3G7idu7l62YiJg82CQpJweAUl xr 5sX/XButBWH4TImzOQBTSvMJHnO9bS8XSg0hjj2u 46iIsFguDf/kLHB4PK265NdzhtMlaVEyBdO3dCveXnZ1Nod2Bs7ikWwzFfvzxx8ytr8feqxepJ50U8PZFhLTTT2dyXDyfv/oqLper7Yu6AMeOHcS63TgHD8ZiiZw/DcOOPhqXMT1iRYG7shLjdLK3vqH7jAyIyBIRWXeAx4Jg9elHTJeLyDIRWVbSjYZEu7qTL76YjywWUlevpmHr1nCH06rk2lpqk0I3wzcU9i0vbGzEU1cX7nA67Y0//YkpCQlkXn5Zu8oOt0ev d4JiQc7HHz eff4I1Xim7Hfd bMMEfyY1MPOYRtDjv1Idy0J1yalhWWukO7rBCCmAwYY2YbY8Yd4PF2K5ftBQY0 z7bd6yl42VAqohE7Xe8pZgeM8ZMNcZM7RvivaJVy6xWK0NvupEGj5vVETw64K6qItUYXBkZ4Q4loLKzsylx woPFXftJHnTpk2M2LgRZ1QU6WedFbR orOyiM3JYX6vVF566aWg9RNKeZ98Sq3bTU6A51h0Vu/evcmLiSGpsLBbjFy1JlzVByHybhO8A5wtIrEiMgQYAXwP/ACM8K0ciME7yfAd4/3JWAqc7rv IqC1ZENFqNMuvpj3LFZ6rV9P/ZrI3LK0eMUKABJGjQxzJIGVnZ39vyqEXXzewLN//SvHJaeQfOopWINcGCr1hPmMjIlh2Vtv4XA4gtpXKHjWrWONw8HUCJwc6xg4iCS3G fe7j2JsHky0CPmDIjIKSKSB8wE3heRDwGMMeuBV4ENwGLgamOM2zcn4BrgQ2Aj8KrvXICbgV KSC7eOQRPhPbVqECwWq2M vWvqHC5WH/bbeEO54D2 IZRM6ZObePMriUjI4Myt7fWfldeUdDY2AiLPsAqQtblVwS9v Tj5mJEOAzp8uWJXeXl9K6tpSwjg/j4 HCH8xOp06cDkL90aZgjCS5Xyf9KEfeIkQFjzJvGmGxjTKwxpr8x5rhmz/3RGDPMGDPKGPNBs OLjDEjfc/9sdnx7caYg40xw40xZxhjuua uIozL7qIN60WknK3URuBFceq1q2nweNhxKHh3cAl0KKiopD0NKBrJwNvvvQSJ8XFYZ80kZjs4M8jju7fj/gpU5ifmspLXbwAUZ4vmUk/InKWFDY3 rg51Hs85C/9NNyhBJWruBh7dDQSGxvyvU8i7TaB6sGsVisTbrmFAqeTLb/5bcTdHzS7d7PT5WLgoEHhDiXgemdl4ejihYfW/PNf9LJaGfmLX4Ssz9QT5jM4KooNH3xAfX19yPoNtB3vvkeDx8O0c84JdygHNGnqVNY2NuLxba/cXTmbVR8UX5nwUNFkQEWUsy 4gNetFuL37KFmyZJwh/MjSRUVlMXHR9Syq0DJHjCACt8a565oy5YtTCgqoqZXLxKnTQtZv8lz5mAsFmZFRfH    HrN9AM vXs9HjZvykSeEO5YDi4 MpSEkmtaISTxdOutriKi4JS/VB0GRARZioqCim33or2 12tv/hrojZx9xdW0tvtxtnN1tJ0MRba8DeZScQvvHAA0yIj6ff eeH9BNVVFoaiTNmcELv3l32VkH1nj30b2jAMXJkyD NtodlzEFYgboInWAcCK7iYgrtoa8 CJoMqAh03gUX8GqUldjiYqreeTfc4QBQ6VvjHD9iRJgjCY7s7GwKHXYcXXAnPqfTifnwI5wiDLjg/JD332v PDItVnZ 8glVVVUh77 zvvvXv7CIMOLss8MdSquGzPVOLduxeHGYIwkO4/HgKi1lV23oqw CJgMqAkVFRTHr1ltZ39jI7j/9CU8ELNva/fU3APT17a/e3WRlZVHodOHqgmu5l7z3HrOjo2mYNBFramrI 0 ePRtjtTI7Lo4333wz5P13VvmSTyjzeJgRofMFmhxy3HFss9up9P2/2N24y8vB5SKvrk5HBpRqcv4FF/CiQHRFBZWvvBrucKhYsxq7x8PwbraSoEl2djZ7nU6w23FXVIQ7nHZZ869/kWy1Mvq668LSv7VXL5IOP5z5vdN46YUXwhJDRxXl5zOovJzSAQOIClK1xkDJzs5mY1QUKXl5mAj4gBBo4Sw4BJoMqAgVHR3NzJ//nO/q6yh8 KGwl8n17NjJDqeT4SO7V8GhJtnZ2RS4nABdqrBLTU0NWZu3UBEfT4pvLXo49Jp3PP1EKPvmGwoLC8MWR3u9  c/08tqZdS5kT0q0MQ5ZgyxxkRsYbLOcDZLBrKyQr/FjiYDKmJdetll/LOmBqmqpuLll8MaS0pZGQVxcUQFeAe8SJGZmUm 05cMdKGtYhc98wxT4uKIPe64sE5 Szr6aIiO5rikJF59NfwjWf6w2 1UvPMudhFGRfgtgibZc4/DYwy73n0v3KEEXNPIQEkYqg CJgMqgvXu3Zup557L9w0NlD79DMb3xyrUXCUlJLvdNGb0D0v/oRAXF4fdV SkKyUDu59/HoBxV18V1jisSUkkzzqyS90qeOrf/ YQEdyTJmFJSAh3OH45bO5cNtgbqfzyi3CHEnD7NinSkQGlfuq6667jqbJSPCUlVIep5GvFypUAJIwdG5b Q6VXZiaNIl0mGSgoKOCg4mKK09OJHTCg7QuCLGXePHoD7rXr2LZtW7jDaVVlZSVf/OlPpEVFMfKyy8Idjt GDh3K5phYkvMLul29AVdxMQ0xMUQnJJCSkhLy/jUZUBFt1KhRJBx2GHvcbsqefCosM933fOH9FJLVTScPNskeMIBiDM69LW78GVEWP/IPhsbE0u/MM8IdCgBJRx4JcbHMS0mO J0Mb/71rznFYsVjyyDpyCPCHU67RE bShRQ8Xn3Gh1wFRdTZbGQlZUVlltemgyoiHf5lVfyZGkJ9vXrafDtHBhKNWvXstfpZHyE7fMeaIMGDWJ3Q2OXGRmoff99nMDIhQvDHQoAloQEUo4 mnm903jp Rcidonmc889x8YXX2RcXByZV1yBdLGKmhPOPZdqt5vtEZ5wtZeruJgSjycs8wVAkwHVBcyfP5/v4uKos1gof/qZkPcftSePXLeLQd1wT4LmvMlAA44uMDJQsHcvE2pqKMzOxhriDV1akzJvHsnGkLpnN6tXrw53OD/xyiuvcNnPfsZvBg8hesAAUk89NdwhtduRRx/NV40NWFeujJgKpYHgLCkmv7FBkwGlWhIdHc05F1/MS6Wl1HzyCc4QLt3y1NeTWl9PdVpaRJdqDYRBgwaR73Jiampw19aGO5xWffbPf9E3KorMM04Pdyg/knj44UhiIvN6pfJCBE0krKur44YbbuDss8/mtzk5ZLlc9PvVTUhMTLhDa7f4 HjKhg0j3umkPgwjhcFgXC7cpWXsqq4Jy RB0GRAdRGXXnopr1RWYDweKl97PWT9NmzahAWwDh8esj7DZdCgQf9bXhjhtQaqFi/Gbgxjzjsv3KH8iCU2lpTZx3B8aiovPfcczjCtgGlSV1fHAw88wNChQ/n73//Ofeecw8l2Bynzjif52GPDGltnHHTeeTg8HrY882y4QwkIV2kpGEN Y6OODCjVmqFDhzLmyCP5we2m4rXXQrbMMP/zzwFInz4jJP2F04 SgfzIvVVQWlLC6MoKijIzsSYlhTucn0g /njiPR6G19by7rvh2VujtraWP//5zwwdOpSbbrqJCTk5/PDnP3PSho3EDh9Oxp13dumRrpPOOosvGxrwfPFF2JYcB1Lz6oOaDCjVhssuu4xnCgpwFxdT89lnIemz/NvvKHQ6GXNY915JANCvXz9KfZPJInlk4NNHH6V/VDT9Tzk53KEcUNIhh2BJSeG0/hn8 9//DmnfbrebRx55hMGDB/PrX/ anPHj fYf/ CxtHQSH3 CuHFjGfjE41jDsHQtkFJSUigYM4YEh4OqpUs73M7u3bu56Wc/4 bBg3liwEBeGnMQH/z8Ktx2ewCjbdv/kgGn3iZQqi0LFixgTXQU1dHRVL78Skj6tGzbxtrGRsaNGxeS/sJJREjKysIZ4bUGyt59D6cxjIuQVQT7k5gYkuccy5Hx8Xzx0Ufs2rUrJP1u3ryZGTNmcM011zAxJ4dlf/sbjyankPL3B3EWF5Pxu98x6OmniUpPD0k8wTb951dS6nKx Z//6tD1Sz/5hLumT e8r77morh4RmVlkSbC4KVL eKQQ3GWlwc44pY5dWRAKf/FxsZyxtln80JJCXVff40jyL9kXeXlJNXWkp USGJiYlD7ihQDBw2iLIKTgerqakaUllLQrx9REfzpNvXkk4l2uTguOZknn3wy6P298cYbTJs2jV07d/LBrbfxr gYEv75L4zdju3uuxn 4WJ6n3Um0o3Kac8/6SQ BZI2bsSxe3e7rv3PK6 w5qKF/KJXKgkzZzL8ow857Nv/Mmf9Or449BB619TwzYknYVyu4AS/H1dxMUaEcrdbkwGl/HHhhRfycmkpRoSKIO9m2Lh2LQCW0aOD2k8kGTRoEHl2e8QmA189 yxZ0dGkHn98uENpVfyUKcQMHswlgwbz2GOPYQ/isPMzzzzD6aefzukjR/LltIMZ9J//YImLI vBvzP0vXdJPfWULrlqoC1Wq5X0iy7E5fGw8d57/b7u8YceouKmXzEnKYmka65m5NNPETNwIOAdHbv88cdZMno0/crKWHnb7cEK/0dcxSXUxUSTmpZGXFxcSPrcnyYDqkuZMWMGvYYMZlVsLFVvvokniFuZFn/9NW5jyDz88KD1EWkGDRrEjro6HHl54Q7lgArefhu3MUy47NJwh9IqESH19NMYarcTX17Oiy  GJR Xn/9da679FKenDSZX9XWYXU6ybz/foa8 R9S5szpcgWF2uvC669nscOOfLoU 549rZ5rjOHe224n f4HmJ6YSPrv7mTANdf8ZCKliHDliy w1OXC s47NIbgNo ruJgKsYRtVAA0GVBdjIhw4YUX8o tW3FXVFC7ZEnQ ir/9ju22u1MPuSQoPURaQYNGkSe04GnvDzs20bvz PxkLpjJ/mJicT17RvucNrUa8ECsFq5Yvhw7r///oBXJFy0aBG3Xnghb44cyQy7nfQrrmDoe /S64T53T4JaJKSkkLvK67A6fGw trrWnyP3W43t116KeNfeokxCQlkP/Qg/c46q8V2k5KS6H/zr3F5PKz45S DFf4 ruLisG1d3KRn/MSobuX888/nm/o66pISg3arwLhcxGzfzmp7IxMnTgxKH5Fo0KBB7PEt1Yq00YHln33GaKsV69Qp4Q7FL1F9 5I0axbHx8WxdeNGPvjgg4C1vXTpUq4/6yyeGziIrJReDHrmafr94gYssbEB66OruPimm3g7Lo7kTZvY9o9//OT5uro6rp43j7mff05WUhLDnn2GVD9qLJxyySUsjoqi1/oNNG7fHozQ93EVF7O3vl5HBpRqjyFDhnD4EUfwZm0t9d99h33HjoD30bhxE9EuF0V9 pDQRbZ3DYRBgwaxx GrNdDGsGuorX7uOSwiHBRhhYZa0/vss4muq f8QYO58847AzI68N///pcLFizg0axsUnv1YvALL5AwdWoAou2aoqOjOe Vl/nGbqfhwYdY9cADGGMwxvDpRx/x2ylTuHLHTlJSezPmP2 QOG2aX 1aLBaGXXctDo HtX 4K2jxexwO3JWV7Kyu1mRAqfa68MILeSw3F2O1Uvl64CsS1v3wPQCxkycHvO1IlpWVxV63dwa1Y3dkJQPuH5ZRK0LGYYeFOxS/JR52KLEjhnNVdjY//PAD//nPfzrV3qpVqzj5 ON5yGbDFhPDwH/ k9ihQwIUbdc1aswYJj77DKvdLmL//TiLxxzEEyNHYb3qKi5BMOPGMX7xB8SOGNGudk 5 GI dDiI /Zb3JWVQYndVVwCQLEzfDUGQJMB1UWdfvrp1EZHsz0tjar/BH4iYelnn7PT4eCgbr5t8f6io6NJysig0WrFsad9y7WCae/evYxpbKRy0CDEag13OH4TEdIWLiSxpISzxozhtttuw9HBn9WNGzcyd84c7u7Tl9EWK1kP3E/C5EkBjrjrmnL44cz/9ltWzZyJOz6ekQkJWMYcRL8H/87EN14nqk fdrcZGxtL1Px5RBnD7iDtNeEqLgLCW2MANBlQXVSvXr04 eSTeTgIEwmN241zzWp qK9nZjfftvhABg0aRJFFcO6JnDkDnz/zjHdjovnzwh1Ku6WccALW9HR PXwEW7Zs4U9/ lO729i fTuzZ8/m sREDomKov/tt5PShfcWCJbeffpwzlNPcsLKFRyxehWHv/sO6XPmdKr08onXXcf6xkaKXwzOlsnOggIACl1OBgwYEJQ /KHJgOqyLrroIj4qKsTRu3dAJxLat2whqtHOBhHGjh0bsHa7ikGDBrGr0R5RIwOFvsl3I89seQZ4pLLExpJ ySXEb9nCLfPn84c//IH169f7ff2OHTs46qijOFUsnBwbR9olPyPt/K4zb6KrGzVqFMtSUkgpK6Nxy5aAt /y7cJa4HIx0FfvIBw0GVBd1rHHHostM5NPLBLQiYR1X38NQNTEiVh6yBKt5gYPHsyW6mqce/MjYr94h8NB37w8SpMSie7fL9zhdEjv884lKiODn0XH0KtXL0499VQq/bgHvX37do466ihmOpxclZREyrx59LvxxuAHrH5kwJln4DGGnS /HPC2nQWFOKKicMfEkB7GUtE97zed6jasVisXXngh9y1bBlZrwLY2Lv/kUzY1NjLh6KMC0l5XM2zYMHbZG8HlwllQGO5w Pazz5gQE4tM6RpLCg/EEhtL32uvwbVxI / 8ka2b9/OKaecQm1tbYvXfPXVV0yfPp3RjY38Ji2NhGnTsN17T4 pIRBJFixcyMqGBio/WBzwtp2FBVRYrQwcODCsO0nqT5Xq0i666CKKHQ4KBg4ISEVCd20djjVr KqujsN7UOXB5oYOHdpseWH4bxWsf/llokW65C2C5nqdcgrxkyeT uabvPDII3z55ZccccQRP7llUFVVxa233sqsWbOYnpbGXzNsxI0YQfY/HsHSDcsKdwXZ2dlsSk2lV0VFwPdEcRUUUuR2h3W AGgyoLq40aNHM2PGDJ7YvRt3RQU1H3/cqfbqv/0vFreb711OJvewZYVNhg0bRp7Tm1Q5IqDWgHvZMhqBvod3nSWFByIWC7a77sI0NDD9q69564032L17N PHj2f27Nlcc801nHbaaQwYMIB7772XW085hfuTU4ju04cBjz2KNTk53C hR0s9fi4ABW 9FdB2nYWF7KmvD t8AdBkQHUDF198MS9t2ICnb18qX32tU23VfvEl9UDS1KnE9NBPYVlZWZRbLLhFwl54qKysjJG1dZRlZXaLzXZihw7BdtcfqP/ eyZ9/TXr16zhjjvuoLi4mBdeeIFVq1ZxxhlnsPKRf3Be7jai09MZ9OwzRPfrmnMlupNjzj6bzfZGCt5fFLA2PQ4H7rIytldX68iAUp115plnEhsXx3cpKZ2aSGhcLqo  ojPq6uZPXdugKPsOiwWC4OGDKEiOjrshYe eu01BsTE0Pvoo8MaRyD1Oukk t5wA9XvvEvjrbdx 0UXsWbNGioqKtj46af8Pj2d2AcfJHbkCAa98DzRYVx7rv5n4sSJrBILibt3464NzL4dTSsJCp1OHRlQqrNSU1M55ZRTuPu7byEqqsMTCeuXLcNUVrK4pobjjjsuwFF2LcOGDSPP5Qr7yEDeu 8BcNC554Y1jkDrc UV2O6 m8b169k /wRyj51D7uxj2XbcXGoWf0j6lVcw LnnOlQoRwWHiBB18DSigOqvvgpIm878/9UY0GRAqQBYuHAh28rKqBo5ssMTCasXL8YhwrZeKT2yvkBzw4YNY2tNNY49ewK 256/jDEkbtlCaUw08UO6X8nd1FNPYdjiD j3q18RP34c8Tk59PvVrxj28Uf0u GGbnFbpLsZe9pp1Lrd7Hyzc2WlmzgLvclAgdMV9tsEUWHtXakAOeaYYxgyZAhP5 /l soqaj76mF4nzPf7euN0UvPRR3zRUM s444L6xKfSDBs2DBW1tXhqanBXVlJVO/eIY9h68aNjBehfMyYkPcdKlF9 pB yc/CHYby01HHHsvzv/glM35YhjGm078n9t0mCHP1QdCRAdVNWK1WrrrqKh777jtMZiZljz O8Xj8vr7mk09xl1fwemkpJ510UhAj7RqGDh3K7qblhQFeSuWvZc8 S4LFwuCTTwlL/0rtLzU1lT19 5JYX48jANsaOwsKaYiKIqF3bxITEwMQYcdpMqC6jZ/97GfExcezODkJ 6ZN1Hzyid/XVr76KtWxsawS6fHzBcA7MrDdYQfAvmNnWGKo eILnMYwbIEmZypy9DpqFgClnVzGDL6CQxZL2OcLgCYDqhtJS0vjvPPO47cff4x1wABK//FPv0YH7Nt3UPfNN7xWUc78E08kLi4uBNFGtiFDhrDX6cQjgiNAZZ7bw l0kllUTEHv3lgSEkLev1ItOXTBAnY6HOR/ FGn23LlF1AQASsJQJMB1c1ce 211DY08O3AAdg3bqTqnXfavKbs0UfxREfzVH4 Z5xxRgiijHzx8fH0y8ykIjY2LMnADx9 yPDoaGJnzgh530q1ZsaMGSxz2InZsgXjcnW4HWMMjr15bK2uZujQoQGMsGPCkgyIyBkisl5EPCIytdnxY0VkuYis9f336GbPTfEdzxWRB8U3c0NE0kTkYxHZ6vtv6Gc6qYiRk5PD/Pnz eVbbxEzfhzFf74fd3V1i fbt  g6r33 CY5GWtaGvPn z/psLsbNmwYezweHDtDnwxseeUVAMaep7vzqcgSGxtL3dChxLjdNK5b1 F23BUVmPoGdjTU99xkAFgHnAp8sd/xUuBEY8x44CLguWbP/RO4DBjhezRVhbkF MQYMwL4xPe96sF  9vfUlZezoeDBuGuqKDwD3cdcHmcMYbCP/we4uK4Y/kyFi5c2GOrDh7IsGHD2FxTg2PnrpDvXmhZtZoKEfp04c2JVPeVNWcOAAUfdXzegDMvD4A8h7PnJgPGmI3GmM0HOL7SGJPv 3Y9EC8isSJiA1KMMd8a72/1Z4GTfectAJ7xff1Ms Oqh5o2bRrz5s3j/556isRLfkb1u 9S8dzzPzmv/Kmnqf/vt3w/YjglTieXXXZZGKKNXKNGjWJtRTnG6cS5d2/I q0oLWVkYyPlgwf1 CWeKjIdNn8 mxobKfn00w630bTvR56zBycDfjoNWGGMsQNZQF6z5/J8xwD6G2MKfF8XAv1balBELheRZSKyrKSkJBgxqwhxzz33UFVVxT1bt5J0zDEU3X03JQ89jMfhwLhclD3xBMV//jOxR83i2kWLOP300xkxYkS4w44oo0aNYqeveFMo5w18 /zz9LJa6aerOlSEmjBhAqs8HuJ37cJjt3eoDWeeN8He63QwePDgAEbXMUErOiQiS4CMAzx1uzHm7TauHQvcB8xpT5/GGCMiLZZLM8Y8BjwGMHXq1PCUVVMhkZOTwzXXXMODDz7IBZ99xqDEBEofeYSyp55CrFY8NTUkHzubv4lQW1vLHXfcEe6QI87o0aPZ7ksG7Dt2kHTkkSHpt2jxYgYYQ875F4SkP6Xay2Kx4Bw9mujdu6lfuZKkGe2f6OrM20NdVBSpNhvx8fFBiLJ9gpYMGGNmd Q6EckG3gQuNMZs8x3eC2Q3Oy3bdwygSERsxpgC3 2E4o7GrLqX3/3ud7zzzjucff75rFixgoGnnkbNkiUYl5OkI45kfXwcfz3sMK644gpycnLCHW7EGTZsGDWAPSYGRwhrDaTu2EFefALj qSHrE l2mvwvONx/fNf7P3gA0Z1IBlw5OVRLETELQKIsNsEIpIKvA/cYoz5uum47zZAtYjM8K0iuBBoGl14B 9kQ3z/bXXUQfUcvXr14rXXXqO4uJjjjjuO mFDybjjdmx33knhgGxOO/10Bg0axN133x3uUCNSTEwMQ4cOpSjKGrLbBNtXr2YEgitnfEj6U6qjZs2bx7rGRiq  LJD1zvz9rKzobFnJwMicoqI5AEzgfdF5EPfU9cAw4HfiMgq36NpI  rgMeBXGAb8IHv L3AsSKyFZjt 14pAKZMmcKbb77J vXrGTduHLfffjs333wz06ZNw Fw8Pbbb9M7DHX3u4pRo0axvbExZMnA6meexSLCsNNPD0l/SnXUiBEj2GC1kFRQ0O4tjY3LhTM/n61VVT07GTDGvGmMyTbGxBpj htjjvMdv8sYk2iMmdjsUex7bpkxZpwxZpgx5hrfqgKMMWXGmGOMMSOMMbONMeXheE0qch1//PF8//33jB8/nnvuuYe//OUvHHHEEXz33XeMH6 fQFszevRoVpWW4SopwV1VFfT Gr75hmpjGDVvXtD7UqozRATJycEK1P3wQ7uudRYUgNvNHqeDYcOGBSfAdoqo2wRKBUtOTg5Lliyhvr6empoa3nnnnYjJyCPZqFGj2FTv/dRjz80Nal8ul4sBZWXs7dMHS5RuqKoi34gTT8Th8bD7/ffadV3TSNsOh4MxEbIrpyYDqkeJi4vTvQfaYfTo0WzxLZ2yb9kS1L7WvP02faxWEg49NKj9KBUoRx13HKsaG6n99rt2XWf37Xi4w Fg1KhRwQit3TQZUEq1aPTo0RS6XDhjYmgMcjKw/Y3/ADBx4cKg9qNUoGRnZ5MbH0dyaSnuykq/r3Ps2ElDlJWEjAySkpKCF2A7aDKglGpRnz59SEtLozguDvuWrUHtK3rdOnaKkHFQZAybKuWPmClTsADV337r9zWO7dvJM95kO1JoMqCUatXYsWPZardj37LlgHs8BEJtSQlD7Haqhw4JSvtKBctBJ51EvcfDrnf9nzdg37GdTTXVmgwopbqOnJwcvi8qwlNTg6uwMCh9rHj6aaJF6H/88UFpX6lgmXXssaxoaMC bJlf57urq3GXlrG5rk6TAaVU1zF /HjWVnuXFQZrEmHZx0uo9XiYdoGWIFZdS3p6OntSe5FSVYWrtLTN8x2 yYM7I2glAWgyoJRqQ05ODlt9KwqCMYnQGEPf3bvJTUggsVevgLevVLAlTJ8OQMWXX7V5buNm7/9Dm 12JkyYENS42kOTAaVUq8aNG0e1x0N9QkJQJhHmf/st6YCZNCngbSsVChNOPplqt5td777b5rmNmzbSaLUg/fvTp0 fEETnH00GlFKtSk5OZsiQIeyJstK4cUPA29/w/AsAjDr77IC3rVQoHD5rFssbG3CvXt3mufaNm9jm9jBx4sTgB9YOmgwopdqUk5PDypoaHNu246lrXx32tri /55tLheTju3QRqdKhV1SUhJF/fqRUleHMz /xfOM203jpk2srKhgUoSNhGkyoJRqU05ODp/n7QVjaNwQuNEBZ3U1WdXVFA/IxmLRX0eq6 p12GEAlCxd2uI5jl27MY2NbGhsjKj5AqDJgFLKD PHj2ddQz0ADevWB6zdNc8/T7QIfefMCVibSoXDlJNPptzlYvf7i1o8p2HVKgDWNTYwY8aMEEXmH00GlFJtmjJlCmVuN43JyTSuWxewdvPfe59at5uZP/tZwNpUKhxmHHIIyx12LOvXt1icq37FcuqtVtyZmWRlZYU4wtZpMqCUatOQIUO866ljogOWDBiPh7QdO9gcH0ef/v0D0qZS4RIbG0vjyJEk2 3Ub9p0wHMaVqxkVWMDh0TgZlyaDCil2iQiTJs2jWWVlTh27cJdXd3pNvO/ JI0YzBTpwUgQqXCb9QFF A2hvWP/fsnz7lKSnBs385/Kys5zDe/IJJoMqCU8svBBx/Mp7t3A9Cwdm2n29v07LO4jGHsRRd2ui2lIsGcM89khd2O87PPfnKroPbzzwH4qq6O4447LhzhtUqTAaWUX6ZNm8bq nqMCA0rVna6vagVy1nrdjHx8MMDEJ1S4RcfH0/pyJGkNTRQsd9eBTWfLqXcasU6dChDhkTehlyaDCil/DJt2jRqPR6q09Op93NTlpbUbt1Kv0Y7FSNG6pJC1a3M NVN1HncrLn77n3HXGVl1H7xBe XlXLSggVhjK5l n hUsov/fv3Z9CgQWzAu0TKOBwdbmvVv733VEecd26AolMqMsycPZvPo2NI37iJhtxcACpefhlcLl6tqOCiiy4Kc4QHpsmAUspvhx56KB/s2oWx2ztVb6Dhs8/Z6nRy5BlnBDA6pcJPRBh2043Uud2sufQyaj77jLLHn BLp5PsmTMjatvi5jQZUEr57cgjj STAm 51frlHbtV0JifT2ZVFXsHDCA2NjaQ4SkVERYsXMhL6WkkFBSQd XPqXK5 M3uXdx1113hDq1Fmgwopfw2a9YsKtxu6tPSqP/u w61seYf/8QiQuZZZwY4OqUig4hwy5tvcq3Hza/y8zluw3p fscdEbmksElUuANQSnUdI0aMICMjg7VRUcz44Qc8DQ1Y4uPb1UbDRx xx27nuAsuCFKUSoWfzWbjvZUref/997ltwAAOOeSQcIfUKh0ZUEr5TUQ48sgjeWPnDozdTt1337Xr vpt2 hXXc2ugQNISkoKUpRKRYakpCTOOuusiE8EQJMBpVQ7HXPMMXyYl4eJjaXuiy/ade2qhx4CYOTFFwcjNKVUB2kyoJRql/nz5 M0hsK fan97PMWN2XZn3G7kU Xssxh59hzzglylEqp9tBkQCnVLpmZmUyZMoXFFeU48/P93riodMknpDoclEyapKsIlIowmgwopdrtpJNO4tHVqyE6mqp33/Xrmk0PPUiZy8XMG24IbnBKqXbTZEAp1W4LFiyg2u2mZOBAqhd9gHG5Wj3fvmcPvXO38U1cHDMieHmVUj2VJgNKqXbLyclh3LhxvFxQgLu0lNrPPmv1/PV3/RGPMdguvQQRCU2QSim/aTKglGo3EWHhwoU8tnIF9OtL2VNPt3ius6iIqC  4IPGRk67/PLQBamU8psmA0qpDjnvvPOQqCi 6p1Gw/Ll1H1/4IqEm267HY/bTey555CQkBDiKJVS/tBkQCnVIRkZGZx//vncvORjLP37U3TXH38yd6Dmyy J vprXmlo4Ge33hqmSJVSbdFkQCnVYTfffDOVDQ0ssmVg37KFonvu3Vd3wJ6by87rb2CLvZHsX9xASkpKmKNVSrVE9yZQSnXY6NGjufLKK7np0Uc56pprqXjhBRw7dhAzfBgVr71OZW0t/ 7dm7evvz7coSqlWiH Vg/rbqZOnWqWLevYFqxKqf prKxk0qRJNDQ08PF11xP7wSLcFZV853Twp9Iy3v3 O4YNGxbuMJXq8URkuTFm6oGe09sESqlOSU1NZdGiRVgsFib 3x2cUl5OzuZN3FhayhPvvauJgFJdgN4mUEp12pgxY1i9ejV/ 9vfWLt2LSeddBLXX389GRkZ4Q5NKeUHvU2glFJK9QARd5tARM4QkfUi4hGRnwQmIgNFpFZEbmp2bK6IbBaRXBG5pdnxISLyne/4KyISE6rXoZRSSnUH4ZozsA44FWhpM/S/AB80fSMiVuAR4HjgIOAcETnI9/R9wF NMcOBCuCSYAWtlFJKdUdhSQaMMRuNMZsP9JyInAzsANY3O3wwkGuM2W6McQAvAwvEW T8aOB133nPACcHK26llFKqO4qo1QQikgTcDPxuv6eygD3Nvs/zHUsHKo0xrv2Ot9T 5SKyTESWlZSUBC5wpZRSqgsLWjIgIktEZN0BHgtauexOvEP tcGIyRjzmDFmqjFmat  fYPRhVJKKdXlBG1poTFmdgcumw6cLiJ/AlIBj4g0AsuBAc3Oywb2AmVAqohE UYHmo4rpZRSyk8RVWfAGHN409cicidQa4x5WESigBEiMgTvH/uzgXONMUZElgKn451HcBHwdugjV0oppbqucC0tPEVE8oCZwPsi8mFr5/s 9V8DfAhsBF41xjRNMLwZ KWI5OKdQ/BE8CJXSimluh8tOqSUUkr1ABFXdEgppZRSkUOTAaWUUqqH02RAKaWU6uE0GVBKKaV6uB47gVBESoBdAWyyD1AawPZ6In0PA0Pfx87T97Dz9D3svEC/h4OMMQesuNdjk4FAE5FlLc3SVP7R9zAw9H3sPH0PO0/fw84L5XuotwmUUkqpHk6TAaWUUqqH02QgcB4LdwDdgL6HgaHvY fpe9h5 h52XsjeQ50zoJRSSvVwOjKglFJK9XCaDLSTiMwVkc0ikisitxzg VgRecX3/HciMjgMYUY0P97DhSJSIiKrfI9LwxFnJBORJ0WkWETWtfC8iMiDvvd4jYhMDnWMkc6P93CWiFQ1 zn8TahjjHQiMkBElorIBhFZLyLXH Ac/VlshZ/vYdB/FiNqC NIJyJW4BHgWCAP EFE3jHGbGh22iVAhTFmuIicDdwHnBX6aCOTn 8hwCvGmGtCHmDX8TTwMPBsC88fD4zwPaYD//T9V/3P07T HgJ8aYw5ITThdEku4EZjzAoRSQaWi8jH /3/rD LrfPnPYQg/yzqyED7HAzkGmO2G2McwMvAgv3OWQA84/v6deAYEZEQxhjp/HkPVRuMMV8A5a2csgB41nh9C6SKiC000XUNfryHqg3GmAJjzArf1zV4t5jP2u80/VlshZ/vYdBpMtA WcCeZt/n8dN/tH3nGGNcQBWQHpLougZ/3kOA03xDiq LyIDQhNat Ps q9bNFJHVIvKBiIwNdzCRzHdLdBLw3X5P6c in1p5DyHIP4uaDKhI9C4w2BiTA3zM/0ZalAqlFXjLt04AHgLeCm84kUtEkoA3gBuMMdXhjqcrauM9DPrPoiYD7bMXaP4pNdt37IDniEgU0AsoC0l0XUOb76ExpswYY/d9 zgwJUSxdSf /KyqVhhjqo0xtb6vFwHRItInzGFFHBGJxvtH7AVjzH8OcIr LLahrfcwFD Lmgy0zw/ACBEZIiIxwNnAO/ud8w5wke/r04FPjRZzaK7N93C/ 4kn4b2HptrnHeBC30zuGUCVMaYg3EF1JSKS0TTfR0QOxvv7UhP7ZnzvzxPARmPMX1o4TX8WW HPexiKn0VdTdAOxhiXiFwDfAhYgSeNMetF5PfAMmPMO3j/UZ8TkVy8k5PODl/EkcfP9/A6ETkJ7yzbcmBh2AKOUCLyEjAL6CMiecBvgWgAY8y/gEXAPCAXqAcuDk kkcuP9/B04Oci4gIagLM1sf JQ4ELgLUissp37DZgIOjPop/8eQ D/rOoFQiVUkqpHk5vEyillFI9nCYDSimlVA nyYBSSinVw2kyoJRSSvVwmgwopZRSPZwmA0oppVQPp8mAUqpVIpLebOvUQhHZ6/u6VkT EaQ bxCRC1t5/gRfbQqlVABonQGllN9E5E6g1hhzfxD7iMJbi32yb7OvA50jvnMONcbUBysWpXoKHRlQSnWIiMwSkfd8X98pIs IyJcisktEThWRP4nIWhFZ7Ku9johMEZHPRWS5iHzYwla2RwMrmhIBEblORDb4drF8GcBXfe0zIGj7uyvVk2gyoJQKlGF4/5CfBDwPLDXGjMdbPnW LyF4CDjdGDMFeBL44wHaORRY3uz7W4BJvl0sr2x2fBlweMBfhVI9kO5NoJQKlA MMU4RWYt334nFvuNrgcHAKGAc8LFvzxUrcKANa2z8eHOqNcALIvIWP966tRjIDFz4SvVcmgwopQLFDmCM8YiIs9lGKh68v2sEWG MmdlGOw1AXLPv5wNHACcCt4vIeN8thDjfuUqpTtLbBEqpUNkM9BWRmeDdw11Exh7gvI3AcN85FmCAMWYpcDPQC0jynTcSWBf0qJXqATQZUEqFhDHGgXcr1vtEZDWwCjjkAKd gHckALy3Ep733XpYCTxojKn0PXcU8H4wY1aqp9ClhUqpiCMibwK/NsZsbeH5/sCLxphjQhuZUt2TJgNKqYgjIqOA/saYL1p4fhrgNMasCmlgSnVTmgwopZRSPZzOGVBKKaV6OE0GlFJKqR5OkwGllFKqh9NkQCmllOrhNBlQSimlerj/ByTjOFC8sFjkAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# import openpyxl\n",
    "# from openpyxl import load_workbook\n",
    "\n",
    "# wb = openpyxl.Workbook()\n",
    "# wb.remove_sheet(wb.active)\n",
    "# wb.create_sheet('DDNN') \n",
    "# sheet = wb['DDNN']\n",
    "\n",
    "# sheet['A1'].value = 'Force_1'\n",
    "# sheet['B1'].value = 'Force_2'\n",
    "# sheet['C1'].value = 'Force_3'\n",
    "\n",
    "# for i in range(len(pred)):\n",
    "#     sheet['A'   str(i 2)].value = pred[i,0]\n",
    "#     sheet['B'   str(i 2)].value = pred[i,1]\n",
    "#     sheet['C'   str(i 2)].value = pred[i,2]\n",
    "\n",
    "# wb.save('Prediction.xlsx')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **Convert jupyter notebook file to python file**\n",
    "Uncomment the script below to convert jupyter notebook file (.ipynb) to python file (.py)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# !jupyter nbconvert --to script DataAnalysis_0804.ipynb"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.9.13 64-bit",
   "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": "3.9.13"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "52634da84371cba311ea128a5ea7cdc41ff074b781779e754b270ff9f8153cee"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}