From 4de35d9048c4072615981d6117d69f9864c8d9cb Mon Sep 17 00:00:00 2001 From: John Kyle Cooper Date: Sun, 11 Jun 2023 15:04:12 +0200 Subject: [PATCH 1/7] add additional mcca example --- examples/example_mcca_2.ipynb | 173 ++++++++++++++++++++++++++++++++++ examples/example_mcca_2.py | 96 +++++++++++++++++++ 2 files changed, 269 insertions(+) create mode 100644 examples/example_mcca_2.ipynb create mode 100644 examples/example_mcca_2.py diff --git a/examples/example_mcca_2.ipynb b/examples/example_mcca_2.ipynb new file mode 100644 index 00000000..3983364c --- /dev/null +++ b/examples/example_mcca_2.ipynb @@ -0,0 +1,173 @@ +{ + "cells": [ + { + "attachments": {}, + "cell_type": "markdown", + "metadata": {}, + "source": [ + "# Example 1 - sinusoidal target in separable noise.\n", + "reproduced from de Cheveigné et al. (2018).\n", + "\n", + "Synthetic data for this example consisted of 10 data matrices, each of dimensions 10000 samples \n", + " 10 channels. Each was obtained by multiplying 9 Gaussian noise time series (independent and uncorrelated) by a 9 x 10 mixing matrix with random Gaussian coefficients. To this background of noise was added a “target” consisting of a sinusoidal time series multiplied by a 1 x 10 mixing matrix with random coefficients. The target was the same for all data matrices, but the mixing matrices differed, as did the noise matrices. The SNR was set to 10−20, i.e. a very unfavorable SNR. The noise is of rank 9 and the signal of rank 1, so signal and noise are in principle linearly separable." + ] + }, + { + "cell_type": "code", + "execution_count": 171, + "metadata": {}, + "outputs": [], + "source": [ + "import numpy as np\n", + "import matplotlib.pyplot as plt\n", + "from scipy import signal" + ] + }, + { + "cell_type": "code", + "execution_count": 172, + "metadata": {}, + "outputs": [], + "source": [ + "# Set the seed for the random number generator for reproducibility\n", + "np.random.seed(5)\n", + "\n", + "# Constants\n", + "num_matrices = 10\n", + "num_samples = 10000\n", + "num_channels = 10\n", + "noise_rank = 9\n", + "signal_rank = 1\n", + "unfavorable_SNR_dB = -20 # SNR in decibels\n", + "\n", + "# Generate noise matrices and mixing matrices\n", + "noise_matrices = [np.random.normal(size=(num_samples, noise_rank)) for _ in range(num_matrices)]\n", + "mixing_matrices = [np.random.normal(size=(noise_rank, num_channels)) for _ in range(num_matrices)]\n", + "\n", + "# Generate sinusoidal target\n", + "t = np.linspace(0, 1, num_samples)\n", + "target_signal = np.sin(2 * np.pi * t) # 1 Hz sinusoidal signal\n", + "\n", + "# Generate signal mixing matrix\n", + "signal_mixing_matrix = np.random.normal(size=(signal_rank, num_channels))\n", + "\n", + "# Prepare data matrices\n", + "data_matrices = []\n", + "for i in range(num_matrices):\n", + " # Create noise for current data matrix\n", + " noise = np.matmul(noise_matrices[i], mixing_matrices[i])\n", + "\n", + " # Create signal for current data matrix\n", + " signal = np.matmul(target_signal.reshape(-1, 1), signal_mixing_matrix)\n", + "\n", + " # Adjust the power of signal to achieve the desired SNR\n", + " noise_power = np.mean(noise**2)\n", + " signal_power = 10**(unfavorable_SNR_dB / 10) * noise_power\n", + " signal = np.sqrt(signal_power / np.mean(signal**2)) * signal\n", + "\n", + " # Add signal and noise\n", + " data_matrix = signal + noise\n", + "\n", + " data_matrices.append(data_matrix)" + ] + }, + { + "cell_type": "code", + "execution_count": 173, + "metadata": {}, + "outputs": [], + "source": [ + "# Concatenate data matrices\n", + "x = np.concatenate(data_matrices, axis=-1)" + ] + }, + { + "cell_type": "code", + "execution_count": 174, + "metadata": {}, + "outputs": [], + "source": [ + "from meegkit import cca\n", + "\n", + "# Compute Covariance matrix\n", + "C = np.dot(x.T, x)\n", + "\n", + "# Compute mCCA\n", + "A, score, AA = cca.mcca(C, 10)" + ] + }, + { + "cell_type": "code", + "execution_count": 175, + "metadata": {}, + "outputs": [], + "source": [ + "# Compute the recovered signal using first SC\n", + "x_recovered = x.dot(A)[:,0]\n", + "# Normalize the recovered signal\n", + "x_recovered = x_recovered / x_recovered.std()\n", + "# Compute variance across SCs\n", + "variance = np.var(x.dot(A), axis=0)" + ] + }, + { + "cell_type": "code", + "execution_count": 176, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAABKYAAAGGCAYAAABBiol3AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAADiZUlEQVR4nOzdeVxU9foH8M8wwADKoiA7Am6I4oqpWGpm4tKilVfLorym5fX2K+V2LSy7ZYvZYrSp19TMFrUyy25uWK6JpgiuuIMsggjIDsMwc35/wAyMDDADM5xZPu/Xa14v58yZc55B+XrOM8/3+UoEQRBARERERERERETUzuzEDoCIiIiIiIiIiGwTE1NERERERERERCQKJqaIiIiIiIiIiEgUTEwREREREREREZEomJgiIiIiIiIiIiJRMDFFRERERERERESiYGKKiIiIiIiIiIhEwcQUERERERERERGJgokpIiIiIiIiIiISBRNTZFISiUSvx759+8QOVcu5c+fw+uuvIz09XexQiKwOxwXD7du3T/NzSUxMbPT6zJkz0bFjx1Yde+bMmQgJCWljhERERGSp1q9fr3UNZm9vDz8/Pzz66KO4dOmS2OFZDPX1mrldw1oCe7EDIOt2+w3Um2++ib179+KPP/7Q2t6nT5/2DKtF586dwxtvvIG7776bN2xERsZxoW0WLlyIgwcPGu14ixcvxgsvvGC04xEREZFl+vLLL9G7d29UVVXhzz//xNtvv429e/fi/Pnz6NSpk9jhkRVjYopMavjw4VrPu3TpAjs7u0bbW6uiogIuLi5GORYRtQ9bHhfUSa3169e36v0TJkzAzp078euvv+KBBx4wSkzdu3c3ynGIiIjIskVERGDIkCEAaq9ZlEol/vOf/+Dnn3/G3//+d5GjMx1BEFBVVQVnZ2exQ7FZnMpHovv8888xatQoeHt7o0OHDujXrx/ee+89KBQKrf3uvvtuRERE4MCBAxgxYgRcXFwwa9YsAEBWVhamTp0KV1dXeHh44PHHH8exY8cgkUga3QAeP34cDz74IDp37gwnJycMGjQI33//veb19evX429/+xsAYMyYMZqS1tbeSBKR4Tgu6DZz5kz06dMHcXFxUCqVze6rUqnw3nvvoXfv3pDJZPD29saTTz6JrKysRse8vQLshx9+wLBhw+Du7g4XFxd069ZN83NVKykpwYsvvojQ0FA4OjoiICAA8+fPR3l5uVE+KxEREYlLnaS6ceOGZltL10xq2dnZeOaZZxAUFARHR0f4+/tj6tSpWsfKyMjAE088AW9vb8hkMoSHh+PDDz+ESqUCACgUCnh7eyMmJqbR8YuKiuDs7IzY2FjNNn2vTSQSCZ577jmsWrUK4eHhkMlk+OqrrwAAly5dwowZM7Ri+vzzzxud//z585gwYQJcXFzg5eWFuXPnorS01JAfLzXAiikS3ZUrVzBjxgzNAHLy5Em8/fbbOH/+PNatW6e1b05ODp544gksXLgQ77zzDuzs7FBeXo4xY8agsLAQy5YtQ48ePbBz505Mnz690bn27t2LCRMmYNiwYVi1ahXc3d2xadMmTJ8+HRUVFZg5cybuu+8+vPPOO1i0aBE+//xzDB48GACrCojaE8cF3aRSKZYuXYrJkyfjq6++apQsaugf//gHVq9ejeeeew73338/0tPTsXjxYuzbtw8nTpyAl5eXzvclJiZi+vTpmD59Ol5//XU4OTnh2rVrWlMtKyoqMHr0aGRlZWHRokXo378/zp49i9deew2nT5/Gnj17IJFIjP75iYiIqP2kpaUBAHr16gVAv2smoDYpdccdd0ChUGiuEwoKCrBr1y7cunULPj4+uHnzJkaMGIHq6mq8+eabCAkJwf/+9z+8+OKLuHLlClasWAEHBwc88cQTWLVqFT7//HO4ublpYtu4cSOqqqo0lVyGXpv8/PPPOHjwIF577TX4+vrC29sb586dw4gRI9C1a1d8+OGH8PX1xa5du/D8888jPz8f//nPfwDUJupGjx4NBwcHrFixAj4+Pvj222/x3HPPtcdfi3USiNrRU089JXTo0KHJ15VKpaBQKIQNGzYIUqlUKCws1Lw2evRoAYDw+++/a73n888/FwAIO3bs0Nr+7LPPCgCEL7/8UrOtd+/ewqBBgwSFQqG17/333y/4+fkJSqVSEARB+OGHHwQAwt69e1v5SYlIX9Y6LqhUKkGhUGg9Ro0aJTz55JONtrdk7969AgDhhx9+EARBEO666y4hMDBQqKysFASh8c8wNTVVACDMmzdP6zhHjx4VAAiLFi3SbHvqqaeE4OBgzfMPPvhAACAUFRU1Gc/SpUsFOzs74dixY1rbf/zxRwGAsH379hY/ExEREZmHL7/8UgAgHDlyRFAoFEJpaamwc+dOwdfXVxg1apTmWkXfa6ZZs2YJDg4Owrlz55o858svvywAEI4ePaq1/R//+IcgkUiECxcuCIIgCKdOnRIACKtXr9bab+jQoUJkZKTmuSHXJgAEd3d3rWtKQRCE8ePHC4GBgUJxcbHW9ueee05wcnLS7P/SSy8JEolESElJ0dpv3LhxvIdsJU7lI9ElJyfjwQcfhKenJ6RSKRwcHPDkk09CqVTi4sWLWvt26tQJ99xzj9a2/fv3w9XVFRMmTNDa/thjj2k9v3z5Ms6fP4/HH38cAFBTU6N5TJo0CTk5Obhw4YIJPiERGcoaxoX9+/fDwcFB63HgwAFs2LCh0XZDV/pbtmwZsrKy8PHHH+t8fe/evQCg+eZSbejQoQgPD8fvv//e5LHvuOMOAMC0adPw/fffIzs7u9E+//vf/xAREYGBAwdq/czGjx/P1WiIiIgs1PDhw+Hg4KC5hurUqRN++eUX2NvbG3TNtGPHDowZMwbh4eFNnuuPP/5Anz59MHToUK3tM2fOhCAImkrtfv36ITIyEl9++aVmn9TUVPz1119aleOGXpvcc889Wg3dq6qq8Pvvv+Ohhx6Ci4tLo89XVVWFI0eOAKi9zurbty8GDBigdcwZM2bo+6Om2zAxRaLKyMjAyJEjkZ2djY8//hgHDx7EsWPHNPN4Kysrtfb38/NrdIyCggL4+Pg02n77NvV85hdffLHRTeG8efMAAPn5+Ub5XETUetYyLkRGRuLYsWNaj8GDB+P+++9vtN3f39+gY48YMQJTpkzBu+++i1u3bjV6vaCgAIDun42/v7/mdV1GjRqFn3/+GTU1NXjyyScRGBiIiIgIbNy4UbPPjRs3cOrUqUY/M1dXVwiCwLGUiIjIAm3YsAHHjh3DH3/8gWeffRapqamaL/UMuWa6efMmAgMDmz1XQUFBk9cp6tfVZs2ahcTERJw/fx5A7eqBMplM6wtHQ69Nbj93QUEBampq8OmnnzY6xqRJk7Q+X0FBAXx9fRvFrmsb6Yc9pkhUP//8M8rLy/HTTz8hODhYsz0lJUXn/rp6lnh6euKvv/5qtD03N1frubqfSlxcHB5++GGdxw8LC9M3dCIyEWsZF1xdXTVNQxtu8/T0bLS9NZYuXYqIiAi88847jV7z9PQEUNt/6/YLw+vXrzfZX0pt8uTJmDx5MuRyOY4cOYKlS5dixowZCAkJQVRUFLy8vODs7Nyo35daS8cnIiIi8xMeHq65RhkzZgyUSiXWrFmDH3/8Ef369QOg3zVTly5dGi22cjtPT0/k5OQ02n79+nUA2tcSjz32GGJjY7F+/Xq8/fbb+PrrrzFlyhStiidDr01uv37s1KkTpFIpYmJi8M9//lPnMUJDQzWx335NCTS+ziT9MTFFolIPCDKZTLNNEAR88cUXeh9j9OjR+P7777Fjxw5MnDhRs33Tpk1a+4WFhaFnz544efKkzhu5htTx3F6ZQUSmx3FBP71798asWbPw6aefYsSIEVqvqac2fvPNN5qpeQBw7NgxpKam4pVXXtHrHDKZDKNHj4aHhwd27dqF5ORkREVF4f7778c777wDT09PzUUaERERWZf33nsPW7ZswWuvvYYzZ87ofc00ceJEfP3117hw4UKTX/CNHTsWS5cuxYkTJzSLygC1VVsSiQRjxozRbOvUqROmTJmCDRs2ICoqCrm5uY0WgGnrtYmLiwvGjBmD5ORk9O/fH46Ojk3uO2bMGLz33ns4efKk1nS+7777zuDzUi0mpkhU48aNg6OjIx577DEsXLgQVVVVWLlypc6pKU156qmn8NFHH+GJJ57AW2+9hR49emDHjh3YtWsXAMDOrn7G6n//+19MnDgR48ePx8yZMxEQEIDCwkKkpqbixIkT+OGHHwAAERERAIDVq1fD1dUVTk5OCA0N1VQhEJHpcFzQ3+uvv45vv/0We/fuRYcOHTTbw8LC8Mwzz+DTTz+FnZ0dJk6cqFmVLygoCAsWLGjymK+99hqysrIwduxYBAYGoqioCB9//DEcHBwwevRoAMD8+fOxZcsWjBo1CgsWLED//v2hUqmQkZGB3bt341//+heGDRtm8s9PREREptOpUyfExcVh4cKF+O677/S+ZlqyZAl27NiBUaNGYdGiRejXrx+Kioqwc+dOxMbGonfv3liwYAE2bNiA++67D0uWLEFwcDB+++03rFixAv/4xz80KwGqzZo1C5s3b8Zzzz2HwMBA3HvvvVqvG+Pa5OOPP8Zdd92FkSNH4h//+AdCQkJQWlqKy5cv49dff9X0vZo/fz7WrVuH++67D2+99ZZmVT71VENqBVFbr5PN0bX61q+//ioMGDBAcHJyEgICAoR///vfwo4dOxqtaDB69Gihb9++Oo+bkZEhPPzww0LHjh0FV1dX4ZFHHhG2b98uABB++eUXrX1PnjwpTJs2TfD29hYcHBwEX19f4Z577hFWrVqltV98fLwQGhoqSKXSRqt4EZHx2NK4MHr0aOGpp54y6D2C0HhVvoYWLVokAGj0M1QqlcKyZcuEXr16CQ4ODoKXl5fwxBNPCJmZmVr73b4q3//+9z9h4sSJQkBAgODo6Ch4e3sLkyZNEg4ePKj1vrKyMuHVV18VwsLCBEdHR8Hd3V3o16+fsGDBAiE3N9fgz0hERETiUK/Kd/uKdoIgCJWVlULXrl2Fnj17CjU1NXpfM2VmZgqzZs0SfH19BQcHB8Hf31+YNm2acOPGDc0+165dE2bMmCF4enoKDg4OQlhYmPD+++9rVvdrSKlUCkFBQQIA4ZVXXtH5OfS9NgEg/POf/9R5jLS0NGHWrFlCQECA4ODgIHTp0kUYMWKE8NZbb2ntd+7cOWHcuHGCk5OT0LlzZ+Hpp58WfvnlF67K10oSQRAEcVJiRKb1zjvv4NVXX0VGRkaLzfeIyDZwXCAiIiIiMi+cykdW4bPPPgNQ23NFoVDgjz/+wCeffIInnniCN59ENorjAhERERGR+WNiiqyCi4sLPvroI6Snp0Mul6Nr16546aWX8Oqrr4odGhGJhOMCEREREZH541Q+IiIiIiIiIiIShV3LuxARERERERERERkfE1NERERERERERCQKJqaIiIiIiIiIiEgUbH5uBCqVCtevX4erqyskEonY4RC1K0EQUFpaCn9/f9jZMddtrjhOka3iGGUZOEaRLeM4ZRk4TpEtM/U4xcSUEVy/fh1BQUFih0EkqszMTAQGBoodBjWB4xTZOo5R5o1jFBHHKXPHcYrIdOMUE1NG4OrqCqD2L8nNzU3kaIjaV0lJCYKCgjS/B2SeOE6RreIYZRk4RpEt4zhluAMHDuD9999HUlIScnJysHXrVkyZMqXJ/fft24cxY8Y02p6amorevXvrdU6OU2TLTD1OMTFlBOpSTjc3Nw5SZLNY0mzeOE6RreMYZd44RhFxnDJEeXk5BgwYgL///e945JFH9H7fhQsXtMaYLl266P1ejlNEphunmJgiIiIiIiIiizFx4kRMnDjR4Pd5e3vDw8PD+AERUZuwux4RERERERFZvUGDBsHPzw9jx47F3r17xQ6HiOqwYoqIiIiIiIislp+fH1avXo3IyEjI5XJ8/fXXGDt2LPbt24dRo0bpfI9cLodcLtc8Lykpaa9wiWwOE1NERERERERktcLCwhAWFqZ5HhUVhczMTHzwwQdNJqaWLl2KN954o71CJLJpnMpHRERERERENmX48OG4dOlSk6/HxcWhuLhY88jMzGzH6IhsCyumiIiIiIiIyKYkJyfDz8+vyddlMhlkMlk7RkRku5iYIiIiIiKzplQqcfDgQeTk5MDPzw8jR46EVCoVOywiEklZWRkuX76seZ6WloaUlBR07twZXbt2RVxcHLKzs7FhwwYAQHx8PEJCQtC3b19UV1fjm2++wZYtW7BlyxaxPgIRNWBRU/kOHDiABx54AP7+/pBIJPj5559bfM/+/fsRGRkJJycndOvWDatWrWq0z5YtW9CnTx/IZDL06dMHW7duNUH0RERERG2zYsUKhIaGwsnJCZGRkTh48GCz+xvjOkif6y9BEPD666/D398fzs7OuPvuu3H27Nk2fVa1n376CSEhIRgzZgxmzJiBMWPGICQkBD/99JNRjk9Eluf48eMYNGgQBg0aBACIjY3FoEGD8NprrwEAcnJykJGRodm/uroaL774Ivr374+RI0fi0KFD+O233/Dwww+LEj8RabOoxFR5eTkGDBiAzz77TK/909LSMGnSJIwcORLJyclYtGgRnn/+ea3MeGJiIqZPn46YmBicPHkSMTExmDZtGo4ePWqqj0FERERksM2bN2P+/Pl45ZVXkJycjJEjR2LixIlaN18NGes6SJ/rr/feew/Lly/HZ599hmPHjsHX1xfjxo1DaWlpmz7zTz/9hKlTpyIrK0tre3Z2NqZOncrkFJGNuvvuuyEIQqPH+vXrAQDr16/Hvn37NPsvXLgQly9fRmVlJQoLC3Hw4EFMmjRJnOCJqBGJIAiC2EG0hkQiwdatWzFlypQm93nppZewbds2pKamarbNnTsXJ0+eRGJiIgBg+vTpKCkpwY4dOzT7TJgwAZ06dcLGjRv1iqWkpATu7u4oLi6Gm5tb6z4QkYXiv/+2O3DgAN5//30kJSUhJyen0dg2c+ZMfPXVV1rvGTZsGI4cOaL3Ofj3RLbKmv7tDxs2DIMHD8bKlSs128LDwzFlyhQsXbq00f6muA7Sdf0lCAL8/f0xf/58vPTSSwBql1n38fHBsmXL8Oyzz7b42XT9PSmVSoSEhDRKSjWMJTAwEGlpaZzWRxbNmsYpa8a/J7Jlpv73b1EVU4ZKTExEdHS01rbx48fj+PHjUCgUze5z+PDhJo8rl8tRUlKi9bAGp7KK8M72VMSsPYrp/01E7OYU/JycDXmNUuzQiKyaPtUIEyZMQE5Ojuaxffv2doyQLNHx/23FnrUrYaHfP9FtqqurkZSU1OiaJTo6uslrFlNdB90uLS0Nubm5WseRyWQYPXp0k8fR51rq4MGDTSalgNqEWGZmZovTGYmIqF5eaRXWHkrD3K+T8LdVhzH7q+NYue8KcourxA6NbJhVNz/Pzc2Fj4+P1jYfHx/U1NQgPz8ffn5+Te6Tm5vb5HGXLl2KN954wyQxiyGzsAL/2XYWf5zPa/TaT8nZWLpDhtcf6IuJ/ZpetYKIWm/ixImYOHFis/vIZDL4+vq2U0RkDfZ/vRYA0GfkGPj36i1yNNRW+fn5UCqVBl2zmOo6SNd51O+7/TjXrl3T+R59rqVycnL0Or+++xER2bLqGhU+33sZ/z1wBVUKldZre1JvYHnCBTwVFYIXx4fByYFVqNS+rDoxBdSWeTek/ua44XZd+9y+raG4uDjExsZqnpeUlCAoKMgY4ba7Q5fy8dzGEyiqUMDeToKJ/fxwVw9PODva40JuCbYkZSO3pAr/+PYEnooKxmsP9IXUrumfDRGZxr59++Dt7Q0PDw+MHj0ab7/9Nry9vZvcXy6XQy6Xa55bS2UnGU4h5zeg1sTQaxZTXAcZIzZ9rqWaW8a9NfsREdmq/DI5/vFNEo6l3wIA9A90x/39/eDv4Yy8Ejl2nsnFX+mFWHMoDYcu52P934fC191J5KjJllh1YsrX17fRN355eXmwt7eHp6dns/vc/q1fQzKZDDKZzPgBt7N9F/LwzIYkVCtVGBDkgeXTBqB7l471Owzwx/Nje+LjPZewYt8VfJV4DbcqFIifPhB2TE4RtZuJEyfib3/7G4KDg5GWlobFixfjnnvuQVJSUpNjkblUdlZnl6E6swQdhvpBwnGDqNW8vLwglUoNumYx1XWQrvMAtZVTDZNEzR1Hn2upkSNHIjAwENnZ2TqnpKp7TI0cOVLvWImIbE1BmRyPrT6CS3llcJXZ452H++H+/n5aXxzMuisUe8/n4d8/nsL53FI8vOJP/PCPEQjwcBYxcrIlVt1jKioqCgkJCVrbdu/ejSFDhsDBwaHZfUaMGNFucYohJbMIz35dm5Sa0NcXm58Zrp2UqiOzl2LhhN749LFBcJBKsO3kdbz1W6qOIxKRqUyfPh333XcfIiIi8MADD2DHjh24ePEifvvttybfExcXh+LiYs0jMzOzHSOul/dpMop+voLKUzdFOT+RtXB0dERkZGSja5aEhIQmr1na6zooNDQUvr6+Wseprq7G/v3723Q9JZVK8fHHHwNoXI2lfh4fH8/G50RETahSKDHzy2O4lFcGHzcZfn7uTjwwwF9nNeuY3t7YOm8Eunl1wPXiKjy17i8UVVSLEDXZIotKTJWVlSElJQUpKSkAapttpqSkaJZJjouLw5NPPqnZf+7cubh27RpiY2ORmpqKdevWYe3atXjxxRc1+7zwwgvYvXs3li1bhvPnz2PZsmXYs2cP5s+f354frV0Vlldj3jdJkNeoMCasCz55bFCL84gfGOCP96cOAACs+zMNPydnt0eoRKSDn58fgoODcenSpSb3kclkcHNz03qISZFbLur5iaxBbGws1qxZg3Xr1iE1NRULFixARkYG5s6dC8B010EtXX9JJBLMnz8f77zzDrZu3YozZ85g5syZcHFxwYwZM9r0mR9++GH8+OOPCAgI0NoeGBiIH3/8EQ8//HCbjk9EZK0EQcDin8/gdHYxOndwxLezdRciNBTU2QXfzB4GXzcnXM4rw4s/nOQiKtQuLCoxdfz4cQwaNAiDBg0CUHuBNmjQILz22msAaptfqi+SgNpv8LZv3459+/Zh4MCBePPNN/HJJ5/gkUce0ewzYsQIbNq0CV9++SX69++P9evXY/PmzRg2bFj7frh2IggC/vV9Cq4XVyHUqwM+fmwQHO31+2cwZVAAnh/bEwDwytbTSMvnjSaRGAoKCpCZmcm+KqQfXk9ajenTpyM+Ph5LlizBwIEDceDAAWzfvh3BwcEATHcd1NL1FwAsXLgQ8+fPx7x58zBkyBBkZ2dj9+7dcHV1bfPnfvjhh5Geno6QkBAAwIcffoi0tDQmpYiImrE1ORs/JGXBTgJ8+tgg9PBuPiml5u/hjDVPDYGj1A57UvPw5Z/ppg2UCIBEYAq0zUpKSuDu7o7i4mLRqxJasjU5Cws2n4Sj1A7b/u9O9PY1LN4apQoz1hzFX2mFGBraGZufGd6qBqlkPSzp37+5Kisrw+XLlwEAgwYNwvLlyzFmzBh07twZnTt3xuuvv45HHnkEfn5+SE9Px6JFi5CRkYHU1FS9b/rE+nvKerl2GXfXuwPhPiG03c5LwIfT7wcATH3lLQT3HyhuMCLiGGUZ9Pl7GjBgAE6dOoWEhATce++97RwhkelwnLIMlvT3dLNUjnuX70dxpQKx43ppigsM8dXhdPxn21nI7O2we8EoBHt2MEGkZClM/e/foiqmqG0KyuRY8us5AMDzY3sYnJQCAHupHZZPGwBnByn+SivElhOc0kfUVs1VI0ilUpw+fRqTJ09Gr1698NRTT6FXr15ITEw0SiUCEZGlsLOrvWxVqVQt7ElEZNve+PUsiisV6OPnhn/c3b1Vx3gyKhh39vCEvEaF1345yyl9ZFJWvSofafv490u4VaFAb19XPDu6dQMUAAR2csEL9/bEuzvO453tqRjXxwfuzg5GjJTIttx9993N/me/a9eudoyGiMg8qSu0mZgiImraX2mF+N+pHNhJgPem9oeDtHW1KBKJBG9OjsCE+IPYf/Emdp7JxcR+bCNBpsGKKRtx9WYZvjta23fitQf6tHqAUnv6rlD08O6IwvJqfHHgqjFCJCIiImqSumKK39oTEekmCALe2V67gvqjQ7siIsC9Tcfr1qUjnh3dDQDw/q4LqFHyiwEyDSambMT7uy6gRiXgnt7eGNHdq83Hc5Da4d/jwwAAaw+lIa+0qs3HJCIi4xPY/ZysBKfyERE1b8eZXKRkFsHFUYr59xreV0qXZ0Z1Q+cOjriaX44fk7KMckyi2zExZQPOXS/BjjO5sJMAL03obbTjRvfxwcAgD1QqlFix94rRjktE1oiLJBBR2zAxRUTUNJVKwEcJFwEAs0d2g7erk1GO6+rkgHl1fari91yCvEZplOMSNcTElA1Ytb82aXRff3+E+RqvWbJEItFUTW38KwMFZXKjHZuIiIioISamiIia9vv5PFzKK4OrzB6zRxp3JeQnhgfD180JuSVV+DmZi1+R8TExZeWuFZTjf6euAwDm1s0PNqYR3T0xINAd8hoVvjqcbvTjExEREQFMTBERNUUQBKzYdxkA8ERUMNycjLswlZODVJPs+u+Bq1Cp2CaAjIuJKSu3+sBVqATg7rAu6OvftuZ3ukgkEs0Kf18lXkO5vMbo5yAiIiLiqnxERLodTStEckYRHO3tMOtO41ZLqT06tCvcnOxx9WY5ElJvmOQcZLuYmLJiRRXVmgZ1c+uSR6Ywvq8vQjxdUFypYEM8ItLyfZAD5g5xRjkbcIuHK5iRleCqfEREuq09lAYA+FtkILq4ykxyjo4ye8REBQMAV2Uno2Niyor9mJQFeY0K4X5uGBba2WTnkdpJMOuu2sz810eu8YKRiDTe6+OE45722CBlDzoiahtO5SMiaux6USV+r6tg+vudISY911MjQmBvJ8Hxa7eQmlNi0nORbWFiykqpVAK+PZoBAHhieFdN+bupPDQoAC6OUlzOK8NfaYUmPRcRWZ6KNlZMJW69jKPb+O0ckS1jYoqIqLFNf2VAJQDDu3VGD2/jLXSli7erE8b39QUAfHv0mknPRbaFiSkrdfhKAdLyy9FRZo8pAwNMfj5XJwdMHugPAJqEGBGRRhty4xUl1TixKwPHt6ejuop97IhsFRNTRETaFEoVNh7LBFC7cl57eHxYVwDA1hPZKGN/YTISJqas1DdHajPYDw0KQAeZfbuc8/FhtYPhjjM5yC/jtB0iMg5lTf1NKGcKG666sgLfL1mEU3t2ih0KUZswMUVEpG332Ru4WSqHV0cZovv4tss5o7p7optXB5RXK7Et5Xq7nJOsHxNTVqiwvBp76uYZPz68a7udNyLAHQOCPKBQCth6IrvdzktERE079utPyDx7CglffCZ2KEajLCpC0U9boSwrFzsUakdclY+ISNsPSbXVUtPvCISjffvc2kskEsyoq5radIwzZcg4mJiyQv87dR01KgF9/d3Q29etXc89NTIQALA1mYkpIiJzIK+oEDsEo8t8di5yFi1C7muviR0KtSOuykdEVO9mqRwHL+UDAB4ZHNiu554yKABSOwlOZRXjys2ydj03WScmpqzQT3XVSg+38wAFAPf384ODVIJzOSW4kFva7ucnImqtW7nlSNqZDoVcKXYoNktVrUTh5guoPJuvtf3s2bP48MMPkZlZ+81w5cmTAICS7dvbPUYSD6fyERHV23byOpQqAQODPNCtS8d2PbdXRxlG9+oCAJwpQ0bBxJSVuXqzDCmZRZDaSfDgAP92P3+nDo4YE+YNgFVTRFSvNfUNgiCgYN2XqExONno8unz3+lEc+fkqjvxypV3O1xxVRQWUJaZfhrmiuAg11dUmP4++yg5moyI5DwVfp2pt/+GHH1BaWorvvvtOpMjIHDAxRURUb2tyFgDg4cGmX+hKl4cG1Z7355RsqFSsZKW2YWLKyvxclwwa2dMLXVxlosSgHqR+4SBFRG1Qtm8f8t57D9mxse163tyrpk8IteTCkDtwcegwqMqN0EOpiWlPZYUFWPnME/jiuVltP4eRKEuaXziDCQnbxsQUEVGtSzdKcSa7BPZ2Etzfv/2LEQBgXB8fdJTZI+tWJY5fuyVKDGQ9mJiyIoIgYGtKbWJKnRwSwz3h3nBzskdOcRWOXC0QLQ4iaj+qykrkvP46yg79abRjKuqmbVmKtva9uXijFKeyimqf1N14y9PS2xZUMzLO1E6HqyguMvi9O24WYVjiOZwoEbn5eF0zbLINTEwREdVSz0y5O8wbnTs4ihKDk4MUEyN86+LJEiUGsh5MTFmR09nFyCyshIujtN2WC9VFZi/FpH5+AIDfTueIFgcRtZ+CtetQtGkzMmfPFjuUFpUeyELBxvMQjFjRmZ2djffffx/JbZh2GP3RATz42Z8oqjCfqXVN+fuZdFyrqkbMqTSxQyEbwlX5iIhqvwhT32NNGSROtZTalLpiiJ1nclGj5NhMrcfElBXZcSYXADAmzBvOjlJRY5lYl5jadTYXSk7nI7J6imzj9ZQrk9fgxR9O4lJe/SovgiCHINQnbFRVNbjxyQmU7DO8qqp4exoqT95E1YVCo8QL1PZAqqiowC+//NLmY+WVNj+dzZxUNZEgyJUrUM3kARkZV+UjIgJSc0pxraACMns7TW9fsQwL7YxOLg64VaHA0TTjXVeR7WFiykoIgoAddZnzCRHiVUupRXXzhJuTPfLLqnE8nYMUEenv0z8u4cekLHx3NAMAoAIgL/oc8qLPoFLVrphXdigbiuvlKNmZ3urzCNXGS5zwRrnemdIKDDx8FuOPXxQ7FLIynMpHRATsOFN7zze6Vxd0kNmLGou91E4zU0cdF1FrMDFlJc7nliJdnTnvLW7mHAAc7e0wTjNI5YocDRFZkutFVVrPq+3r/6tSVNW+JiiZCDIHlTrK9n+8UdsANbW8qtFrRsceUzaFiSkiovp7K3XrFLFN7Fd7z7fzzA3OlKFWY2LKSqgHqFG9uqCjyJlzNXUzvJ1ncrk6H5GN02cEEAQBNQqFyWMxlgoL76VgjCqv9v4JsDLNtjExRUS27nJeKS7nlcFBKsE94eIXIwDAiO5ecHWyR36ZHElcnY9ayeISUytWrEBoaCicnJwQGRmJgwcPNrnvzJkzIZFIGj369u2r2Wf9+vU696mqaodveo1oZ13p5EQzmMandldPL3SU2SO3pArJmUVih0NEZu7Htxfjk5hHYFdd0fROZpKY+DG3EN0OnML67HyxQ7E6lzvaobyFNomFnTrh4F13ocS1Y/sERWaBiSkisnU7TtcWI9zVwwtuTg4iR1OrdqaMDwBgOxe+olayqMTU5s2bMX/+fLzyyitITk7GyJEjMXHiRGRkZOjc/+OPP0ZOTo7mkZmZic6dO+Nvf/ub1n5ubm5a++Xk5MDJyak9PpJRXM4rw8UbtZnzseE+Yoej4eQgxT110wp3neV0PiJqXsbpFAiCCi7XUwHUV1lVOnmKF1QTnkut/X/n5YvWszxyya7dyF64EKo2fDFTXl4OZY3SoPdcOpaInEsXAABH7JV49M4OmH5nh2bfkzA+GtcDA3Bg5MhWx0qWh6vyEZGt2143S2aimUzjU5sYUb/wFaubqTUsKjG1fPlyPP3005g9ezbCw8MRHx+PoKAgrFy5Uuf+7u7u8PX11TyOHz+OW7du4e9//7vWfhKJRGs/X1/zqTrSx++pNwAAUd294O5sHplztei+tYkydYxEZLtq8vNR+O23UJaW6v0eARKc6/t0q86nKi9v3WqB5nRBZeJYFFJ75HbxhyAIyH7hBZRs+xWF679Cza1bKNu/H4JS/ySTXC7H+++/j6N//aX3ewqyMrDtg7fx3av/AgDsdKgBAOQ663d5Ut6RFVO2hKvyEZEtyy6qRGpOCewkwDgzKkYAgJE9veDkYIec4iqk5uh/nUekZjGJqerqaiQlJSE6Olpre3R0NA4fPqzXMdauXYt7770XwcHBWtvLysoQHByMwMBA3H///UhOTm72OHK5HCUlJVoPMf1+Pg8AMNYMmp7fbmTPLrC3k+DKzXKk55eLHQ4RGUiorsb1l+NQ/Ov/2nysjL//HTfefAs5r7yqtT3n0gWUFuqeEidItOd05Wek6X2+S3ePweWx96I6Pb2JPczn5rZCqYJ8eBcoerqZIBfV9AG/f2AWvn5kHjbm1K+eWlNQgPTpjyLz2bm49c03jd6jqJbrPFZBQUGzUaTs+g1n9/+uta3oBkv+SX+cykdEtuyPunu+wV07oVMHR5Gj0ebkIMVdPbwAAH+cZ0ECGc5iElP5+flQKpXw8dHODvv4+CA3t+VpYjk5OdixYwdmz56ttb13795Yv349tm3bho0bN8LJyQl33nknLl261OSxli5dCnd3d80jKCiodR/KCIorFJomc/eYYWLK3dkBd4R0BlA/mBKR5SjasgXFP/+M6//+d6PXDhSWIvZ8BkoUNXodq+LKNZzqOwv7L+cg+/w5AEBe+lV89+q/sPofM/U6xtZli/WO/WxAAPaPHoUSPb+8ENOWG4UQ3B2h7Obarue97tsVAPBtjnZSSVE3Rb5kx85G76muaKYHWBPKCgvw+7qV2LniI6hUhk31I1JjYoqIbNneunspc1iBXZd7etfNlOE9H7WCxSSm1CS3LQ0tCEKjbbqsX78eHh4emDJlitb24cOH44knnsCAAQMwcuRIfP/99+jVqxc+/fTTJo8VFxeH4uJizSMzM7NVn8UYDly6CaVKQA/vjgjq7CJaHM1RJ8z2XuAgRWRpagoKm3xt2skr+C6nEEs2p0BAyxdJWQGjkesqRa5rBTb9ZyEAIPvCOR17tlwyVH605elipwYOQK6fHy6IXNWqD4XIK5cqBSClZzgqZKbpr1hdVdnq9woNVmoULGxhEjIuJqaIyFZVKZQ4fKW2utwcixEAYEzvLgCAlMwiFJTprq4maorFJKa8vLwglUobVUfl5eU1qqK6nSAIWLduHWJiYuDo2HzZo52dHe64445mK6ZkMhnc3Ny0HmJRZ87NdYACoFnK9MjVApTJ9ausICIz4+CC6sz6ngHKBvPNsu0FCEJks28XANTYu0BQNZ3oMkRlC1OuG6pp8ia25S81dKk42XKSvSI5GQo9qnnNRUppBRbEvoZ/zX8Fl1064mxoz9YfrMGPtSg3B5Wl1VAq9KuSUtZo/x8hVFfj0l31Dc4N6XlF1ofNz4lI7cCBA3jggQfg7+8PiUSCn3/+ucX37N+/H5GRkXByckK3bt2watUq0wdqJIlXClClUMHP3Qm9fdu3ulpffu7O6ONX25Jg34WbYodDFsZiElOOjo6IjIxEQkKC1vaEhASMGDGi2ffu378fly9fxtNPt9xAVxAEpKSkwM/PvFY60EWpErDvYu0v/Zgw801MdfPqgBBPFyiUAg5d4iBFZIk6jnsbeZ+noOL8TXycfgM9Dpw22blaqh0qLaxCju8wzfMflx2HUtH0jaoxapGuX7+Ozz//HBcvXkThxgvN7lt55iyuPTYDl+8eY4Qz6yetQo4HT1zCnoISqFQqZGVlQWgi8aaQV+HHtxcjZfdvjV47H9IDUweOwnMLlyDfvVOb4yq4no91/z6EXz5O0fn67f20Khs0xlcplbi0azuqGjbL16NCmqwXK6aISK28vBwDBgzAZ599ptf+aWlpmDRpEkaOHInk5GQsWrQIzz//PLZs2WLiSI3jjwbT+PSZLSSWsXUFCWzhQoaymMQUAMTGxmLNmjVYt24dUlNTsWDBAmRkZGDu3LkAaqfYPfnkk43et3btWgwbNgwRERGNXnvjjTewa9cuXL16FSkpKXj66aeRkpKiOaY5O5lVhMLyarg62WNISNtvIExFIpFo5hxzkCKyTBLHDgCAIx9/g6VpOahslxtD3Rdev3z4FeR18QDAjbQSXEkx7tiiqtSeerZx40bcvHkT3333XYvvrTxxwuDztfUic965a/iruBxPnLqKffv2Yc2aNajyD9G5b8qu33DtVDJyL19s9pg3Onu1Oda89NpplJWl9dPxlIKAN69cx+784mbfe+zXn/DrN2twtLv5f1FE7YOr8hGR2sSJE/HWW2/h4Ycf1mv/VatWoWvXroiPj0d4eDhmz56NWbNm4YMPPjBxpG0nCILmHuoeMy5GAOpn8Ry4eBMKJb9EIP1ZVGJq+vTpiI+Px5IlSzBw4EAcOHAA27dv16yyl5OTg4y6hq1qxcXF2LJlS5PVUkVFRXjmmWcQHh6O6OhoZGdn48CBAxg6dKjJP09b/ZFaO0CN6tUFDlLz/qtUD1J/nL/JC0oi0pC0Yjrdjcu/QlWtnVQRlAIqT53ClfETUPrH3jbFpLx1CxcGDcatTZs02+Ry8+6VUNCgAf2ff/4JAKhx99S5b2ml/v2erru4IC/PsKRfSyP8L3nF+DwjD0+ebn6FxXN1K/iVuJim7xVZHlZMEVFrJSYmNlrdffz48Th+/DgUDXoZNmQuK7FfvFGG7KJKyOztcGcP3V8amYsBgR7w7OCIUnkNjqffEjscsiDmnc3QYd68eUhPT4dcLkdSUhJGjRqleW39+vXYt2+f1v7u7u6oqKjAnDlzdB7vo48+wrVr1yCXy5GXl4ddu3YhKirKlB/BaPZdtIzMOQDcEdoJzg5S5JfJcT63tOU3EJHVUUK/HnMSPZLXglDeaFvGnGdQfe0asubN0zsmuVyOQ4cOIfbUZewaVLuARHVWFgAg9/U39D5OW7UmYS+oVCj/6y8oyxr/LJry561SzPCOwME77m1x37IOHbC7axBWrFhhcGy61NhJccutM3Lkum8AAEDSUlrrtp+TqpV9wsgyMTFFRK2Vm5urc3X3mpoa5Ofn63yPuazEvq9uAamo7p5wdpSKEoO+7OwkGNWrtgn6QbZwIQNYXGKKahWWV+NMdm3WfmQv886cA4DMXoph3ToDAA5d0j34E5F5KVcq8Yp/DyRGDNJr/7MBXlA0Ub1ZUXwLl90uQClP0fv89i4SJA8aAEGiQo1zR5R1j0BNh6YXm1BVVOh+QRBQkn8Tl48fvS0BJGDnzp3Ytv8gvisow1+9nFDlIH6iQ9CzK9at7zbiwjPP4oeXXoJSz6bgiy9lAwCORN7d4r7lHTu0uE9D9T/b21bPhQqlnTPwzfTnsWZGLA7cKmv6GHokmipLa//vS+raC2tG3o+zZa1f8Y8sCxNTRNQWulZ317VdzVxWYj90ufbeaVTPLqKc31B31VV1qeMm0gcTUxZKvVxob19XeLtaxjQH9SB1kIMUkUX47FoetnXywaJ/LtRr/2te7kj1q50+VnHiBCrPntW8Vnzjhu43NdOvyC3UEZd79UBFx1xUBodBcHRCZddeOvdVNZPLEQB88c+/45f338Slo39qvZaeng5VgxjqLlGbPpgJKBQKqBoklk4dPYQdEyfgYs/mV8Yr/vlnHIkajvNdvFBaqnt6we933of4MdOhsHcwXsAG9JiSSCSo6JCJaucS3KybWniwqEFi6rYKqJrqqhaPuWL2DADAsdA+UNlJ8Z+6ZBtZP67KR0St5evrq3N1d3t7e3h66p76bg4rsVcplPgrrXZF45E9zb8YAQDuqovzdHYxbpVXixwNWQompiyUuurI3OcZNzSyLst/9GoBqvRcOpzIFrS05LEgCHj99dfh7+8PZ2dn3H333TjbIOljKtcbTLmqlAKbuzqg0Fl3IrwS1SgPDUdOUDBqbt3CtRmPI/2RqfWfoZUxVEvtUdQBLSZDvjqcrtfxMs+ZbjXBerWfNsPHD9NSLmNfTm0J/q95RfglT7vfQmVpKZa9+y527tql2XYu+S+UuLsjOXIwbpU1XV0EALc61S580dRMwBP9opDn5onz3fsBABTK+h3LQ/sY9rFaoOsbZ0EQoHBsvsm5WmVpCZRN9PkgAlgxRUStFxUV1Wh19927d2PIkCFwcDDilzdGlnTtFuQ1Kvi4ydDDu6PY4ejFx80JvXw6QhCAP6+wIIH0w8SUBRIEAQfrElN3WUjmHAB6+XSEt6sM8hoVkq6xGR6RWktLHr/33ntYvnw5PvvsMxw7dgy+vr4YN24cSkvbr1/bR2EyvB/uhEXjInW+fsL+KlROHVAU3B3KJno1tMaXd96HL8dEocyx+crQU9nayY9qI5fb53Z0x4GeA1DVROWRUNW40ue1Z2qnrT16/jqSDuzFnLPpePbsNZTU1Cfm//vPmahpZhqevBWJGpWycS8voS5pdPlmfaJL5eTS7HH0mVbXVr988JbmzxXF2n+HhdezTH5+sixclY+I1MrKypCSkoKUlBQAQFpaGlJSUjQLYd2+WvvcuXNx7do1xMbGIjU1FevWrcPatWvx4osvihG+3tTT4e7s4dXmFXzb0109agsS2MKF9MXElAW6VlCB7KJKOEglGBbaWexw9CaRSDSJtIMcpIg0mlvyWBAExMfH45VXXsHDDz+MiIgIfPXVV6ioqMB3331n9FhqFApU6ajSSfSyBwCUODnqfh9aV8GgVDRf4q1OqOS53T7W6bgxbXCzWp1+TefxVFUNV9fT/wLv+3534px/KP7s0V/n68oGSRVBVXvc/E71Mf/+3VeaP1c2WD65pm61P30avutLaG5eowkkJCRg8+bNmmSBIWfP8/S9bUv938mATndjYsBs2Esa/JuzoItyMj5WTBGR2vHjxzFo0CAMGlTbBzM2NhaDBg3Ca6+9BqDxau2hoaHYvn079u3bh4EDB+LNN9/EJ598gkceeUSU+PWlTuzcZUGzZID6HsgHL+XzywTSCxNTFkidOR/ctRNcHO1FjsYwIzWJKa7SQKSPtLQ05Obmai1xLJPJMHr0aBw+fNio56pUqjBoxyE8uGETSgtanzyuqNZvqu7Z/b9j34Y1rT5PQ3aq2nOqJBIo7G8bFxtcDxVv2dKm89xycW3ilQYJE1Xj5Mk1N939Kwz1+++/Y8OGDXo3OzeWqitFEBRKCCoBykK51mt//vknUlNTkZaZhUoH3YnLpnz1t+c0f76RdlkrqdXbYxjcHD3R021wW0InK8LEFBGp3X333RAEodFj/fr1AHSv1j569GicOHECcrkcaWlpmDt3bvsHboBb5dU4c732iy9LS0wNC+0MR6kdsosqkZav/+rBZLuYmLJAf162zMw5UN8T6+z1EhSUyVvYm4jUjTp1LXF8exPPhuRyOUpKSrQeLdl+swg3XTvhVPgQXD1xTO/V4W63e+1PjbZJdBxr54qPWnV8XcZknQAA/D52LH6a+gguZdSv1KYs0p46LIcCR+0vIa+45eSbIAgQahpPjWtOTVFHON+5QGtbwqjJBh2jKQcPHsTVq1dx6dIlqNrxG8j8L06j4NvzKPwuFcWrLujc57lKB3w1YhJKG0wRLM7LgEqZp9c5dnz2oc7t/TuP1vyZ37raNiamiMiWJF4tgCDUtUNxs4zFrtRcHO0xONgDAFfnI/0wMWVhlCoBh68UALCs/lJq3q5O6O1bW3HwZ93nIKKW6VriuLleA0uXLoW7u7vmERQU1OI5jHWr13/71216f2tSD4NvXgIAFHrVVib9nlS/CmDF0aMAADuJFC7OvjjicBGn7TOw/o/NLR436x/zcOmukQbFosjzhH2XcAjStjdTFSCg+H+/oeinrSjMqf/GMSkpCV/3i4DC0bAKJUPcXvhVdb4QlWeaHrdz6y4prnn5abad3fstBDv9E3stTdQTJFK9j0XWh6vyEZEt0fQUruvXZGnUC1+xhQvpg4kpC3MmuxjFlQq4OtmjX4C72OG0irrSK5GrNBC1yNe3tgePriWOb6+iaiguLg7FxcWaR2Ybm4G3mCzSo/WPYGeHag8vqKTiTEEe4/so+kc8jwJJgwbgyttafDd4UmMnRem+fVAWFWkd51hnfZMjbe+HJCgUuP7ii8hZtAgb3ziq2X7p0iWDjhPmfofB577ka5wehi02WZe0fCmS2K0vEsKHQGVnWdPXybhYMUVEtkQzS6ancdoBtDf1Pd+RqwVQtnP/S7I8TExZGHUp5IjunrCXWuZf3/ButYPrkauFIkdCZP5CQ0Ph6+urtcRxdXU19u/fjxEjRjT5PplMBjc3N61HS6obNAk1NgESVPmGQO4XgsquvZrYx3TnBgAvp0BkO0tw1ttHc67Swiqd5y3u4Ir7P1qLN2a/0Oi1f9yhnWipUhp2k6w0ZDqajl5SAoAi545acQs6quca9ntyc6i9OHRUNd9sXgxlvQe3mKw8GdQTV7wDke/aVI8v27FixQqEhobCyckJkZGROHjwYLP779+/H5GRkXByckK3bt2watWqRvts2bIFffr0gUwmQ58+fbB161aDz1tWVobnnnsOgYGBcHZ2Rnh4OFauXNm2D3sbrspHRLYis7ACGYUVsLeTYFioZSam+vq7oaPMHqVVNUjNabmlBNk2y8xs2LAjV2unUYzobnnT+NTuCO0MiQRIyy9HbnHjJdaJbE1zSx5LJBLMnz8f77zzDrZu3YozZ85g5syZcHFxwYwZM4wax83PPjfq8RoSANS4eQCor6BRSeywdfwMHB04Csl9huKzmYtww9Ov6YNo0VGN1MTNanmH+mNOHtURv/Tth/O+wU0HCmD3sLsgd5Rhf+TwFiNZm924+jNfUtJkou2l1GtIP3kCNQpFi8duSGVXDRWAb4dFY9PQe3E8pDcAQN5EUudGo5UMATeFeV4YKjvWVgC3lG5Q2dn2qnybN2/G/Pnz8corryA5ORkjR47ExIkTtVaeaigtLQ2TJk3CyJEjkZycjEWLFuH555/HlgaLACQmJmL69OmIiYnByZMnERMTg2nTpuHo0foKPX3Ou2DBAuzcuRPffPMNUlNTsWDBAvzf//0ffvnlF6N9flZMEZGtSKxreTIwyAMdZJZZLWwvtcMdIZ0A1N/DEjWFiSkLolCqkHSttonvsG7GmWIhBndnB/T1r63eOJrGQYqopSWPFy5ciPnz52PevHkYMmQIsrOzsXv3brgauXrkYliY1nNT1iQonVxwOaQ3Lof2wYHh0dgz6kFUOblgx5iHAQCdqoyXQMnrMqjRthz3pr99LHMNxA3vSL2Pf0OunWBSQYWfZcdQY6f7J5hQVI5pp9Lw5bcb9D4HACgcinHOPxRldYm9pODeOBHUE1/edb9mmxgUdlIo7Nj7qT0sX74cTz/9NGbPno3w8HDEx8cjKCioycqkVatWoWvXroiPj0d4eDhmz56NWbNm4YMPPtDsEx8fj3HjxiEuLg69e/dGXFwcxo4di/j4eIPOm5iYiKeeegp33303QkJC8Mwzz2DAgAE4fvy40T4/E1NEZCuO1N0jWfI9H9Bwpgzv+ah5TExZkNPZxaioVsLDxQG9vC17OsPwUE7nI1JracljiUSC119/HTk5OaiqqsL+/fsRERFh9DiKOnVq8jWVkdNU5aF9UO0oa/L1xy7s0ftYqpobRsuiKcq3Q6kqRolrgN7vuX31wkq0PF0uI7A73vLt0/KxGxz6zzBPHOo5QOv1v7r11S/INrpu1/RYvXbkA1g78gGdr10ODsPlLvr/LKlp1dXVSEpKQnR0tNb26OhoHD58WOd7EhMTG+0/fvx4HD9+HIq6ir2m9lEfU9/z3nXXXdi2bRuys7MhCAL27t2LixcvYvz48a37wDowMUVEtuJo3T2SpU7jU1Mnpv5KK2SfKWoWE1MWRD1ADQ3pDDsLn86gHqSOMntOZDYue9cnEE5s155+UyGRG+08B3oOwIaoCah07qjzdbuaKgzNTdXa1vBSpqJrGBSS+i3VZT+2cEb9L4RUijQoyraiRn5a7/fc3u+mSqLfFD25zFmv/Q5HRWH3vWPxZ7hhU7jbfPnX4ABX7fJQikqDD7F1Ygz29LkDFQ5NJyENJbltZb5mFqe0Kvn5+VAqlY0WPfDx8Wm0OIJabm6uzv1ramqQn5/f7D7qY+p73k8++QR9+vRBYGAgHB0dMWHCBKxYsQJ33XWXztjkcjlKSkq0Hi3hqnxEZAuyblUgu6gSUjsJIoOb/tLQEqj7TJWwzxS1gIkpC3JUU9Jp2ZlzoL7P1NX8ctwoYZ8pIrFdLK9Chqev5nnh9Swo8w1JHOufBjnnH4pKRyecDms8xQ4A7MtvwW5sLDI7aS+PrIIEx4PDkOUfgqyODZI/Qm3S7JZ7D80mmVCG2/3ppd90M0FVDAil9edtIfFRsjuh1evvCVJ73HLuiIO9Bja5T3pICPK7eLfyDK3XsKH6eftsbHY6DBValxCQ2zs0+doFnyCUO3XQLyYAHUK1Ex221gdbclsmThCERtta2v/27focs6V9PvnkExw5cgTbtm1DUlISPvzwQ8ybNw979uiufly6dCnc3d01j6CgoCY/gxorpojIFqiLEfoFuFtsfyk19pkifTExZSGUKgHH0+v6S4Va9lxjQLvPFAcpIvFlVzWefpZWWNSmY7aU0BGaSVZMu68Hfut/p9a2835dcTwkHL8MHIkaSeNsRHL/5zV/DnNpvArhC5G6+zA1DFPX6nbngsIabWtInq/d/Dyto32j6WslHXSviljeayD29h7c5LE/rxLwxagHcc4vpNkYGjoWEo7rzfTQUsv2aL4Cq0bHMYqlKpQbuZ3U3t6R+OKhZ1Du3EGvBF+OrNy4AVgILy8vSKXSRtVReXl5jaqZ1Hx9fXXub29vD09Pz2b3UR9Tn/NWVlZi0aJFWL58OR544AH0798fzz33HKZPn67Vz6qhuLg4FBcXax6ZmZkt/gy4Kh8R2YK/0uqm8Vl4fyk1rshO+mBiykKcu16CMnkNXJ3sEe7X8rLvloB9pojMW7KOVd0MkdWp+bFK0NEw+6aXH3boWAlPAuB0QPc2xdNa2b4Dmn39VJ8IbBh+Ly741FZ8zB3hjT197oCimcRbQ4omVtUDgI1utTf+f/bsr2e0QFJoOLYNHInbVy6U2TlD0mDbdffmE1MqmZPWc6VEgnH3dsLoe03T43DX6If0qru72dEdB3vo//OwFo6OjoiMjERCQoLW9oSEBIwYMULne6Kiohrtv3v3bgwZMgQODg7N7qM+pj7nVSgUUCgUmsSRmlQqbbK6SSaTwc3NTevRElZMEZEtUM+SGW7h/aXU6vtMFbDPFDXJsmsDbYh6gLojpDOkFt5fSm1YN0+sOZTGPlNEZijLN7jV75U7uuFc+EzYFTe4kdVR4dSUzC6Nqz8KOrjh1m1VR2VS4Ky7FKr02jFRaOU0s5Yo5UkAHmry9bQu/gCAPHdPhN1ouepDH5UOjnCs0a9Xlb7C3IfBSdr61fuqHByNGE1jNz19EZKn3ccqxUOKLUHaCb4tkWO0nh8qKkNySQUGuYm3MmF7iY2NRUxMDIYMGYKoqCisXr0aGRkZmDt3LoDaKqTs7Gxs2FC74uPcuXPx2WefITY2FnPmzEFiYiLWrl2LjRs3ao75wgsvYNSoUVi2bBkmT56MX375BXv27MGhQ4f0Pq+bmxtGjx6Nf//733B2dkZwcDD279+PDRs2YPny5Ub7/ExMEZG1u1FShfSCCthJgMgQy+4vpXZ7n6mIAHexQyIzxMSUhTiiWZnBOko6gdom7g37TPm4ObX8JiJqFxunzNF6bsj3W6cjnkGJWyhqnIqNFk9SSO9G26aM6oAiRzuMUt6F8PyTKOzyl+a1ckn91ESFKsdsy4MLXVyhlGhHV9DBDT8Mucfo55JIJLCTmO9/+zU6KsdmD9Mv2TQt5TIujbL+Kqrp06ejoKAAS5YsQU5ODiIiIrB9+3YEB9cmknNycpCRkaHZPzQ0FNu3b8eCBQvw+eefw9/fH5988gkeeeQRzT4jRozApk2b8Oqrr2Lx4sXo3r07Nm/ejGHDhul9XgDYtGkT4uLi8Pjjj6OwsBDBwcF4++23NckrY2BiioisnbrFSR9/N7g56Vd5be7spXYYEtIJ+y7cxJGrBUxMkU7me4VKGiqVgGPp6rnG1lHSCQDuLg7o4+eGs9dLcDStEA8O8Bc7JCJqSjPNlQVoJ65K3EJNHk6uqxuKHGtvUs/2GojiQ8ehktYno07ap2v+XKPMQnO1Pg1jl7TQuybf1cPwYJvx/R1jG20779vVqOdoSo3UyM2i2qjCpSOA0hb20v3vsFRpO4mKefPmYd68eTpfW79+faNto0ePxokTJ5o95tSpUzF16tRWnxeo7VX15ZdfNnuMtuKqfERk7Y6q+0tZyTQ+teHdPLHvwk0cTSvE7JHdxA6HzJC5folMDZzPLUVxpQIdHKWI8LeO/lJqd4TUVoAlpbPPFJGlUuoxhe72qiAAjabmGeKcb/slslV69ooyN4KOHE5uB5nmzyeDerZbLMaagK6rOT3ZDlZMEZG1U7c4saZZMgA0K/OduHaLC1iQTkxMWQB1f6nIkM6wl1rXX9mQukHq+LVbIkdCRM0plzk382rLFxgJYSFGi+V2jnZO8O/cr5Xvbjn28p7NNz+3FNc7OqPKQf8qKcFo6ST96UqmNfTzoFHtEwiZJa7KR0TW7GapHFdulkMiAYZaWWIqIsAdjvZ2KCivRlq+ba6uS82zriyHldJM47OyAQoAhgTXfqbUnNpVB4nIeqhMmNewt6ufie5s3xE9A8Y0s3fzbKEGp8hRgjM+5v9/SGqX+r4Tb/WVNbMn2SJWTBGRNTted88X5uMKDxfTLjjS3mT2UgwIrP0//ng6CxKoMYtLTK1YsQKhoaFwcnJCZGQkDh482OS++/btg0QiafQ4f/681n5btmxBnz59IJPJ0KdPH2zdutXUH0NvgiBofnmHBFvHygwN+bo7IbCTM1QCkJzBQYrImiT3CtLeoGM6X2vZS+ov2PJkuo+bHNQDv/WLgtKumfPelpWyrqli9Z9lZQ/HZvuE6XW0dihSyXGtb3b+c6B1XZRT2zExRUTWTD2DZIiVrMZ3uyF1LVyOX2MLF2rMohJTmzdvxvz58/HKK68gOTkZI0eOxMSJE7VWoNHlwoULyMnJ0Tx69qzvq5GYmIjp06cjJiYGJ0+eRExMDKZNm4ajR4+a+uPoJetWJfJK5bC3k2BAkIfY4ZiEOuHG7DmR9ajokImrXU3XuFNpV59kKXOo/fNfIeE4FdBds/1otwhkdvbBheCwRu+/6NsVx4PDIEhartS84tW+CzOcDuxh9GMq7SR6dALTtrvvUJ29wVqjpSl6RPpgYoqIrFmSOjEVbP4Vzq2huedjCxfSwaISU8uXL8fTTz+N2bNnIzw8HPHx8QgKCsLKlSubfZ+3tzd8fX01D2mDlYji4+Mxbtw4xMXFoXfv3oiLi8PYsWMRHx9v4k+jnxN1VUR9A9zhZEBvEEvC7DmR5VM0SBS5y8pQ7pqGGpnpbh4VEu1jZ7hIcCI4DId7NO41VSPVvQDt8ZBwVDk2/m9weJcH0LVDH83zhL5D2xitOOS3NW13lDbXJ0zH+x0c8VdouFFi2XzHvUY5Dtk2rspHRNaqSqHE2evFAIBIK5wlA9R/rqs3y1FQJhc5GjI3FpOYqq6uRlJSEqKjo7W2R0dH4/Dhw82+d9CgQfDz88PYsWOxd+9erdcSExMbHXP8+PEtHrO9qKuIIrta5wAF1JerJmcUocaGlvwmshYHfZwQd/cw5LrV/i736GD63+MCB+1Kpwpp0yU5UuhOTAGAXKqdvHGSuqBrx3A4yNybeIfl2B82SOv5mQDDK79OBvVEjRGnYBK1BSumiMhancoqhkIpoIurDIGdDPsiyVJ4uDiip3dHAPXVYURqFnO1mZ+fD6VSCR8fH63tPj4+yM3N1fkePz8/rF69Glu2bMFPP/2EsLAwjB07FgcOHNDsk5uba9AxAUAul6OkpETrYSpJVj7XGAB6ebvC1ckeFdVKnM8tFTscImqlo6F92+1c5/1C9N7Xr0P3Jl/7bng0yrRWHJTgtP11nLXPan1wZuovr6YTdM1J69K+UxmJmsJV+YjIWtVP4+ukqQ61Rup7Wiam6Hatu0oV0e2/qIIgNPnLGxYWhrCw+t4iUVFRyMzMxAcffIBRo+qXnDbkmACwdOlSvPHGG60J3yBl8hqcz61NellrSScA2NlJEBncCfsu3MSx9EJEBFh+pQIRmRc7ND0V+oJvV82f105+CqUd3TAo42J7hGURtBN3ROJhxRQRWaukupYm1nzPB9T2z9r4V6Zm1XkiNYupmPLy8oJUKm1UyZSXl9eo4qk5w4cPx6VLlzTPfX19DT5mXFwciouLNY/MzEy9z2+IlIwiqAQgsJMzfNycTHIOc8FmeERkKl/39kfKsMl67XvLvRNqpFLkuJuucTsRtQ4TU0RkjQRB0FQQWX1iqq5i6nR2MaoUSpGjIXNiMYkpR0dHREZGIiEhQWt7QkICRowYofdxkpOT4efnp3keFRXV6Ji7d+9u9pgymQxubm5aD1OwlQEKaNAAPb2QJfpEIjDW793OvsOwsZsHAEBy2yGzPbxgqt/uYofmy953DRps0PEqHRzbEg4RmQATU0RkjdLyy3GrQgFHezv09bfumSNdO7ugi6sMCqWAU1nFYodDZsSipvLFxsYiJiYGQ4YMQVRUFFavXo2MjAzMnTsXQG0lU3Z2NjZs2ACgdsW9kJAQ9O3bF9XV1fjmm2+wZcsWbNmyRXPMF154AaNGjcKyZcswefJk/PLLL9izZw8OHTokymds6LiNlHQCwIBAD9jbSXCjRI6sW5UI6uwidkhENqWmXNHmY+R4eAEA0r2AuTpaNP064C7ceelUm8+jyz/vqB8z5FIHlMvaVmVa4OrRxoiIyNi4Kh8RWSP1jJEBge5wtLeYupFWkUgkGBLcCTvO5OL4tUIMDe0sdkhkJiwqMTV9+nQUFBRgyZIlyMnJQUREBLZv347g4GAAQE5ODjIyMjT7V1dX48UXX0R2djacnZ3Rt29f/Pbbb5g0aZJmnxEjRmDTpk149dVXsXjxYnTv3h2bN2/GsGHD2v3zNaRUCUjJKAJgG4kpZ0cp+ga442RmEY5fK2RiiqidZfx1A/Ay7jEFHb36LvkEGvckOnx5130mP0dTst2N/EM0A2lefhiUeanlHYlMjBVTRGSNTmhmydhGkiZSnZhKZwsXqmdRiSkAmDdvHubNm6fztfXr12s9X7hwIRYuXNjiMadOnYqpU6caIzyjuXijFKXyGnRwlCLMx1XscNrFkOBOOJlZhOSMIjw0yPQ3r0RUT1muMHpiyhanw/068C6xQzC6PDfbuFAm88dV+YjIGh23ofYtQH0Ll+SMWy0uOka2w7prBS2Yur/UwK4esJfaxl/ToK4eAIDkukoxIiIyDyeCeqLa3kHsMMjGsWKKiKxNUUU1LueVAQAG190LWbs+fm5wtLfDrQoFrhVUiB0OmQnbyHhYIFsr6QSAQV1rvyVIzSlBZTVXaSCi9pHf0bobjRrDX936Yl/YILHDIBvHxBQRWRv1F/LdvDrAs6NM3GDaiaO9HSL8axcPS87kdD6qxcSUmbK1kk4A8Hd3grerDDUqAWeuc5UGIkt23rerzh5T5pgE4lQ1/eS6e4odAtk4JqaIyNqoF7sabEP3fEB9QQJnypAaE1Nm6GapHBmFFZBIgIFBHmKH024kEonm86ZwkCKyaPvCButM+KjspCJEQ0TWgKvyEZG1SbLBYgSg/h43JbNI1DjIfDAxZYbUv6A9vTvC3dm2enposucs6yQiIqIGWDFFRNZEqRJwOqt2lsjgrraVmFL3Fj53vQRVCrZwISamzFJKXVJmQKCHuIGIgA3QiYiISBeuykdE1uRSXinKq5Xo4ChFD++OYofTrgI8nNFF3cIlmy1ciIkps3Qys/aXc6CNrMzQUP9Ad9hJgJziKuQUV4odDhEREZkJVkwRkTU5WTdLpl+gO6R2jftyWjOJRIJBddP5WJBAABNTZkelEjSDlC31l1JzcbRHb9/aVRrYZ4qo/aiULKMmIvPGxBQRWZMUzT2fbU3jU2MLF2qIiSkzczW/DKXyGjg52CHMx1XscEShrhRjMzyi9lNYWCp2CEREzWJiioisibpSaGCQ+a1Y3B646BU1xMSUmUmpm8bXL8Ad9lLb/OthWSdR+yuqLhM7BCKiZnFVPiKyFhXVNbh4o/ZLQVutmFK3cLleXIXc4iqxwyGR2Wbmw4ypG5/b4jQ+NXVZ56nsIiiUvPgkag8lTjfEDoGIqFmsmCKihlasWIHQ0FA4OTkhMjISBw8ebHLfffv2QSKRNHqcP3++HSOudzqrGCoB8HVzgq+7kygxiK2DzB5h6hYunM5n85iYMjO2PtcYALp5dYCbkz2qFCpcyOX0IqL2wVWuiMi8cVU+IlLbvHkz5s+fj1deeQXJyckYOXIkJk6ciIyMjGbfd+HCBeTk5GgePXv2bKeItaXYcE/hhrgiO6kxMWVGqhRKnM+pTcQMsNG5xgBgZyfBAM10PmbPiYiIiBVTRFRv+fLlePrppzF79myEh4cjPj4eQUFBWLlyZbPv8/b2hq+vr+YhlUrbKWJtJ7OKAEBzz2OrBrKFC9VhYsqMnL1ejBqVAK+OMgR4OIsdjqg0qzRwkCJqF7a1SDERWSImpogIAKqrq5GUlITo6Git7dHR0Th8+HCz7x00aBD8/PwwduxY7N27t9l95XI5SkpKtB7GkqJpfO5htGNaosF1FVNs4UJMTJmR5AYDlLrBp60axJX5iNqVVKUUOwQiomYxMUVEAJCfnw+lUgkfHx+t7T4+PsjNzdX5Hj8/P6xevRpbtmzBTz/9hLCwMIwdOxYHDhxo8jxLly6Fu7u75hEUFGSU+PNKqnC9uAp2ktoG4Lasm1dHuLKFC4GJKbNyMqt2RT5bXTK0oYGBHgCAq/nluFVeLW4wRCJ7/fXXGzXr9PX1Neo5bDsVTkSWgKvyEVFDt3+RLwhCk1/uh4WFYc6cORg8eDCioqKwYsUK3Hffffjggw+aPH5cXByKi4s1j8zMTKPErf7ivae3KzrI7I1yTEtlZydpMJ2PLVxsGRNTZqR+RT7bbXyu1qmDI0I8XQAAp7KLRY6GSHx9+/bVatZ5+vRpsUMiImpXrJgiIgDw8vKCVCptVB2Vl5fXqIqqOcOHD8elS5eafF0mk8HNzU3rYQxsfK5N/XNQF2mQbWJiykwUlMmRWVgJAOjPiikAQP+6qqlTnM5HBHt7e61mnV26dBE7JCKidsVV+YgIABwdHREZGYmEhASt7QkJCRgxYoTex0lOToafn5+xw2uRJjFV17rE1mnu+eoawpNtsu3aQTOiXpmhe5cOcHNyEDcYM9E/0B3bTl5n9pwIwKVLl+Dv7w+ZTIZhw4bhnXfeQbdu3ZrcXy6XQy6Xa54bs2EnEZEYWDFFRGqxsbGIiYnBkCFDEBUVhdWrVyMjIwNz584FUDsNLzs7Gxs2bAAAxMfHIyQkBH379kV1dTW++eYbbNmyBVu2bGnXuFUqAafq7m0G1CVkbN2Auj5bl/PKUC6vsfnpjbaKf+tmon5lBk7jU1Mvn8rsOdm6YcOGYcOGDejVqxdu3LiBt956CyNGjMDZs2fh6emp8z1Lly7FG2+8YcBZWIFAROaNiSkiUps+fToKCgqwZMkS5OTkICIiAtu3b0dwcDAAICcnBxkZGZr9q6ur8eKLLyI7OxvOzs7o27cvfvvtN0yaNKld475yswxl8ho4O0jRy6dju57bXHm7OcHXzQm5JVU4k12MYd10X9uSdWNiykwks6Szkb7+brCTAHmlctwoqYKPm5PYIRGJYuLEiZo/9+vXD1FRUejevTu++uorxMbG6nxPXFyc1mslJSXNriZT6cgbPSIyb0xMEVFD8+bNw7x583S+tn79eq3nCxcuxMKFC9shquap7/n6BbrDXsquOmr9A92Re64Kp5mYsln8bTADgiDgpDoxxZJODRdHe/TycQUAzc+HiIAOHTqgX79+Rm3YKXBZPiIyc1yVj4gs3Uk2PtdpABug2zwmpsxAWn45SqpqILO3Q28/V7HDMSv96+Ycn+IgRaQhl8uRmppq1Iad1VL2tiMi88aKKSKydFyRT7f6e74icQMh0TAxZQbUjc/7+rvBgSWdWtSrNJzkIEU27MUXX8T+/fuRlpaGo0ePYurUqSgpKcFTTz0ldmhERO2Gq/IRkSWrUihxPrcUQH2FENXqH+ABALhWUIGiimpxgyFRWFwWZMWKFQgNDYWTkxMiIyNx8ODBJvf96aefMG7cOHTp0gVubm6IiorCrl27tPZZv349JBJJo0dVVZWpP4qGZmUGDlCNqFerOJ1dzAtRsllZWVl47LHHEBYWhocffhiOjo44cuSIpsEnEZEtYMUUEVmyczklUKoEeHWUwd+dvXMbcndxQIinCwDOlLFVFpWY2rx5M+bPn49XXnkFycnJGDlyJCZOnKi14kJDBw4cwLhx47B9+3YkJSVhzJgxeOCBB5CcnKy1n5ubG3JycrQeTk7tN1icrvvl6xfg3m7ntBRhvq5wlNqhqEKBjMIKscMhEsWmTZtw/fp1VFdXIzs7G1u2bEGfPn3EDouIqF0xMUVElqz+ns9N0zOP6qlnynA6n22yqMTU8uXL8fTTT2P27NkIDw9HfHw8goKCsHLlSp37x8fHY+HChbjjjjvQs2dPvPPOO+jZsyd+/fVXrf0kEgl8fX21Hu1FqRJw9noJgPq5tVTP0d4O4f61TZvZDI+IiMh2MTFFRJbsdHZdYoqLXemkvhfmPZ9tspjEVHV1NZKSkhAdHa21PTo6GocPH9brGCqVCqWlpejcubPW9rKyMgQHByMwMBD3339/o4oqU7pyswyVCiU6OEoR6tWx3c5rSQaom+FxZT4ikznrHyp2CEREzeKqfERkydQVU/05S0YndVsbVkzZJotJTOXn50OpVMLHx0dru4+PD3Jzc/U6xocffojy8nJMmzZNs613795Yv349tm3bho0bN8LJyQl33nlns8uwy+VylJSUaD1aSz2Htm+AO6R2LOnUpb6sk9lzIlPJ8fASOwQiomaxYoqILFVFdQ0u5dU2Pu/HWTI69fV3g50EuFEix42S9uv3TObBYhJTarfPxxUEQa85uhs3bsTrr7+OzZs3w9vbW7N9+PDheOKJJzBgwACMHDkS33//PXr16oVPP/20yWMtXboU7u7umkdQUFCrP8/puoww+0s1TV0xdeZ6MZQqNkAnIiKyRVyVj4gs1bnrJVAJgLerDD5ubHyui4ujPXr5uAIATnKmjM1pVWKqqKgIa9asQVxcHAoLCwEAJ06cQHZ2tlGDa8jLywtSqbRRdVReXl6jKqrbbd68GU8//TS+//573Hvvvc3ua2dnhzvuuKPZiqm4uDgUFxdrHpmZmfp/kNuo5xqzv1TTunXpiA6OUlRUK3E5r0zscMiKiTG2ERGRflgxRWR+eO2kH97z6Uf98+FMGdtjcGLq1KlT6NWrF5YtW4YPPvgARUVFAICtW7ciLi7O2PFpODo6IjIyEgkJCVrbExISMGLEiCbft3HjRsycORPfffcd7rvvvhbPIwgCUlJS4Ofn1+Q+MpkMbm5uWo/WqFGqNI3PWTHVNKmdBBEB6mZ4ReIGQ1ZLrLGNiMgQK1asQGhoKJycnBAZGYmDBw82u//+/fsRGRkJJycndOvWDatWrWq0j3qlT5lMhj59+mDr1q2tOm9qaioefPBBuLu7w9XVFcOHD29y5eTWYGKKyLzw2kl/9SvyeYgbiJlTt3DhPZ/tMTgxFRsbi5kzZ+LSpUtwcqovQ5w4cSIOHDhg1OB0nXvNmjVYt24dUlNTsWDBAmRkZGDu3LkAaiuZnnzySc3+GzduxJNPPokPP/wQw4cPR25uLnJzc1FcXJ+BfeONN7Br1y5cvXoVKSkpePrpp5GSkqI5pildyiuDvEaFjjJ7hHh2MPn5LFl99rxI3EDIaok5thER6WPz5s2YP38+XnnlFSQnJ2PkyJGYOHFik8mftLQ0TJo0CSNHjkRycjIWLVqE559/Hlu2bNHsk5iYiOnTpyMmJgYnT55ETEwMpk2bhqNHjxp03itXruCuu+5C7969sW/fPpw8eRKLFy/WGk/bis3PicwLr530d0qzIl/rChpshfqe73R2Madt2xiDE1PHjh3Ds88+22h7QECA3k3IW2v69OmIj4/HkiVLMHDgQBw4cADbt29HcHAwACAnJ0frIum///0vampq8M9//hN+fn6axwsvvKDZp6ioCM888wzCw8MRHR2N7OxsHDhwAEOHDjXpZwHqSzojAtxgx8bnzWIDdDI1Mcc2IiJ9LF++HE8//TRmz56N8PBwxMfHIygoCCtXrtS5/6pVq9C1a1fEx8cjPDwcs2fPxqxZs/DBBx9o9omPj8e4ceMQFxeH3r17Iy4uDmPHjkV8fLxB533llVcwadIkvPfeexg0aBC6deuG++67T6uvZ1upK6YA9pkiMge8dtJPubwGV27WtiOJ4CyZZvX2dYOj1A5FFQpkFFaIHQ61I4MTU05OTjpXobtw4QK6dOlilKCaM2/ePKSnp0MulyMpKQmjRo3SvLZ+/Xrs27dP83zfvn0QBKHRY/369Zp9PvroI1y7dg1yuRx5eXnYtWsXoqKiTP45gAZLhtYlXahpA+p+Rqk5JZDXKMUNhqyS2GMbEVFzqqurkZSUhOjoaK3t0dHROHz4sM73JCYmNtp//PjxOH78OBQKRbP7qI+pz3lVKhV+++039OrVC+PHj4e3tzeGDRuGn3/+udWfV5eGiSlWTRGJj9dO+jl7vQSCAPi5O8HblY3Pm+Nob4dwv9oG6CxIsC0GJ6YmT56MJUuWaC5oJBIJMjIy8PLLL+ORRx4xeoDWTFPSycx5i4I6O6OTiwMUSgEXckvFDoesEMc2IjJn+fn5UCqVjRZ88fHxabIyITc3V+f+NTU1yM/Pb3Yf9TH1OW9eXh7Kysrw7rvvYsKECdi9ezceeughPPzww9i/f7/O2ORyOUpKSrQeLWHFFJF54bWTfk5xFXaD1M+UKRI1DmpfBiemPvjgA9y8eRPe3t6orKzE6NGj0aNHD7i6uuLtt982RYxWSaFUITWHjc/1JZFI0E/TDI/ZczI+jm1EZCo1NTXYs2cP/vvf/6K0tPbLlevXr6OszPCVZtV9ltQEQWi0raX9b9+uzzGb20ddvTR58mQsWLAAAwcOxMsvv4z7779fZ7N1AFi6dCnc3d01j6CgoCY/gxorpojMC6+d9HOaxQgGUfeZ4j2fbbE39A1ubm44dOgQ/vjjD5w4cQIqlQqDBw/Gvffea4r4rNbFG6WorlHB1ckewZ4uYodjEQYEuuPAxZs4lVkEDA8WOxyyMhzbiMgUrl27hgkTJiAjIwNyuRzjxo2Dq6sr3nvvPVRVVTWZuLmdl5cXpFJpo+qovLy8RtVMar6+vjr3t7e3h6enZ7P7qI+pz3m9vLxgb2+PPn36aO0THh6OQ4cO6YwtLi4OsbGxmuclJSUtJqeYmCIyL7x20o8mMRXIxJQ+BgR5AADOZBdDqRIgZS9mm2BwYkrtnnvuwT333GPMWGxKfX8p92a/6aR66m8Z1IM7kSlwbCMiY3rhhRcwZMgQnDx5UpMMAoCHHnoIs2fP1vs4jo6OiIyMREJCAh566CHN9oSEBEyePFnne6KiovDrr79qbdu9ezeGDBkCBwcHzT4JCQlYsGCB1j4jRozQ+7yOjo644447cOHCBa1zXbx4UbNAze1kMhlkMpm+Hx+AdtUWE1NE5oPXTk0rrVLg6s1yAKyY0lf3Lh3h4ihFRbUSafll6OHtKnZI1A70Skx98skneh/w+eefb3UwtuSUZkU+DlD6Us83vpRXhiqFEk4OUnEDIovHsY2ITO3QoUP4888/4ejoqLU9ODgY2dnZBh0rNjYWMTExGDJkCKKiorB69WpkZGRg7ty5AGqrkLKzs7FhwwYAwNy5c/HZZ58hNjYWc+bMQWJiItauXYuNGzdqjvnCCy9g1KhRWLZsGSZPnoxffvkFe/bs0ap0aum8APDvf/8b06dPx6hRozBmzBjs3LkTv/76q9aiNG3Fiiki8fHayTBnsmtbtwR4OMOzo2HJeFsltZOgr78bjqXfwqmsYiambIReiamPPvpI6/nNmzdRUVEBDw8PAEBRURFcXFzg7e3NAUhPZ+oSU/0DPMQNxIL4uMng1VGG/DI5zuWUYHDXTmKHRBaOYxsRmZpKpYJS2Xg12aysLLi6GnaxPX36dBQUFGDJkiXIyclBREQEtm/frqlKysnJQUZGhmb/0NBQbN++HQsWLMDnn38Of39/fPLJJ1pNiUeMGIFNmzbh1VdfxeLFi9G9e3ds3rwZw4YN0/u8QG0F2KpVq7B06VI8//zzCAsLw5YtW3DXXXcZ9Bmbw8QUkfh47WSYM+wv1SoRAe44ln4Lp7OL8fDgQLHDoXagV2IqLS1N8+fvvvsOK1aswNq1axEWFgagdknQOXPm4NlnnzVNlFZGXqPUND7vz7nGepNIJOgX4Ia9F27iTHYxE1PUZhzbiMjUxo0bh/j4eKxevRpA7f9lZWVl+M9//oNJkyYZfLx58+Zh3rx5Ol9bv359o22jR4/GiRMnmj3m1KlTMXXq1FafV23WrFmYNWtWs/u0BVflIxIfr50Mc4r9pVpFncg7wxYuNsPgVfkWL16MTz/9VDP4AEBYWBg++ugjvPrqq0YNzlpdzC2DQinAw8UBgZ2cxQ7Homj6THGVBjIyjm1EZAofffQR9u/fjz59+qCqqgozZsxASEgIsrOzsWzZMrHDsyismCIyL7x2atnprCIALEYwlPqe7+z1EihV/CLCFhjc/DwnJwcKhaLRdqVSiRs3bhglKGt3KrsIQO0vHBufGyaCDdDJRDi2EZEp+Pv7IyUlBZs2bUJSUhJUKhWefvppPP7443B25pdThmBiisi88NqpecWVCqQXVAAAIvyZmDJEtwYN0K/eLENPH/aZsnYGV0yNHTsWc+bMwfHjxzVl1MePH8ezzz7LpUH1xLnGrXd7A3QiY+HYRkSm4uzsjL///e/47LPPsGLFCsyePZtJqVbgqnxE5oXXTs07W3fPF9TZGZ06OLawNzWkboAOsCDBVhicmFq3bh0CAgIwdOhQODk5QSaTYdiwYfDz88OaNWtMEaPVOVU3DY0lnYZTN0BXqgScq+vTRWQMHNuIyBSWLl2KdevWNdq+bt06TuUzEBNTROaF107NO8XFrtqEM2Vsi8FT+bp06YLt27fj4sWLOH/+PARBQHh4OHr16mWK+KxOlUKJC7mlAOp/2Uh/bIBOpsKxjYhM4b///S++++67Rtv79u2LRx99FC+99JIIUVkuOzs7qFQqJqaIzACvnZqn7onLe77WYQN022JwYkqtV69eHHRa4XxuKWpUAjp3cESAB8v4W6NfgDv2XrjJBuhkEhzbiMiYcnNz4efn12h7ly5dkJOTI0JElk2dmOKqfETmg9dOuqn7CnOWTOvc3gBdasfezNbM4MRUS8sA6ypXp3qnG/SXYuPz1ulX12eKZZ1kTBzbiMgUgoKC8OeffyI0NFRr+59//gl/f3+RorJc6gborJgiEh+vnZpWVFGNzMJKAGx83lpsgG5bDE5M3bp1S+u5QqHAmTNnUFRUhHvuucdogVkrLhnadursuboBupODVOSIyBpwbCMiU5g9ezbmz58PhUKhGUt+//13LFy4EP/6179Ejs7yMDFFZD547dQ09RfoIZ4ucHdxEDkay6RugH4s/RZOZxczMWXlDE5Mbd26tdE2lUqFefPmoVu3bkYJypqd4lzjNlM3QM8vk+NcTgn7TJFRcGwjIlNYuHAhCgsLMW/ePFRXVwMAnJyc8NJLLyEuLk7k6CyPutqciSki8fHaqWm85zOOiAB3TWLq4cGBYodDJmTwqnw6D2JnhwULFuCjjz4yxuGsVpVCiUt5ZQBYMdUW6gboAJvhkWlxbCOitpJIJFi2bBlu3ryJI0eO4OTJkygsLMRrr70mdmgWiRVTROaN10611PcovOdrGzZAtx2tbn5+uytXrqCmpsZYh7NK53JqG7d5dZTB181J7HAsmroB+ik2QCcT49hGRMbQsWNH3HHHHWKHYfGYmCIyf7x2qq+Y6hfgIW4gFq4+McUG6NbO4MRUbGys1nNBEJCTk4PffvsNTz31lNECs0anNQOUGxuft5G6ATqz52QsHNuIyBTKy8vx7rvv4vfff0deXl6jhMrVq1dFiswyqRNTXJWPSHy8dtKtoEyO7KLaxud962Z5UOuwAbrtMDgxlZycrPXczs4OXbp0wYcfftjiygy2TpM5r0uqUOuxAToZG8c2IjKF2bNnY//+/YiJiYGfnx+/mGojVkwRmQ9eO+mmbnzezasD3JzY+Lwt2ADddhicmNq7d68p4rAJmrnGbILXZmyATsbGsY2ITGHHjh347bffcOedd4odilVgYorIfPDaSTf1PV8/9pcyCjZAtw0GNz+/5557UFRU1Gh7SUmJzS8L2pyK6hpcyisFwEHKGBo2QD/NPlNkBBzbiMgUOnXqhM6dO4sdhtXgqnxE5oPXTrrV95fiPZ8xqH+OvOezbgYnpvbt26dZ7rihqqoqHDx40ChBWaNz10ugEgBvVxl82PjcKDSDFPtMkRFwbCMiU3jzzTfx2muvoaKiQuxQrAIrpojMB6+ddFPfmzAxZRzqn+PZ67UN0Mk66T2V79SpU5o/nzt3Drm5uZrnSqUSO3fuREBAgHGjsyKnuWSo0bEBOhkDxzYiMqUPP/wQV65cgY+PD0JCQuDgoN1v5MSJEyJFZpmYmCISH6+dmnazVI6c4ipIJEBfJqaMgg3QbYPeiamBAwdCIpFAIpHoLM10dnbGp59+atTgdFmxYgXef/995OTkoG/fvoiPj8fIkSOb3H///v2IjY3F2bNn4e/vj4ULF2Lu3Lla+2zZsgWLFy/GlStX0L17d7z99tt46KGHjBr3aS4ZanQNG6BXVivh7MgG6GQ4cxnbiMg6TZkyRewQrApX5SMSn7lcO5nivrCt1F+Yd+/SER1lBrdzJh3YAN026P3bkpaWBkEQ0K1bN/z111/o0qWL5jVHR0d4e3tDKjVtYmDz5s2YP38+VqxYgTvvvBP//e9/MXHiRJw7dw5du3bVGfOkSZMwZ84cfPPNN/jzzz8xb948dOnSBY888ggAIDExEdOnT8ebb76Jhx56CFu3bsW0adNw6NAhDBs2zGixn9I0weOSocZyewP0yGA2QCfDmcPYRkTW6z//+Y/YIVgVVkwRic8crp1McV9oDOwvZRrqBuinstgA3VrpnZgKDg4GIO6FwPLly/H0009j9uzZAID4+Hjs2rULK1euxNKlSxvtv2rVKnTt2hXx8fEAgPDwcBw/fhwffPCBZgCKj4/HuHHjEBcXBwCIi4vD/v37ER8fj40bNxol7jJ5Da7cLANQ+0tFxqFugL73wk2cyS5mYopaxRzGNiIi0g8TU0TiM4drJ1PcFxrD6ewiAExMGZv658kWLtZLr8TUtm3bMHHiRDg4OGDbtm3N7vvggw8aJbDbVVdXIykpCS+//LLW9ujoaBw+fFjnexITExEdHa21bfz48Vi7di0UCgUcHByQmJiIBQsWNNpHPWgZw7nrJRAEwM/dCd6ubHxuTP0CPbD3wk02QDcBQRDwys9nEObjimlDgqxyqqQ5jG2GMLRknYjEp1Qq8dFHH+H7779HRkZGo0bBhYWFIkVmmbgqH5G4zOHayVT3hcbAvsKmof55qhugS+0kIkdkXb45cg0FZdW4r78veniLM1VSr8TUlClTkJubC29v72Z7JUgkEiiVSmPFpiU/Px9KpRI+Pj5a2318fLQa7jWUm5urc/+amhrk5+fDz8+vyX2aOiYAyOVyyOVyzfOSkpJmYz+VVQSAmXNTYPbcdDIKK/Dd0Qw4Su3w2NDGJdHWwBzGNn0ZWrJORObhjTfewJo1axAbG4vFixfjlVdeQXp6On7++We89tprYodncVgxRSQuc7h2MtV94e0Mvee7UVKFGyVy2EmAPv5s32JMoV5sgG5K3x/PxKmsYnT37iBaYspOn51UKhW8vb01f27q0R43bupvytQEQWi0raX9b99u6DGXLl0Kd3d3zSMoKKjZmPsHemDmiBCM7+vb7H5kOHVi6uKNUlRWi5s4sDbqb3zC/VzhaK/XUGFxzGlsa0nDkvXw8HDEx8cjKCgIK1euFDs0ImrGt99+iy+++AIvvvgi7O3t8dhjj2HNmjV47bXXcOTIEbHDszhMTBGJy5yunUxxX9iQofd8dhIJ5t/bE48PC4aLIxufG5O6ATpQ38eLjKO6RoXzOaUAgP4iLtRmMXebXl5ekEqljbLgeXl5jbLfar6+vjr3t7e3h6enZ7P7NHVMoLYPVXFxseaRmZnZbOxDQzvj9Qf74pFINmozNnUDdJUAnMtp/lsMMox6JUn2RROfumT99hL05krW5XI5SkpKtB5E1P5yc3PRr18/AEDHjh1RXFw7tt5///347bffxAzNInFVPiIy1X3h7Qy95+viKsP8e3vhzSkRBnwa0pf6noQtXIzr4o1SVCtVcHd2QFBnZ9Hi0CuV+8knn+h9wOeff77VwTTH0dERkZGRSEhIwEMPPaTZnpCQgMmTJ+t8T1RUFH799Vetbbt378aQIUM084ijoqKQkJCg1Wdq9+7dGDFiRJOxyGQyyGSytnwcMhI2QDcd9aBvzVNQzWFs00drStaXLl2KN954oz3CI6JmBAYGIicnB127dkWPHj2we/duDB48GMeOHeO1RCuwYopIXOZw7WSq+8Lb8Z7PvLCFi2mo7/kiAtyarTg0Nb0SUx999JFeB5NIJCa9eYuNjUVMTAyGDBmCqKgorF69GhkZGZg7dy6A2qx2dnY2NmzYAACYO3cuPvvsM8TGxmLOnDlITEzE2rVrtVbbe+GFFzBq1CgsW7YMkydPxi+//II9e/bg0KFDJvscZFxsgG58giBoBn1rrpgyl7FNX4aUrMfFxSE2NlbzvKSkpMUSdCIyvoceegi///47hg0bhhdeeAGPPfYY1q5di4yMjEaLr1DLmJgiEpe5XDuZ4r6QzBsboJvGaTO559MrMZWWlmbqOPQyffp0FBQUYMmSJcjJyUFERAS2b9+uWbI0JycHGRkZmv1DQ0Oxfft2LFiwAJ9//jn8/f3xySefaC0JOmLECGzatAmvvvoqFi9ejO7du2Pz5s0YNmxYu38+ah119vw05xsbTUZhBUqqauAotUMvK24uaC5jW0taU7LOb/mIzMO7776r+fPUqVMRGBiIw4cPo0ePHmax2qel4ap8ROIyl2snU9wXknlr2AD9ys0yq75HaU9nzGSWTJu6srXUMM4U5s2bh3nz5ul8bf369Y22jR49GidOnGj2mFOnTsXUqVONER6JQP1LdCmvtgG6s6NU5Igsnzpz3tuKG583R4yxrTmtKVknIvM0fPhwDB8+XOwwLBYrpojMk7XcF5L5UjdAP5Z+C6ezipmYMoKGjc/FTky16o5z7dq1iIiIgJOTE5ycnBAREYE1a9YYOzYivbABuvGZS0lnezPnsS02NhZr1qzBunXrkJqaigULFmiVrBOR+di2bRsUCoXmz809yDBMTBGZF3O+diLrwwboxqVufO7mZI+unV1EjcXgiqnFixfjo48+wv/93/8hKioKAJCYmIgFCxYgPT0db731ltGDJGqORCJB/0B3/HE+jw3QjcRcSjrbk7mPbS2VrBOR+ZgyZQpyc3Ph7e2NKVOmNLmfRCJplyXVrQlX5SMyH+Z+7UTWR91nig3QjaNhT2GxZ4oYnJhauXIlvvjiCzz22GOabQ8++CD69++P//u//+MARKKICKhNTJ1in6k2EwRB06/LlhJTljC2NVeyTkTmo2E1Dyt7jIsVU0TmwxKunci6qO9N2ADdOE6pixECxb/nM3gqn1KpxJAhQxptj4yMRE1NjVGCIjIUlw81HltpfH47jm1EZGwKhQJjxozBxYsXxQ7FajAxRWQ+eO1E7U3dAL1SUdsAndrGnGbJGJyYeuKJJ7By5cpG21evXo3HH3/cKEERGer2BujUerba+JxjGxEZm4ODA86cOSN6ebw14ap8ROaD107U3tQN0AGuyN5W5tT4HGjlqnxr167F7t27NavKHDlyBJmZmXjyyScRGxur2W/58uXGiZKoBeoG6PllcpzLKWGfqTaw1cbnAMc2IjK+J598EmvXrsW7774rdihWgRVTROaF107U3iIC3GtX5ssuxiORgWKHY7HMqfE50IrE1JkzZzB48GAAwJUrVwAAXbp0QZcuXXDmzBnNfvx2kNpTwwbop7OKmJhqA3Mq6WxPHNuIyBSqq6uxZs0aJCQkYMiQIejQoYPW67xZMwwTU0Tmg9dOJAZ1A3SuzNc25tT4HGhFYmrv3r2miIOozdQN0E9nl4gdisUSBAFn6n5+tpaY4thGRKbQ8Mbt9l5T5nAhaGm4Kh+R+eC1E4lBfY9yjg3Q2+S0mRUjtGoqH5E5YgP0tsssrERxpcLmGp8TEZkKb9yMixVTRES2Td0AvaK6tgE671la54yZtW8xODFVVVWFTz/9FHv37kVeXl6jC4MTJ04YLTgiQ9zeAN3ZUSpyRJbnVHYRANtrfA5wbCMisgRMTBGZD147kRjUDdCPpd/C6axiJqZaobpGhdS6xufqqZFiMzgxNWvWLCQkJGDq1KkYOnQoy9DJbGg3QC9GZHBnsUOyOLbc+JxjGxGZyrFjx/DDDz8gIyMD1dXVWq/99NNPIkVlmbgqH5H54LUTiYUN0NvG3BqfA61ITP3222/Yvn077rzzTlPEQ9Rq2g3QmZhqDVttfA5wbCMi09i0aROefPJJREdHIyEhAdHR0bh06RJyc3Px0EMPiR2exWHFFJH54LUTiYUN0NvG3BqfA4DBc3UCAgLg6spyOTJP6kHqVBYHKUPZcuNzgGMbEZnGO++8g48++gj/+9//4OjoiI8//hipqamYNm0aunbtKnZ4FoeJKSLzwWsnEkv/QA8AwNnrxahR8v8DQ5lb43OgFYmpDz/8EC+99BKuXbtminiI2kSTmGL23GAZhRU23ficYxsRmcKVK1dw3333AQBkMhnKy8shkUiwYMECrF69WuToLA9X5SMyH7x2IrGEenaAq8weVQoVLuWViR2OxTHH9i0GT+UbMmQIqqqq0K1bN7i4uMDBwUHr9cLCQqMFR2SofgEeAIArN8tQJq9BRxkXntSXusos3AYbnwMc24jINDp37ozS0toGowEBAThz5gz69euHoqIiVFRUiByd5WHFFJH54LUTicXOToKIAHckXi3AqawihPu5iR2SxZDXKJGaUztLZkBd5Zk5MPiu/bHHHkN2djbeeecd+Pj4mM2cRCIA6OIqg7+7E64XV+FMdjGGd/MUOySLcSqrCEB9aayt4dhGRMaUkpKCgQMHYuTIkUhISEC/fv0wbdo0vPDCC/jjjz+QkJCAsWPHih2mxWFiish88NqJxNQ/UJ2YKsb0O8SOxnJcyC2FQinAw8UBQZ2dxQ5Hw+DE1OHDh5GYmIgBAwaYIh6iNusX6I7rxVU4lVXExJQB1BVT/cxkydD2xrGNiIxp8ODBGDRoEKZMmYLHHnsMABAXFwcHBwccOnQIDz/8MBYvXixylJaHq/IRmQ9eO5GY+rG3cKto7vnMqPE50IoeU71790ZlZaUpYiEyCnXFDwcp/SlVgmZ1BnMq6WxPHNuIyJj+/PNPDB48GB988AG6d++OJ554Avv378fChQuxbds2LF++HJ06dTL4uCtWrEBoaCicnJwQGRmJgwcPNrv//v37ERkZCScnJ3Tr1g2rVq1qtM+WLVvQp08fyGQy9OnTB1u3bm3TeZ999llIJBLEx8cb/PlawoopIvPBaycSk/qe5XxuCeQ1SnGDsSDqWTLmds9ncGLq3Xffxb/+9S/s27cPBQUFKCkp0XoQiY3LhxouLb8M5dVKODtI0b1LB7HDEQXHNiIypqioKHzxxRfIzc3FypUrkZWVhXvvvRfdu3fH22+/jaysLIOPuXnzZsyfPx+vvPIKkpOTMXLkSEycOBEZGRk6909LS8OkSZMwcuRIJCcnY9GiRXj++eexZcsWzT6JiYmYPn06YmJicPLkScTExGDatGk4evRoq877888/4+jRo/D39zf48+mDiSki88FrJxJTYCdndHJxgEIp4EJuqdjhWAxznSVjcGJqwoQJSExMxNixY+Ht7Y1OnTqhU6dO8PDwaNU3f0TG1r+uAfq1ggoUVVSLG4yFUA9Qff3dYC+1vcbnAMc2IjINZ2dnPPXUU9i3bx8uXryIxx57DP/9738RGhqKSZMmGXSs5cuX4+mnn8bs2bMRHh6O+Ph4BAUFYeXKlTr3X7VqFbp27Yr4+HiEh4dj9uzZmDVrFj744APNPvHx8Rg3bhzi4uLQu3dvxMXFYezYsVrVTvqeNzs7G8899xy+/fbbRk2QjYWr8hGZD147kZgkEgn61VX9nORMGb1UVis1qxj2N7PElME9pvbu3dvka8nJyW0KhsgY3F0cEOzpgmsFFTidXYyRPbuIHZLZUyembLXxOcCxjYhMr3v37nj55ZcRFBSERYsWYdeuXXq/t7q6GklJSXj55Ze1tkdHR+Pw4cM635OYmIjo6GitbePHj8fatWuhUCjg4OCAxMRELFiwoNE+6sSUvudVqVSIiYnBv//9b/Tt27fFzyOXyyGXyzXP9a2uYMUUkfngtROJrX+AOw5cvInTWUUAgsUOx+ydyymGUiWgi6sMvm5OYoejxeDE1OjRo7WeFxcX49tvv8WaNWtw8uRJzJ8/31ixEbVa/0APXCuowKksJqb0Ub8in3llztsTxzYiMqX9+/dj3bp12LJlC6RSKaZNm4ann35a7/fn5+dDqVTCx8dHa7uPjw9yc3N1vic3N1fn/jU1NcjPz4efn1+T+6iPqe95ly1bBnt7ezz//PN6fZ6lS5fijTfe0GvfhpiYIjIfvHYisfVnA3SDaIoRzKzxOdCKqXxqf/zxB5544gn4+fnh008/xaRJk3D8+HFjxkbUav0D1INUkbiBWACFUoWz12u/qbblxJQaxzYiMpbMzEy8+eab6N69O8aMGYMrV67g008/xfXr1/HFF19g+PDhBh/z9gtJQRCavbjUtf/t2/U5ZnP7JCUl4eOPP8b69ev1vtCNi4tDcXGx5pGZmanX+7gqH5H54bUTiUU92+PijVJUVrMBekvMeZaMQRVTWVlZWL9+PdatW4fy8nJMmzYNCoVCs5oLkbng8qH6u3SjDPIaFVxl9gjxtM3G5xzbiMjYxo0bh71796JLly548sknMWvWLISFhbX6eF5eXpBKpY2qo/Ly8hpVM6n5+vrq3N/e3h6enp7N7qM+pj7nPXjwIPLy8tC1a1fN60qlEv/6178QHx+P9PT0RrHJZDLIZDI9Prk2VkwRmQdeO5E58HGToYurDDdL5Th7vRhDQjqLHZJZM+dZMnpXTE2aNAl9+vTBuXPnNN/2ffrpp6aMTcutW7cQExMDd3d3uLu7IyYmBkVFRU3ur1Ao8NJLL6Ffv37o0KED/P398eSTT+L69eta+919992QSCRaj0cffdTEn4ZMLSLAHRIJkFNchbzSKrHDMWuns4sA1Cbz7OzMq6SzPYg9thGRdXJ2dsaWLVuQlZWFZcuWtSkpBQCOjo6IjIxEQkKC1vaEhASMGDFC53uioqIa7b97924MGTJE05y8qX3Ux9TnvDExMTh16hRSUlI0D39/f/z73/82qI+WPpiYIhIfr53IXEgkEgxgQYJeSqsUuJpfDsD8VuQDDKiY2r17N55//nn84x//QM+ePU0Zk04zZsxAVlYWdu7cCQB45plnEBMTg19//VXn/hUVFThx4gQWL16MAQMG4NatW5g/fz4efPDBRqWlc+bMwZIlSzTPnZ2dTfdBqF10lNmje5eOuJxXhjPZxbint3k1dzMnJ810ydD2IvbYRkTWadu2bUY/ZmxsLGJiYjBkyBBERUVh9erVyMjIwNy5cwHUTo/Lzs7Ghg0bAABz587FZ599htjYWMyZMweJiYlYu3YtNm7cqDnmCy+8gFGjRmHZsmWYPHkyfvnlF+zZsweHDh3S+7yenp6aCiw1BwcH+Pr6tjkhdzuuykckPl47kTnpF+CBPal5OJ3NxFRzzmSXQBCAAA9neHU0vGLZ1PROTB08eBDr1q3DkCFD0Lt3b8TExGD69OmmjE0jNTUVO3fuxJEjRzBs2DAAwBdffIGoqChcuHBB50WPu7t7o2/3Pv30UwwdOhQZGRla5eYuLi7w9fU17Yegdtc/0B2X88pwMrMY9/TWPc2BgNOaJnge4gYiEjHHNiIiQ0yfPh0FBQVYsmQJcnJyEBERge3btyM4uHYlopycHGRkZGj2Dw0Nxfbt27FgwQJ8/vnn8Pf3xyeffIJHHnlEs8+IESOwadMmvPrqq1i8eDG6d++OzZs3a6639Dlve2LFFJH4eO1E5qR/UO2X6yfZW7hZmlkyAeZZjKD3VL6oqCh88cUXyMnJwbPPPotNmzYhICAAKpUKCQkJKC0tNVmQiYmJcHd317pIGj58ONzd3ZtcIlmX4uJiSCQSeHh4aG3/9ttv4eXlhb59++LFF19s8bPI5XKUlJRoPcj8qBugM3veNHmNEudzbbvxuZhjGxGRoebNm4f09HTI5XIkJSVh1KhRmtfWr1+Pffv2ae0/evRonDhxAnK5HGlpaZoqp4amTp2K8+fPo7q6GqmpqXj44YcNOq8u6enpJlmRi4kpIvHx2onMifqe7+rNcpRWKUSOxnypZ8moE3nmxuBV+VxcXDBr1iwcOnQIp0+fxr/+9S+8++678Pb2xoMPPmiKGJGbmwtvb+9G2729vZtcIvl2VVVVePnllzFjxgy4ublptj/++OPYuHEj9u3bh8WLF2PLli06L8gaWrp0qabXlbu7O4KCggz7QNQu+gd5AKht8saSf93O55RCoRTQycUBgZ1sewqrGGMbEREZhqvyEZkPXjuROfDsKEOAR+19DAsSmmbus2QMTkw1FBYWhvfeew9ZWVla/Qr09frrrzdqPH77Q90PStfywy0tkaymUCjw6KOPQqVSYcWKFVqvzZkzB/feey8iIiLw6KOP4scff8SePXtw4sSJJo/X2iWOqX318XODvZ0E+WXVyClmA3RdTmXXLxmq7xLftqCtYxsREZkGK6aIzBOvnUhM6pkfp9kAXaeiimpkFFYAMN+pfHr3mGqOVCrFlClTMGXKFIPe99xzz7W4Al5ISAhOnTqFGzduNHrt5s2bTS6RrKZQKDBt2jSkpaXhjz/+0KqW0mXw4MFwcHDApUuXMHjwYJ37tHaJY2pfTg5S9PJxxbmcEpzKKoK/h21XBOlyKrMIgO1O42tJa8c2IiIyDSamiMwbr51IDP0DPbDjTC5X5muC+ucS4ukCdxcHkaPRzSiJqdby8vKCl5dXi/tFRUWhuLgYf/31F4YOHQoAOHr0KIqLi5tcIhmoT0pdunQJe/fubbRijC5nz56FQqGAn5+f/h+EzFb/QPe6xFQxJkTw7/R26nJXc82cExERNcRV+YiI6HbqL9lP1TX4Jm2ae75AD3EDaUabpvK1l/DwcEyYMAFz5szBkSNHcOTIEcyZMwf333+/1op8vXv3xtatWwEANTU1mDp1Ko4fP45vv/0WSqUSubm5yM3NRXV1NQDgypUrWLJkCY4fP4709HRs374df/vb3zBo0CDceeedonxWMq5+gWyA3pSK6hpcvFHbnHJAXT8uIiIic8aKKSIiul1E3ZfsmYWVuFVeLXI05udk3SyZAWY8S8YiElNA7cp5/fr1Q3R0NKKjo9G/f398/fXXWvtcuHABxcW1CYisrCxs27YNWVlZGDhwIPz8/DQP9Up+jo6O+P333zF+/HiEhYXh+eefR3R0NPbs2QOpVNrun5GMb0BdVvhUVjG/Xb3NueslUAmAt6sMPm5OYodDRETUIiamiIjodu7ODgj16gCgvocu1bOEWTKiTuUzROfOnfHNN980u0/DxENISEiLiYigoCDs37/fKPGReerl4wpHqR2KKxXIKKxAsGcHsUMyG+q5xv3NuKSTiIioIa7KR0REuvQLcEdafjlOZxVhdK8uYodjNvJKq5BTXAWJpL6yzBxZTMUUUWs42tsh3L+24f1JNsPTciqrCAAbnxMRkeVgxRQREemivqfhPZ829UqFPbp0RAeZ+dYlMTFFVq9/gHr50CJxAzEz6oqpfkxMEYA+2VfFDoGIqEVMTBERkS7qWSCnmZjSctJC7vmYmCKrx+x5Y8UVClzNLwcADORUPgL/MyAiy8BV+YiISJe+/m6wkwC5JVXIK6kSOxyzkVLX+HyQmS92xXsRsnrq7PmZ7GIoVbyQBYCUuuqxEE8XdOrgKG4w1KKQkBBIJBKtx8svvyx2WERE7Y4VU0R069YtxMTEwN3dHe7u7oiJiUFRUVGz75k5c2aja6nhw4e3T8DULjrI7NHDuyMAFiSoCYKgWZFvYFAncYNpARNTZPV6eHdEB0cpKqqVuJxXJnY4ZiElowgAMNDMM+dUb8mSJcjJydE8Xn31VbFDIjIbd6SdEzsEaidsfk5EM2bMQEpKCnbu3ImdO3ciJSUFMTExLb5vwoQJWtdS27dvb4doqT2pV2RXJ2NsXVp+OYorFXC0t0NvP1exw2mW+Xa/IjISqZ0E/QM9kHi1AMkZtxDma96/lO0hJfMWAGAAE1MWw9XVFb6+vmKHQWSWOshZsm8rWDFFZNtSU1Oxc+dOHDlyBMOGDQMAfPHFF4iKisKFCxcQFhbW5HtlMhmvpazcoK6d8ENSFpLr7nVsnXoaX4S/Gxyk5l2TZN7RERnJwK4eAOp/OW2ZIAianwMrpizHsmXL4OnpiYEDB+Ltt99GdXW1cU/Afi1EZAGYmCKybYmJiXB3d9ckpQBg+PDhcHd3x+HDh5t97759++Dt7Y1evXphzpw5yMvLM3W41M7U9zYnM9nCBUCDez7znsYHsGKKbIR6kEqum8JmyzIKK3CrQgFHqR36+LuJHQ7p4YUXXsDgwYPRqVMn/PXXX4iLi0NaWhrWrFnT5HvkcjnkcrnmeUlJSXuESkRkUkxMEdm23NxceHt7N9ru7e2N3NzcJt83ceJE/O1vf0NwcDDS0tKwePFi3HPPPUhKSoJMJtP5Hl5LWZ5ePh3h7CBFmbwGV26WoZePbc+U0SSm6oo0zBkrpsgmqFchuJhXijJ5jbjBiEw9QIX7u0FmLxU3GBv2+uuvN2rCefvj+PHjAIAFCxZg9OjR6N+/P2bPno1Vq1Zh7dq1KCgoaPL4S5cu1TQFdXd3R1BQUHt9NKJ2JwG/FbUVXJWPyDoZcl2k7jXXkCAIOrerTZ8+Hffddx8iIiLwwAMPYMeOHbh48SJ+++23Jt/DaynLYy+106zInmLjBQlVCiVSc2qTqea+Ih/AiimyEd5uTgjwcEZ2USVOZRZhRA8vsUMSjbpqzBIGKGv23HPP4dFHH212n5CQEJ3b1avIXL58GZ6enjr3iYuLQ2xsrOZ5SUlJsxdUTV/KERGZD1ZMEVknfa+LTp06hRs3bjR67ebNm/Dx8dH7fH5+fggODsalS5ea3MfQaykyDwO7euBoWiGSM29h2h22+/d19noJFEoBnh0cEdjJWexwWsTEFNmMgV09kF1UiWQbT0yxv5R58PLygpdX6/4dJicnA6i9qGqKTCZrsjRdN1YfEJH546p8RNZJ3+uiqKgoFBcX46+//sLQoUMBAEePHkVxcTFGjBih9/kKCgqQmZlp5GspMgeD6vop2XoLl4b3fM1VE5oLTuUjmzGIfaYgr1Hi3PXakk4mpixDYmIiPvroI6SkpCAtLQ3ff/89nn32WTz44IPo2rWr0c4TnnPNaMciIjIVVkwR2bbw8HBMmDABc+bMwZEjR3DkyBHMmTMH999/v9aKfL1798bWrVsBAGVlZXjxxReRmJiI9PR07Nu3Dw888AC8vLzw0EMPifVRyEQG1fVTunijFOU23MLF0ooRmJgimzGowcp8ttqbIjWnFNVKFTq5OCDY00XscEgPMpkMmzdvxt13340+ffrgtddew5w5c7Bx40ajnqdzRalRj0fUniS2OaTbJCamiOjbb79Fv379EB0djejoaPTv3x9ff/211j4XLlxAcXExAEAqleL06dOYPHkyevXqhaeeegq9evVCYmIiXF1tuzm2NfJxc4K/uxNUAnAqq1jscESTknkLgGU0Pgc4lY9sSF9/d9jbSZBfJkfWrUoEdba9xExKRu0ANcBCSjoJGDx4MI4cOSJ2GEREZoGJKSLq3Lkzvvnmm2b3afgltLOzM3bt2mXqsMiMDOzqgeunc5GceQtR3XX3Y7VmBWVyZBZWAgD6B3qIG4yeWDFFNsPJQYo+/m4A6ksbbY2llXQSEelF4OWMreCqfERE1BJ1nylbXZlPfc/XvUsHuDs7iBuMnnglRzZloI33mWJiioiskVOV7S5oYWtYMUVERC1RT19LttEWLvX3fJ3EDcQATEyRTanvM3VL3EBEcKu8GukFFQCYmCIia8OpybaCq/IREVFLIupauNwsleN6cZXY4bQ7TWLKQvpLAUxMkY1RZ43PXC9BdY1tXdSmZBUBAEK9OsDDxVHcYMgsdSovETsEotaxwW9DbRUrpoiIqCXOjlL09qttbJ+cYVsFCSqVoElMDbKgYgQmpsimhHi6wMPFAdU1KqTm2NZNuHqONaulqCl2vLknIjPHxBQREenDVvtMXc0vR2lVDWT2dgjztZxVJ5mYIpsikUga9Jmyrex5MvtLUUuYmCIzIlUq9d6XE/lsBxNTRESkD809n40teqW+x+0X4A4HqeWkeywnUiIj0WTPbWiQUqoEJF+rHaQigy2nCR61H69S3uSReRl1KUXsEMgMcVU+IiLSh7q38JnsYptq4ZKkvucLsax7PiamyOY0XKXBVlzKK0WpvAYdHKXobUElndR+hqar0BFOYodBpLfI9PNih0AiYMUUERHpI9SrA9ydHSCvUeF8ru20cFEnpoYEdxY5EsMwMUU2Z2CgBwDgWkEFbpbKxQ2mnRxPrx2gBnb1gL0FlXRS+3FUCpAJDmKHQRYgJD8HD6QcEjsM2An1iYmQfp4iRkLtiavyERGRPhq2cFEna6xdUUU1LuWVAQAGW9CKfAATU2SD3F0c0MunIwDbGaROqEs6u1pWSSe1J3bpIf2E56QjoDhf7DC0uLhypVFbwYopIiLS15C6FibHbeSeL7mu0XuoVwd4dpSJG4yBLCYxdevWLcTExMDd3R3u7u6IiYlBUVFRs++ZOXMmJBKJ1mP48OFa+8jlcvzf//0fvLy80KFDBzz44IPIysoy4SchcxBZV9qYdK1Q5Ejax3HNXGPLKukkskT9si4j6sppscPQ26CMi2KH0Ei3m9lih0BmiokpIiLSl7rPUlL6LZvoTXi87t7WEnsKW0xiasaMGUhJScHOnTuxc+dOpKSkICYmpsX3TZgwATk5OZrH9u3btV6fP38+tm7dik2bNuHQoUMoKyvD/fffD6UBqwGR5bkjxHay53mlVcgorIBEUt8EkIhMp2NVJQZkXRE7DJvCej/bwcQUERHpa2CQB+ztJMgtqUJ2UaXY4ZhckgUvdmUvdgD6SE1Nxc6dO3HkyBEMGzYMAPDFF18gKioKFy5cQFhYWJPvlclk8PX11flacXEx1q5di6+//hr33nsvAOCbb75BUFAQ9uzZg/Hjxxv/w5BZUDeDO5NdjCqFEk4OUpEjMh31NL4wH1e4ObGHEJGlcq0sR6lzBxMc2TK/QXSpto0egaSNq/IREZG+XBzt0dffDSezinE8/RYCO7mIHZLJKJQqzarzQywwMWURFVOJiYlwd3fXJKUAYPjw4XB3d8fhw4ebfe++ffvg7e2NXr16Yc6cOcjLy9O8lpSUBIVCgejoaM02f39/RERENHtcuVyOkpISrQdZlqDOzvB2lUGhFHDSylfnU2fOB1vgAEVkTiKyr+q9b4iyi1HPfUfaOUSf+8uox2yrvx3/Q9Tz98irn3bf355JClvBiikiIjLEkLpWJsetvIVLak4JqhQquDnZo3uXjmKHYzCLSEzl5ubC29u70XZvb2/k5uY2+b6JEyfi22+/xR9//IEPP/wQx44dwz333AO5XK45rqOjIzp10r5h9/Hxafa4S5cu1fS6cnd3R1BQUCs/GYlFIpFgiI1M5zuuWTKUiSmi9jJC0XQlb2tEZlyEs6LaKMfqnXNNe0Mrczqe5eJ+KWMnCHhm/y+Ydeh/cOdcPpvBVfmIiMgQmgbo6VZ+z5deP43Pzs7yLoxETUy9/vrrjZqT3/44fvw4gPoLkYYEQdC5XW369Om47777EBERgQceeAA7duzAxYsX8dtvvzUbV0vHjYv7//buPLypKv0D+DdJkzTd0iV0L21ZWyhrEShrESygLAqyuKDOKOooIww44zCOgqMOozM4uG+DgMtPGGRRhEEBKRQoCIWy7xRKS3fadKP7/f1RGpo2e5Nm6ffzPHkgN+eee3LTnt775pz3LIZardY8rl+/buI7JkfSlAD9yFXXjZ5X1dbjVLYagHPONSZqbtLxfXY7tr1HB1lTW8cWiZrVEFWmHSwLrWzfYIEYAmT1de16TLIvjpgiIiJzNCVAP59XBvWtWju3xnbSMp03vxRg5xxT8+bNw+zZsw2WiYqKwokTJ5CXl9fqtYKCAgQFBZl8vJCQEERGRuLixYsAgODgYNTU1KC4uFhr1FR+fj6GDRumtx65XA653LmWX6TWmhKgp10rRkOD4JSRZWNOZqtRWy9A5SVDZ3/XnVNNHYNHTTVGXDyOm54+OBMa3a7HtsbooN7ZV3A6rIsVWmNdben53j+Yj8n3hFutLea5EyALCwuzUxuovTEwRURE5gj0dkdkgAeuFVXiWGYxEnu2nonl7ARBQNpV507fYtcRUyqVCjExMQYf7u7uSEhIgFqtxq+/3smvcejQIajVaoMBpJaKiopw/fp1hISEAADi4+MhlUqxY8cOTZmcnBycOnXKrHrJOcWG+EAhlaC0qg4X88vt3RybaL4yg6FRgEQAIBj4ERlzLq39GmJA3I0MjLp43N7NaOWujLNGy7T1NzCiPqCNNVinHaJmSafFNkztJDJSt6RZO0JDQ23XEHIoDEwREZG54l18Ot8NdRVyS6sgEYvQP8LX3s2xiFPkmIqNjcWECRMwd+5cHDx4EAcPHsTcuXMxadIkrRX5YmJisGnTJgBAeXk5XnzxRaSmpuLq1atITk7G5MmToVKp8MADDwAAlEolnnzySSxatAi7du3CsWPH8Oijj6JPnz6aVfrIdUklYs0vrqsmw2s+15jIGH1xgB65meiZxynLhrg11JtULqhUd1/TIzfTms2xqbDiQns3gTowrspHRETmalqR3XXv+RrfV68QH3jI7DopzmJOEZgCgG+++QZ9+vRBUlISkpKS0LdvX3z11VdaZc6fPw+1ujGfjkQiwcmTJzF16lT06NEDjz/+OHr06IHU1FR4e3tr9vn3v/+N+++/HzNnzsTw4cPh4eGBLVu2QCKRtOv7I/vQTOdzwei5IAg46uRzjcl1SeusP8c/oFyNsWePaG2LLMpFrxsZVj9Wc2HF+cYL3eZVVWnxcYQ2Z4fSrVt+FmJyrppcXtysHRyHSe2NI6aIiMhcTfd86ddLUFvven8/ms+ScVZOE07z9/fH119/bbBM82/PFAoFfvrpJ6P1uru74/3338f777/f5jaS84m/vXzoYReMnl/ML8fNihq4S8XoE+Zr7+aQg6upKGrX4z2YthvfDkmyap2e1bfQPT8Lu2IH3dnYDqMqAstaB7btPZYjJucazoVE6nytZdv8K8uQeCEd50Ki2nxcQ9NBiayBq/IREZG5unbyglIhhfpWLU7fKHXa6W76HLrSeC87ONrfzi2xnNOMmCKyhYGdfSEWAddv3kJeaZW9m2NVh640BhoGdvaDzI2/6mRYTXkeKmrVBsv0z7xgteMp2zByyFrca6rt3QQbMhwaM2eElE+V6X2jvQNy5Po4YoqIiMwlFoswSJNnyrUGJNysqMH5vDIADEwROS1vdyl6BvsAAA67WCd1KKPx/QyJtk7CZHJ1AmqFGoMlhmacaae26Cayc9gjtKTApHI9LMjH5V5vfKiRNd/96AvpmJqegif2bzVa1rTsWUTtg4EpIiKyRPzt6Xyuds/X9H66BXpB5SW3c2ssx8AUdXhDbkeWf81wnU5KEARNYMqZI+dE5rD1LLLowhy9rzUPGinqDAf4dJG08022CECIugjuVs71VdNgzfAZx19RawxMERGRJZru+Q5fLUaDVa9X7MsVpvEBDEwRYWiXxl/ig1faN8eOLV0tqkRBWTVkEjEGdPa1d3PICajmzYO42Y1esNr1V17rm3XJrPL6A1+6X1l+tBIDb9ahz9k0eFTfwoDM83prSLx0oU2J0c0hEsy9oTc95MfRVbb30UcfITo6Gu7u7oiPj0dKSorB8nv27EF8fDzc3d3RpUsXfPLJJ63KbNiwAb169YJcLkevXr00Kxybetza2lq89NJL6NOnDzw9PREaGorHHnsMN27caPsbboGr8hERkSX6hPlCIZXgZkUNLuaX27s5VvPr1cZ72CEMTBE5t8G3p7pdyCtHYblr5Jxpyi/VP8IX7lKuMEnGefTvD0l9nea5oqb1qJ97awa2Z5Nas+Q+VKQ/qDLg+kXL22KUgNEF9fjs8C088f1XmHPwJ3gayGklryrDrMO7dL7WLT/Luk2rt134SNxQZ7wQWWzdunVYsGABXn75ZRw7dgwjR47ExIkTkZmZqbN8RkYG7r33XowcORLHjh3DX/7yF7zwwgvYsGGDpkxqaipmzZqFOXPm4Pjx45gzZw5mzpyJQ4cOmXzcyspKHD16FK+88gqOHj2KjRs34sKFC5gyZYrVzwFHTBERkSVkbmIMuj2dz1UGJJRW1eLMjVIAwNAuzp2+hYEp6vD8PWWICfYG4DrT+TT5pbo4d+ScHItMsPNCriYO3DF15FF7LiAnavFvS+W1JUbrEKordG73uVUBv4pS7eMZumkXmf45+pQZb5dW1QIDU7b0zjvv4Mknn8RTTz2F2NhYrFixAhEREfj44491lv/kk0/QuXNnrFixArGxsXjqqafw29/+Fv/61780ZVasWIF77rkHixcvRkxMDBYvXoyxY8dixYoVJh9XqVRix44dmDlzJnr27ImhQ4fi/fffR1pamt6gmaW4Kh8REVmqKXjjKoGpI1dvokEAogI8EOTjbu/mtAkDU0RwvU7qV+aXog5sxpFf2lyHe001oGOq0PBLJ3SWNxTkSvfua8IR9Q8HC6v3x+iaXmio0B04H3rlNDoX5ZlwjEYikemJMWf8uMrksgCzQtlSTU0N0tLSkJSUpLU9KSkJBw4c0LlPampqq/Ljx4/HkSNHUFtba7BMU52WHBcA1Go1RCIRfH19db5eXV2N0tJSrYcpOGKKiIgs1ZTC5VDGTZfIM+VKOYUZmCKCa+WZun6zEtklt+AmFiH+9rKoREbZaPhQD7UVk2ubeP0gbzYl0a1BDGlxvs5yPQ38Cex//aLOqX4eJg8IunNC13cbb0Jpmd7X5JCie0MI3Gt1T8ELPJeGbie1AwRuBqbrSeo8jbYHAGJyrsFfXQTBwM/GmZpcreciES8rbKWwsBD19fUICgrS2h4UFITc3Fyd++Tm5uosX1dXh8LCQoNlmuq05LhVVVX485//jIcffhg+Pj46yyxbtgxKpVLziIiI0PPOtTEwRURElnK1PFNNic9dYRV2XkESwbXyTDVFzvuEK+Ehs/PUK3IeNvjS6OlL1ehdYsHULkHAH080BokHnUtrUxv86/SPDtIXbxl79gj6ZF9utV1+sxB31XbRuY+x0zcqeQ8AaCWYb0lsJCm5761b8L5VAffaasReOI7hv+7ExF++Q0hBNrpcv4gXOgdqyg68ehYBeqbhSRrkkFSUGWmxbr2zr2g9946q0n4PHh469wsoLrDoeKbo1KmTzep2RKIWedMEQWi1zVj5lttNqdPU49bW1mL27NloaGjARx99pLddixcvhlqt1jyuX7+ut2xzDEwREZGlXCnPVEV1HU5mqwG4RvoWBqaI4Fp5pn7NaFqZwfkj5+RozIte+dRaFu0SA5iVI8Pvv3gDQ84c1lkmsLT176nKzHxI+nTPz4JEENDy/cpvFsANli0mEJKbi2H79mNwxhkoK3V8QycSQSIIGHT1bKuXmlohr6nHQ7/uwGMH/ofJe/Zi2NFkxF1I15Qb6uul+b+itgYzjibrbIsAQFRv+ki25qv4uTVoj8Tykpt2Pj5a9meTj2cKj+o7AbH4+Hir1u2oVCoVJBJJq1FK+fn5rUYzNQkODtZZ3s3NDQEBAQbLNNVpznFra2sxc+ZMZGRkYMeOHXpHSwGAXC6Hj4+P1sMUXJWPiIjawlVSuBzNLEZ9g4AwXwXC/XR/MehMGJgius1VOqnUK66xZCh1XAKAvbnfwb1GezRO87EZXlW3NP+fceQXDM44gwGZF8w6jsjI/MW+WZfhXa7GA8nbMemHLRDVVhksr033TbNXTRUeOrxTR2Ma29I/U/9KgZ5VtRhw7BiUN7tBJGox9U/kbUbbzGQgACAR22YOqLRCCsWNDL2vq8pLbHJcRyaTyRAfH48dO3Zobd+xYweGDRumc5+EhIRW5X/++WcMGjQIUqnUYJmmOk09blNQ6uLFi9i5c6cm8GVtHDFFRERt4Sp5plIvu9Y9H+f5EN02tIs/Vh+46tSBqcyiSly/2ZhfyhWS4FH76nQzDznBnRufGJlWZlOCgJxbrafSNSdqFvgJqChFQIXuxMlV9RXw1zG6CgAEmf68TkDjqKMXP38Doy5kAQAaUI+rRQcAjDW4nz6B+Y25riTVt3S82vh+3Aycd7HQgJjzF3AjJBQ1LV6TeU/Xei6CYIXZmUJTZa3IczPRIJXBp0vnNh9Fl/ZcMdGZLFy4EHPmzMGgQYOQkJCAzz77DJmZmXj22WcBNE6Py87OxpdffgkAePbZZ/HBBx9g4cKFmDt3LlJTU7Fy5Up8++23mjrnz5+PUaNG4a233sLUqVPx/fffY+fOndi3b5/Jx62rq8ODDz6Io0eP4scff0R9fb1mhJW/vz9kRn7XzMFV+YiIqC1a5pnqGWzDL/dsaP+lxlyRw7up7NwS6+CIKaLbhrhAnql9tzuogZ394Cln3JnMk3hwu+b/buVqO7akmTZGVyrrSjHw5EEMSt/X6jWhR6zR/UUAghbfmYbW6XyyScetqGs8f9W4E4SS19Rg2ncbEHrKcN6s4b/uRFBBtknHAQCR2B9iifUD0VK1/iC9rDgf7vmNAbu47Dsjm0ZWt+53RuVbMQG+DobyK7maWbNmYcWKFfjb3/6G/v37Y+/evdi2bRsiIyMBADk5OcjMzNSUj46OxrZt25CcnIz+/fvj9ddfx3vvvYfp0+8EMocNG4a1a9di1apV6Nu3L1avXo1169ZhyJAhJh83KysLP/zwA7KystC/f3+EhIRoHoZW7rMER0wREVFbuEKeKXVlLU7czi/FwBSRi/FrlmfKWTspV4ucU/vx8nOHl/Qena8Jt2/8C6pMS07cXE1DG4O8bcwjI4ga8yKNObgd92//Gp1Ki++86GZa8Nb/8cc1/xeb2J7U/C24XHYcp2QHtbZL6+ogMlJHwrE9eGzDxzpe0befjQIzt9sZUah75bUm4mbD4B+8JcV9cmDgyVTNtjdOtJ4CGbvzYKttZJrnnnsOV69eRXV1NdLS0jBq1CjNa6tXr0ZycrJW+dGjR+Po0aOorq5GRkaGZpRTcw8++CDOnTuHmpoanD17FtOmTTPruFFRURAEQecjMTHRau8dYGCKiIjarimFy4HLhXZuiWVSrxRCEIBugV4IVrrbuzlWwcAUUTNNAZ19F52vk2poEDSd64juTHxO5vEP9cT4Z+43WCb95m6z6sy/dQ3FNTltaFVLbQtSdb96Dvec1Z1M3RTjTmVg1LlM4wUBVNaX4kjhdoj9ZJg5KPzOCxIJOpVWGtxX5jUTEllvzXPBgriTuKrC5LKeWYaXS558ZK/JdckhwqveQO8LxzTbPLTzpWPgyQPwKlYj7prhoGVInvmBUHJ9DEwREVFbNd3zHbhchLp65/t7sv9S4yCKES40GIGBKaJmRnZv/OVOuVjodCv+nMkpRXFlLTxlEvQN97V3c8hEb775JoYNGwYPDw/4+vrqLJOZmYnJkyfD09MTKpUKL7zwAmpqWmYZarueQ4INvm5u1qLsyotosGeuKh18qioxJT0FMw/vMpr8vCVZfQO8qk2blhZdUIVOpRW4+zfP4O0H+2m2S0NCMObf7+OB69qfX/O2SKWBkHqO11Gr4fPvJ70zAsz9xlVIiwyNdLqzmp5bVb2BcoCixlAASXv1Qr8g46vCBBQXGC0DAH6lN/HEf983Wk6pVJpUH7kGrspHRERt1SdMCaVCirKqOhzPcpD0FWZomiUzrKvrDEZgYIqomSHRAZBJxMguuYUrhaaPOHAETR3U0C4BkEr4q+0sampqMGPGDPzud7/T+Xp9fT3uu+8+VFRUYN++fVi7di02bNiARYsW2bRdIfmm5zjSp8eVU6itLtH52rDaHvC+ZZ/fsVB1EfwryxAQFGK0bLCn8SXsO9/Ma7UtpKQQd2XkwtPXT/sFkQhe8YPw8hntYE+A553Em53cdSdyN2aAjwd+X5iF+1PLIa6v0+SA0kUsjbLoGMbEDg81WqZSosA5r54mDYDrpOPcNjdu3Dj07NnT1OaRC+CIKSIiaiuJWITh3RqDOikXTfvCzFE03aeKRcBQBqaIXJNCJtEkw3O26Xz7mF/KKb322mv4wx/+gD59+uh8/eeff8aZM2fw9ddfY8CAARg3bhyWL1+Ozz//HKWllgUwDNkzOAZ/6xaKQSf2GywXdf1iq20edXciDT/U58KrshwhepJ49/bqArEJI5YkZkxJM1dgeKTRMhNWfmnwdZ+KMoQYSBJuKk/5nRFMMcobFtfzzM0b6JNpwmg6kcR4ER3bVGUlBvcRi0UQSyQQDHy2o2Y9ih8XWrayYUsjRozoUMnPiavyEZFpo81bEgQBS5cuRWhoKBQKBRITE3H69GnbNpQc2sjunQA43z1f02CEfhG+8HGX2rk11sPAFFELTZ2UM0XPq2rrcfjqTQDAiO4MTLmS1NRUxMXFITT0zkiU8ePHaxIQW1tPT3c8HREImSzeYDlfv9arwNULd6aE3TV2PB56/V969w/+/UAIYn0BhTsBLrdK6wffmpgS0JB5eGo9dwvRHmUVfcPMPEgmHNNfVoHRe/+geW7OhCWxuxz9TnxoXpv01QU33HW5MUg28/AujD5/DN10BBqlLeIDYTG9ENy1u956/TxkTrs0M9kfR0wRkbHR5rq8/fbbeOedd/DBBx/g8OHDCA4Oxj333IOysjIbtpQcWVN+pmPXS1BaZdsVhK2pKTDlSvmlAAamiFppyjOVerkINXXOceF7NLMYVbUN6OQtR/dAL3s3h6woNzcXQUFBWtv8/Pwgk8mQm6s/h1B1dTVKS0u1HuZwcx+u+b9P06ilZjldYiuLW+6CBtwJTIlEIoT2iIFIonvlO4mXzFZryVnsD5FB+M3JFDP3ss27kDRYlkPMd8YMhEfKDZaR1muHuiRuui8FfKQh6FR+CwDgX1mG2NxrkBdk4ze/+Y1WufAK7XMgFksw9snnjLZV3CIwqbpV1+yZo/10kKNgYIqIjI02b0kQBKxYsQIvv/wypk2bhri4OKxZswaVlZX4v//7Pxu3lhxVhL8Huqg8Ud8gIPWyc6zILgiCJvG5q82SYWCKqIVeIT4I8JShoqYexzJb33w7or0X7kTOOa3F/pYuXQqRSGTwceTIEZPr0/WZCoJg8LNetmwZlEql5hEREWHWexCJRHhktxpx1y+hf+ZFSEsKIK6+pXldLADPFLc8fuv29HCTAQC8jUwBs4eWyc9f6hIC/1uGV8xrSWZuEvpmh3w0486+MVKZ0V3ltXpuxJv9HIg9PBD1re6L7LHHjmCgjwd+46YdvO7cq/Xot8amth6rJS/MQWSk9hRIWYNlfY5UoT2dsGdJ1Z1ji70hkgRB4n4XemRmWFQ/uSYGpojIXBkZGcjNzUVSUpJmm1wux+jRo3HgwAE7tozs7c7CV84xU+ZMTikKy6uhkEowoLOvvZtjVQxMEbUgFos00+FSnGTOcfL5fABAYs9Odm4JAcC8efNw9uxZg4+4uDiT6goODm41Mqq4uBi1tbWtRlI1t3jxYqjVas3j+nUzp5wB6JJfjwlb18Dv9CG451yzaAxL57sfxOCN2yFLMbRCHDD2rGmBOrfatg217nrubJv2b3kS3OrrdJfTu/udCuZdrMayXSfx38XPQSk2/uc4RF1u1rFa6nv5CrbF98DUh2PhH3xniqKuEVNetxow9ZT+5OlNRFLDo7MMCe3ma6BiMeQ+j0CqGIn3//26xccg18NV+YjIXE3XUS2vm4KCgmw6+pwc350ULs5yz9cYQBveTQW5m/F8oc6EgSkiHZwpz9SNkls4l1sGsQgY1Z2BKUegUqkQExNj8OHu7m5SXQkJCTh16hRycnI0237++WfI5XLEx+vPAyWXy+Hj46P1sITIYBpr48Tunjjh2QcVEsM5hbrrXUFO++iK+no95fTR3j84x3BicXMHHD6/vjE5emzONSirKtE7+woAQFJvfDSHmwAkXitCp5JiuDcLTOlrgriNN+Ky0sZzJ5VLoAr21VvuiSvVWPBDCfxvVbd6zc1jHAAgJiIaSrkCcdMeNOnYntW16H7lNGQ1VZgS2HhsubvuaZ7NBeYdgUQQoNCbj4w6Go6YInJN1h5trkvLkea2Hn1Ojm9o1wC4iUW4VlSJa0WOvyL77nONgxHGxLjePZ/TBKaKi4sxZ84cTccwZ84clJSUGNxHX6f2z3/+U1MmMTGx1euzZ8+28bshR9c0rPNEthpF5a1vzhxJU+R8QGc/+Hkanw5EjiUzMxPp6enIzMxEfX090tPTkZ6ejvLyxtExSUlJ6NWrF+bMmYNjx45h165dePHFFzF37lyLg02WCgjvrPVcGuSh9dzW4xdE1jiA0Py/batw6NFkhBU2XiDI62vx19TtGHnpBABAKgjo/OWa1jvpuQCWi0SYO3cunn32WYhbXjib0Ba5p3aQp/+1PHTPval96OZtv2sEpNV+8C6JbVVXRV2p3uCYm7wvAOC+f72Pp75cD6ncXWfZ5nEkiZ8InjW1ePPj5TgYEwRfqe6AVEWtWuv5yH1/RNzZVTpaTx0ZV+Ujck3WHG3eUnBwMAC0Gh2Vn59v89Hn5Ni85G4YGNm4IvueC449IKGksgZHb6eZSewZaOfWWJ/TBKYefvhhpKenY/v27di+fTvS09MxZ84cg/vk5ORoPb744guIRCJMnz5dq9zcuXO1yn366ae2fCvkBIJ83NErxAeCAOw+79id1O7b0/jGcBqfU3r11VcxYMAALFmyBOXl5RgwYAAGDBig+VZQIpFg69atcHd3x/DhwzFz5kzcf//9+Ne/9K94Zystv1WUBpuXaH+iSgkAGOnXuF9bwkJdChpHPnlWmZcTyhiRyPI/i25KpdZzz8GDzdo/LCxMc/FsDh+VAo8sHaq1LbSkHN3z9OfIU7gr4FvcB+5VrfuN7IoLZh2/vjyn1bY4LwUGe3ngPpEcXgmNSxl3Kr+F0MhovfVU1JVo/h/Z2x/SOut+tuQaOGKKyDVZc7R5S9HR0QgODsaOHTs022pqarBnzx4MGzZM737WGn1Ojm3M7SDPrrP5dm6JYXsvFqJBAHoGeSPMV2Hv5lidUwSmzp49i+3bt+M///kPEhISkJCQgM8//xw//vgjzp8/r3e/4OBgrcf333+PMWPGoEuXLlrlPDw8tMopW9xcUMc0Lrapk8qzc0v0q66r1ywZ6oqR845g9erVEASh1SMxMVFTpnPnzvjxxx9RWVmJoqIivP/++5DLLc/tYw65z2Sd26P6DcTogBbT84zMg3svtjPei+2M//SOMvn4LYNXotvT2boU3sDUY3sxM2230Tq6D9F90SmCCJ/21k7kLZFKTW5bS/IWScF1H1Sk+7mhaXpNI0SKG5OAi9206xgxazYU3rpHSz63fRMAYNiRX4y37bawLPMCU6i71WqTRCTCD3f1wMrEWIjdzR/t5CbXzpvANR2oSVNuqaqqKqxYsQLffPMNkpOTUW/2NF8iclbGRpsDQExMDDZtavwbKBKJsGDBAvz973/Hpk2bcOrUKTzxxBPw8PDAww8/bK+3QQ6i6Z4v9XIRKqrNyx3anpJvT+NLdMFpfABgPMGDA0hNTYVSqcSQIUM024YOHQqlUokDBw6gZ8+eRuvIy8vD1q1bsWZN66kV33zzDb7++msEBQVh4sSJWLJkCby9DedDIdc3NjYI7/1yCXsvFKC6rt4hE8wdzihGZU09Ar3l6B3Kb3HI+tzce6K6dAsAwEPpq9nu7uWNob5eeL1bGF65lK13/+bhFm83CWYG614BzjQCmqZ0iQCElN40WBoAIuL6YfLAGXjnoSkAALcWuZ/cWkQ84uLisG/fPpNaEylqMWLDhATmllDE+qN82UMQqssAAA+9OgQb3tyCktszEmKGj9a77+wDv6Ch4AK8KsoAWW+95dyzr+CxA/9DZWUZorIvA36mt48pqKm9bNy4Ec8//zyAxtEOf/jDHzSv+fj44LHHHkPXrl0REBCAoqIiBAQEoKCgAEVFjUtr+/v7Izg4GGFhYRg5ciQkEsf7u05Exr366qta93QDBgwAAOzevVvzxd758+ehVt+ZIv6nP/0Jt27dwnPPPYfi4mIMGTIEP//8M+/5CN0CvdDZ3wOZNyuRcrEQE+LMH71uaw0NApJvTzUc46KDEZwiMJWbm4vAwNYfQGBgoMGVFJpbs2YNvL29MW3aNK3tjzzyiGZ456lTp7B48WIcP35ca6hnS9XV1aiuvpN3iCs0uKY+YUp08pajoKwah67cxKgejhed3t1sNT5DyRuJLKXwkmL8k68hbetmJD39e+CUdvLweJ87eaasFaCwSi6p29ykMojEYkxZ9BdUlZdD6emLPf/bprd8YmKi0cDUIxs/wY3gzuh3Uzthu0wmg1IqRY1aDfdbVbp31vd7auj3VyKCUH3n74xvoAe8A9w1gSljvCvKjJaRlt5E4MkDmhFpAOCGtq2A2BZdPSybrkGua+PGjXjwwQf1rsZXWlqKDz74wOT6/Pz8MHnyZISHhwNoDFoFBgYyoEXkBFavXo3Vq1cbLNOyrxCJRFi6dCmWLl1qu4aRUxKJRBgbG4hV+69i19k8hwxMnchW42ZFDbzlboiPNOPbQydi18DU0qVL8dprrxksc/jwYQCtc5sAxldSaO6LL77AI4880mpu8ty5czX/j4uLQ/fu3TFo0CAcPXoUAwcO1FnXsmXLjLabnJ9YLMLYmECsPXwdu87mOVxgShAEzTRDV42ck/1MWdAfh76/gsRHekIV7o3o/k0rABpY1a6dgqNTNn+PvL6JCAm7B1JIsE6WCmOLtnUf3DidLy/P8NRcNzfjfxZD87MQmp8FcYtk8ADwWHQ0cl9/o9UKel7jxqJ85y74/+YJo/U3ibpxHVdDIzA9yLYXIH3HTcSFQ/u1glIRsX7ocuoCbLn0w7zIQHzXIg/WQ5s/x9Ve8Xhh1OPIvL1NwhX5Orz6+nrMnz9fb1DKEsXFxfjyyy/N3s/cgBaDXUREjm9cbBBW7b+K3efz0dAgQOxg1x5N93wje6gglThFNiaz2TUwNW/ePKMr4EVFReHEiRM6byYKCgoMrqTQJCUlBefPn8e6deuMlh04cCCkUikuXryoNzC1ePFiLFy4UPO8tLSUy4e6qLGxQVh7+Dp2ns3H0immB0Lbw4W8clwtqoTMTexwQTNyfhEx/oiI0T/trunvtczGf7glDfXwEYtQXlcPr+rGXEaKqipE/foTPEf5Ybb/IIyzsO4ALxkya80LvYiMrA4nFolaBaUAIHzFCtRkZkIWrT/5d0sfvfUK5AdS0cvTHefMaqVubmLdOXgi+/bHtMWvYeOyJZptU+YPQNHqwchPMZ7Dy1IxngpkjOqL6L2NKxmKBAHhudcQU1sBT8lvNeVarlJIHU9KSgqysrKMF2wHlga0DPHz88PUqVNx9913mxXY0vWaWCxGYmIiEhMTGewiIjLRXVH+8Ja7obC8BulZJRjY2bFGJf10unGY/D29jMc+nJVdA1MqlQoqlcpouYSEBKjVavz6668YfHuVo0OHDkGtVhtcSaHJypUrER8fj379+hkte/r0adTW1iIkJERvGblc3m6Jh8m+RnRTQe4mRnbJLZzPK0NMsOPkcdp+qrGDGtVdBU+5U8zKJRfw+86B2FVUitkhjUGr3l4KDCu+gfqMizjcb7h5lZkQcBAB2N8zEO9+9jkkWgEfAULlRdzwj9e3q1GhvgocK7DemCBDgWuRmxvkLRbeMEZRU41YL+utujIg9M43bP4hngjt7guFd2Oy9/CY1vmn/B99FG4BKoh+2WW1NrSk0PGtn2ezXGZNGJrq2HJyWq/86EqKi4tNmppkqjfeeAOenp6YMWOGxcGuTp06cTQXEXUYMjcxRvXshK0ncrDrbJ5DBaauFJTjQl453MQi3B3DwJRdxcbGYsKECZg7dy4+/fRTAMDTTz+NSZMmaSU+j4mJwbJly/DAAw9otpWWlmL9+vVYvnx5q3ovX76Mb775Bvfeey9UKhXOnDmDRYsWYcCAARg+3MwbLHJJCpkEw7up8Mu5fOw8k+dQgammyHlSb8ebB02u6+WuoXi5a6jmuUgkwpNZp3Hu1z3mB6ZaiHSX4VpVDSKzL2ltl4lFcGtaGr5FhOKjRwbi0IYjsDTDlRVnBrUvMxqeeOYa1B5ydBsyVrNNJBbhgUV3RgVLdSzBLXJzg3LyJIiM5NwSWekkhsX2RkR2H9zz9Dyr1Eeuw9CXhaRbRUWFVYJd3t7eSEpKQkJCglZAy5RRXAxwEZEzGRcbiK0ncrDzTD7+OD7G3s3R+Ol048yxhK4BUCosXz3a0TlFYApoXDnvhRdeQFJSEgBgypQprZJctlx9AQDWrl0LQRDw0EMPtapTJpNh165dePfdd1FeXo6IiAjcd999WLJkCf94ksb43kH45Vw+tp3Mxby7u9u7OQCA6zcrcSanFBKxCONiXTdyTs6m9biWhC4BAAC5m/H58JsGdMPHf3oF8pIr+gu1CILc2ycEhze2jtP0z7yA9M49sLiLduBWKm37H/Qu8YNxJe1XDLx3CrDVdlPdrMWjtg4e6jqHmoqsi4fSFzOXLGv9gp52M1jRcYwcORLh4eEOM52vIykrK8OGDRuwYcOGNtXTctVES6csMthFRLZyd88gSCUinM8rw6X8cnQL9LJ3kwAA228PRnDEpOzW5DSBKX9/f3z99dcGy+hKivn000/j6aef1lk+IiICe/bssUr7yHUl9QrGXzadwpmcUmQUViBa5WnvJmlGSw2O8oe/p8zOrSHSL8LfA/v/fLdJ3/CEusvw8M8/YFu/rm0+7tCMMxh6/QLixvxVa7u/vz9GjBjRaiGM5sJ8fZBdUgpZoe5E71MW/gUluTfgHxaBshUrgIOpbWiphcEiBw8yWcLcd/Tkk0/apB3keCQSCd59912Dq/KRYzN31URTtAx2MbBFRG2h9JBieDcVks8XYNvJHLww1v4DEnLUt3D8eglEItfOLwU4UWCKyF78PGUY3k2FvRcaO6nnx3Szd5M0+aXG93btDoqcg5usMedeYur/8MuISXgmQjsZf5ivnjxJFt5gijw8IFRWwnOE4amDEj31jxtnOF16bEggio+mQnw72Xqret3cEHB7RT6fCePbGJhyfYp+/SDppIIs3PJFQn4TpsIHmfla20xZQZFcx7Rp0/Ddd99h/vz5HDlFANoe7DJnFBcDWkQdw319QhwqMPXz7Wl8Azv7IdBb/5eqroBXdUQmuK9PMPZeKMDWE/YPTOWqq5CW2bjEOvNLkSMYPutR5F2+gD8O7o+3E3ohTG7idDkLR/103bYVlYcOwefeewEAkyZNwpYtWzBq1Cjs3bvXojqbixk2Eoc2roVfaHir1yT++lcqdHZT//gKfvrkXdw7b5HWdpGF+buaiOVydN+9G2jDjdxL0SFI9PfGg+mX29QWcm7Tpk3D1KlTkZKSgs2bN2P16tWtUjgQmcrSwFZ4eDjeffddTJs2zQatIiJ7SuoVjL9ITuJcrmNM59t2snHxj44wGIGBKSITJPUKxssOMp1vy/EbEARgUKQfQvWNRCFqR15+/njsn9aZoiFWKo2WkQYHQzl1quZ5fHw8YmJi4OnpaZXAlCoiEs988iUU3t6abRGff4bCTz9F6BtvtLl+R9Vt0BB0/fwbm+SiEhkY3dTdQ46LldWYFqRnBRyRCFKxCCP8vNGlIBtXOoWhX9Yl3WXJ5UkkEiQmJiIxMRHLly9HSkoKsrOzUVBQoDWdS9e0ruzsbGzZsgU3b96087sgZ5aVlYUHH3wQ3333HYNTRC5G6SHFiG4q7HaAUVM3Sm7hUEbj36v7+oYaKe38GJgiMoGfpwzDHGQ63/fHswEAUweE2a0NRNagaypEl+83A/Of0ruPvpCJp2frYHFbAixeftojo7xGjoTXyJEW16ePSFIFAPAZn6TZJg1r+++22MsLDeXl8IgfhLLt201vjx1yV20f1AMZldXo7WU80P5pbGes3fo/PD7tfts3jBxeU5DKHPX19ZpgVl5enkkJuBnQIl0WLFiAqVOnclofkYu5t0+IQwSmthxvzHU6ONpff1oMF8LAFJGJJvUJwd4LBfgh/QaeS+xqlxu4S/nlOJVdCjexCPf14YpU5NwUCgWKa+q0tkmD23d6qr3SKHd6pi8qTxTAe8w8COWPQhp655swv9mzUZt9A57NAmGyyEjUXLsGiI2vbggA0Rs3QL1lC/wffdSswJQ9eEokiPP2MKlsnz590Lt3b4hNPA9ELVkSzAIsC2gx2OW6BEHA9evXkZKSYtHPExE5rubT+S7klaFHkLfxnWzg+/TGwNTU/q4/WgpgYIrIZOPjgvHX70/hfF4ZTt8oRVyY8SlH1vbD7cj5qB6duBofOT17BHcdhTxaCXn07T7ERzsoI5LJELT4z1rbIj79BPkr3oXqmdurzBpJHC/r3Bmdnn/eau21B5FUCqG2Fu69emltZ1CK7MHSgJYhzYNdTVMRTQ1s6Xpt//792LFjB8rKyqzaTtIvJyfH3k0gIitTekgxpmcgfj6Thw1pWVh8b2y7t+FiXhnO5JRCKhHh3riOMRiBgSkiEykVUozvHYwtx2/gu7Ssdg9MCYKAH9JvT+PrIJFzImoki4pC+Ip/W7SvyN0dQlUVvEZZfyoioH96ZVtFb9qI4m/XIqApGEfkYqwd7Fq4cKFVgl27d+/G+vXrUV5ebrW2uaqQkI5xw0jU0TwYH46fz+Rh47Fs/HF8T7hJ2vdLsabBCKN7dIJfBxmMwMAUkRkejA/HluM3sDk9G4vvjYHcrf3yChy5VoyrRZVQSCUYF+v6KzMQNXH3ss8QanM56giw7sm7UZOdDUXv3japXxYVZZN65d26IfiVv9qkbiJXZY1g15w5c/D5558jOTkZycnJaGhoaBXQMjaK68qVK/jyyy9ddtVEkUiE8PBwjLRB7kEisr8xMYEI8JShoKwaKRcLMSYmsN2OXd8gYENaFgBgSv+Ok1OYgSkiM4zopkKQjxx5pdXYfS4fE9pxaOW3v2YCACb3C4GnnL+65Pq6DhqCIffPhNzDslUwfX19rdsgJyXx9YWiDedCGh4O3L7x1MXjrruA3GKL6ycixyORSDB27FiMHTvW4jr+/e9/61010ZIpi44W7FqxYgUTnxO5KKlEjKn9w/DF/gx8l5bVroGpvRcLcENdBV8PKZJ6dZzBCLy7JTKDRCzCAwPC8cmey1h/JKvdAlPqylpsPdGYx2D24M7tckwiewuM6oKQ7j0t3n/27NlWbE3HJfHxMRiYErXjyFEich62yMtlKNjVXoGtiIgIrFixAtOmTbPGWyIiBzU9vjEwteNMHkoqa+Dr0T5T6tbeHozwwIAwuEs7zjUWA1NEZnowvjEwtft8PrKKKxHuZ9pqUm2xOT0b1XUN6BnkjQERvjY/HpGz69SpE1QqVbsf1zspCbmv/Q0eQ4a0+7GJiFxdW4NdloziUiqV+O1vfwsA2Lx5MyZNmsSRUkQdQO9QJXqF+OBMTinWH8nC3FFdbH7M/NIq7DybDwB4qIMNRmBgishM3QK9MLxbAPZfKsLXBzPx54kxNj2eIAiaaXyzB0c4bB4bIkfQdJPRq8VKbvp0Vljv2y+FQgG3gAD0TD8GkVxutXqJiMg6LA1sLViwAKWlpYiJiWFQiqgDeSwhEn/eeBJfHryK346IhkRs2/uw9WlZqG8QMLCzL3oEOUeOVWvhmstEFng8IQoAsPZwJqpq6216rINXbuJcbhncpWI8MKDjJMAj1ydYsI/I3d3g608++SRmzZqFUaNGmVRfP28PvB/bGZsHdLOgNY0eeOAB9OzZE8OGDQMAiN3dGUAmInIhAQEBAKAZUUVEHcPU/mFQKqS4fvMWdp/Lt+mxausb8PXBawA63mgpgIEpIouMjQ1CuJ8CJZW1muU8bWXlvisAgOkDw9ttbjORo3IzMj3Pw8MDsbGxZn2jPSPYH0N9vSxuU79+/fDQQw9B3o6jpEY+/AQAIP6++9vtmEREHRUDU0Qdk0Imwey7IgAAa1Kv2vRY207mIEddBZWXHFP6h9r0WI6IgSkiC0jEIswZGgkA+GJfBgTBkrEfxl0uKMfOs/kQiYAnR0Tb5BhEzkQk5p8tAAjp3hPzv9qIxMeesndTIAJHhxGRa2NgiqjjenRoJEQiIOViIc7nltnkGIIg4POUxsEIjydEQt4BF5bhFT6RhWbf1Rlecjecyy3TJKmztpX7MgAAY2OC0KWT5SM6iBzdfffdZ+8mOB03GUdQEhG1h6bFNBiYIup4Ivw9MDEuGADw4e5LNjnGwSs3cSq7FO5SMR65Pfiho2FgishCSg8p5iQ0dhzv/3LR6qOmsoorsf7IdQDA3JEcLUWu7a677rJ3E4iIiHTiiCmiju35MY25SH88cQNXCsqtWrcgCFix8wKAxtXf/T075hePDEwRtcFTI6KhkEpwIkuNPRcKrFr3B79cQm29gGFdAzCkS4BV6yYiIiIi0zQFpgoLC+3cEiKyh96hSoyLDUSDAHy4+7JV6z5wuQiHMm5CJhHjuUTLF+NxdgxMEbVBgJccjwxpXDXhnz+dR32DdUZNXSuqwPq0LADAoqQeVqmTyNG82rUxseNT4boTmtsodRsREZFZOGKKiH5/d3cAwKZjWVbLNSUIApb/fB4A8PCQzgj1VVilXmfEwBRRGz03phu83d1w+kYpNtwOJrXV37edRX2DgNE9OiE+0t8qdRI5mvuD/HBqeBxe7xZm76aQiYKifezdBCKidsfAFBH1i/DFxLhgNAjA6z+esUoal60nc3A0swRyNzGeS+xqhVY6LwamiNrI31OG+WMbI+hv/3QeZVW1bapvz4UC/HQ6DxKxCC/fF2uNJhI5LJXMDSIRV3VzFtP/GG/vJhARtTsGpogIABZPjIVMIsa+S4XY1cbFryqq6/Dm1rMAgOcSuyHQx90aTXRaDEwRWcFjCVGICvBAYXk1lv3vnMX13Kqpx9IfTgMAnhgWhR5B3tZqIhFRm4nErYOIg5SedmgJEVH74ap8RAQAnQM88JsRUQCAV78/hdI2DEh4d9dF5KirEOGvwDOju1iphc6LgSkiK5C5ifH3aX0AAP93KBN7LUyEvux/Z5FRWIFAbznmj+tuzSYSEdnE5E5KfNY7CgeHcoQnEbmm5iOmrL0KMxE5lxfu7o7O/h64oa7Cmz+etaiOg1eK8HnKFQDA0sm94S6VWLOJTomBKSIrGdZVhccTIgEAi9YfR666yqz9fz6diy9TrwEA/jWjH3zcpVZvIxGRuXx8DOeVEolEmBLoiyiFvJ1aRETUvpoCUzU1NaioqLBza4jInjzlbvjXjH4QiYB1R67j+/Rss/YvKq/Gov8ehyAAswZFYGxskI1a6lycJjD15ptvYtiwYfDw8ICvr69J+wiCgKVLlyI0NBQKhQKJiYk4ffq0Vpnq6mr8/ve/h0qlgqenJ6ZMmYKsLOsksKaO56WJMegZ5I2Csmo8/dURVNbUmbTf6RtqLFiXDgD47fBojOrRyYatJCIyXVJSEnr37o1HH33U3k0hAB999BGio6Ph7u6O+Ph4pKSkGCy/Z88exMfHw93dHV26dMEnn3zSqsyGDRvQq1cvyOVy9OrVC5s2bTL7uKZccxE5K7lcDqm08QvDt99+G1999RVWrFiBr776Cu+88w5efvllvPzyy1i+fLnO11555RXs2rUL9fX1dn4nRGQNg6P9NcnK//TdCaRfLzFpv6raejz9VRqyS24hKsADr0zuZcNWOhc3ezfAVDU1NZgxYwYSEhKwcuVKk/Z5++238c4772D16tXo0aMH3njjDdxzzz04f/48vL0bc/csWLAAW7Zswdq1axEQEIBFixZh0qRJSEtLg0TCIXVkHg+ZG/7z+CBM+WAfTmSp8cSqw/jiibvgJdf/q3YqW43HvvgVlTX1GNldhcX3xrRji4mIDPP09MSMGTPs3QwCsG7dOixYsAAfffQRhg8fjk8//RQTJ07EmTNn0Llz51blMzIycO+992Lu3Ln4+uuvsX//fjz33HPo1KkTpk+fDgBITU3FrFmz8Prrr+OBBx7Apk2bMHPmTOzbtw9Dhgwx+bimXHMROaONGzdi/vz5qK1tzCXz+uuvW1TPG2+8gYCAAHz22WeYNm2aNZtIRHaw8J6eOJdThl3n8vHYykNY89vBGNDZT2/5ypo6PPNVGtKuFcPH3Q3/edzwPWJHIxKcbKL06tWrsWDBApSUlBgsJwgCQkNDsWDBArz00ksAGkdHBQUF4a233sIzzzwDtVqNTp064auvvsKsWbMAADdu3EBERAS2bduG8ePHm9Sm0tJSKJVKqNVqo1MeqGM4mlmMx1f+irLqOnQP9MJ7Dw1AbIj2z4YgCNicno2/bjqFipp69AlT4uunhkCpcK4pfPz5dw7O9DktnzUJADB0+kMYPvMRAMDSpUsBAJ06dcLzzz9vr6aRE3Kmn31jhgwZgoEDB+Ljjz/WbIuNjcX999+PZcuWtSr/0ksv4YcffsDZs3dyYDz77LM4fvw4UlNTAQCzZs1CaWkp/ve//2nKTJgwAX5+fvj2229NOq4p11zGuNLnRK5j48aNePDBB62eV2rDhg1awSn+/DsHfk7UUnl1HZ744lccuVYMd6kYSyf3xqy7IlqtOH0pvwwvfJuOMzmlUEgl+OKJu5DQNcBOrbaMrX/+nWYqn7kyMjKQm5uLpKQkzTa5XI7Ro0fjwIEDAIC0tDTU1tZqlQkNDUVcXJymDJElBnb2wzdzh6CTtxwX88tx33speP7/jmLTsSwkn8/H6v0ZuP+jA/jDuuOoqKlHQpcAfDPX+YJSRO0lLi4OADBy5Eg7t4TIPmpqapCWlqZ1zQI0TrXUd82Smpraqvz48eNx5MgRzegPfWWa6jTluKZccxE5m/r6esyfP98myc7nz5/PaX1ELsBL7oY1vx2M0T06oaq2AX/eeBKT3t+HlfsykHw+H9+nZ2PB2mMYvyIFZ3JK4e8pw9dPDXa6oFR7cNmxY7m5uQCAoCDtZGJBQUG4du2apoxMJoOfn1+rMk3761JdXY3q6mrN89LSUms1m1xI33Bf/G/+SPx10ylsP52LrSdysPVEjlYZhVSC5xK74rkx3SDRsQw7ub4333wTW7duRXp6OmQymc7RoC2/dQGAjz/+GM8++2w7tNB+5AqF5v/Tpk3DuHHjTM4xSORqCgsLUV9fr/O6Rt81S25urs7ydXV1KCwsREhIiN4yTXWaclxTrrla4rUUObqUlBSb5Z3NyspCSkoKEhMTbVI/EbUfT7kbVj1xFz5PuYL3dl3E6RulOH3jTKtyY2MC8fdpfRDk426HVjo+uwamli5ditdee81gmcOHD2PQoEEWH6PlDZ0gCDpv8swps2zZMqPtJgIAlZccn8yJx6lsNb5Pz8bx62qUVtUiyMcdw7oGYNrAcHTy5kpWHZmp+fNWrVqFCRMmaJ4rlcr2aJ5d3PP0PFw+cgj9xt+n2SYWixmUIoL51zW6yrfcbkqd1irThNdS5OhycnKMF3Lg+omo/YjFIjwzuitmDIrAxqNZ2HepEHml1fCSSxAXpsTU/mHoH+Fr72Y6NLsGpubNm4fZs2cbLBMVFWVR3cHBwQAav8ULCQnRbM/Pz9d8oxccHIyamhoUFxdrjZrKz8/HsGHD9Na9ePFiLFy4UPO8tLQUERERFrWTOoa4MCXiwlw3kECWa7oxW716tcFyvr6+mn7N1fUdOwF9x04wXpCoA1GpVJBIJK1GRzW/rmkpODhYZ3k3NzcEBAQYLNNUpynHNeWaqyVeS5Gja/6z7Iz1E1H78/eU4amRXfDUyC72borTsWuOKZVKhZiYGIMPd3fLhrpFR0cjODgYO3bs0GyrqanBnj17NEGn+Ph4SKVSrTI5OTk4deqUwcCUXC6Hj4+P1oOIyJbmzZsHlUqFu+66C5988gkaGhoMlq+urkZpaanWg4icl0wmQ3x8vNY1CwDs2LFD7zVLQkJCq/I///wzBg0aBKlUarBMU52mHNeUa66WeC1Fjm7kyJEIDw+3Sd3h4eHMmUhE1IzT5JjKzMzEzZs3kZmZifr6eqSnpwMAunXrBi8vLwBATEwMli1bhgceeAAikQgLFizA3//+d3Tv3h3du3fH3//+d3h4eODhhx8G0DgV5sknn8SiRYsQEBAAf39/vPjii+jTpw/GjRtnr7dKRKTl9ddfx9ixY6FQKLBr1y4sWrQIhYWF+Otf/6p3H06TIXI9CxcuxJw5czBo0CAkJCTgs88+Q2Zmpibf3OLFi5GdnY0vv/wSQOMKfB988AEWLlyIuXPnIjU1FStXrtSstgc0JmEeNWoU3nrrLUydOhXff/89du7ciX379pl8XFOuuYicjUQiwbvvvmuTVfneffddSCQSq9ZJROTUBCfx+OOPCwBaPXbv3q0pA0BYtWqV5nlDQ4OwZMkSITg4WJDL5cKoUaOEkydPatV769YtYd68eYK/v7+gUCiESZMmCZmZmWa1Ta1WCwAEtVrdlrdI5JT489/akiVLdPZXzR+HDx/W2mfVqlWCUqk0qf5//etfgo+Pj8EyVVVVglqt1jyuX7/Oz4k6JFfroz788EMhMjJSkMlkwsCBA4U9e/ZoXnv88ceF0aNHa5VPTk4WBgwYIMhkMiEqKkr4+OOPW9W5fv16oWfPnoJUKhViYmKEDRs2mHVcQTDtmssQV/ucyHVs2LBBCA8PN/p33ZRHQECAzt8v/vyb54033hASEhIEhUJh8rWTrnvJIUOGmHVcfk7Ukdn6518kCDZYA7WDKS0thVKphFqt5lB06nD4899aYWEhCgsLDZaJiorSmqq8evVqLFiwQOeqfC3t378fI0aM0Lmalj78nKij4s++c+DnRI6svr4eKSkpyM7ORkFBAQICAlBUVISAgAAUFBSgqKgIAODv74/AwMBWr4nFYiQmJiIxMVHnSCn+/JtnyZIl8PX1RVZWFlauXGnStdMTTzyBvLw8rFq1SrNNJpPB39/f5OPyc6KOzNY//04zlY+IyFmoVCqoVCqb1X/s2DG4u7tzlToiIqJ2IJFIkJiYaO9m0G2mLhzTklwu7zALyRA5GwamiIjsyFj+vC1btiA3NxcJCQlQKBTYvXs3Xn75ZTz99NOQy+X2bTwRERGRk0hOTkZgYCB8fX0xevRovPnmmwgMDNRbvrq6GtXV1ZrnXEiGyHYYmCIisqNXX30Va9as0TwfMGAAAGD37t1ITEyEVCrFRx99hIULF6KhoQFdunTB3/72Nzz//PP2ajIRERGRU5k4cSJmzJiByMhIZGRk4JVXXsHdd9+NtLQ0vV/0cSEZovbDHFNWwPnG1JHx59858HOijoo/+86BnxN1ZPz5B5YuXWo0CHT48GEMGjRI89yc/Jwt5eTkIDIyEmvXrsW0adN0ltE1YioiIqJDf07UcTHHFBEREREREbmsefPmYfbs2QbLREVFWe14ISEhiIyMxMWLF/WWkcvlTJtA1E4YmCIiIiIiIiK7sfXCMS0VFRXh+vXrCAkJabdjEpF+Yns3gIiIiIiIiMgUmZmZSE9P11o4Jj09HeXl5ZoyMTEx2LRpEwCgvLwcL774IlJTU3H16lUkJydj8uTJUKlUeOCBB+z1NoioGY6YsoKmNF1cqYE6oqafe6arc2zsp6ijYh/lHNhHUUfGfso8xhaOAYDz589DrVYDACQSCU6ePIkvv/wSJSUlCAkJwZgxY7Bu3Tp4e3ubfFz2U9SR2bqfYvJzK8jKykJERIS9m0FkV9evX0d4eLi9m0F6sJ+ijo59lGNjH0XEfsrRsZ8isl0/xcCUFTQ0NODGjRvw9vaGSCTSWaZpFYfr169zFQcr47m1HVPOrSAIKCsrQ2hoKMRizg52VMb6Kf4e2RbPr+0YO7fso5wDr6Xsj+fXtgydX/ZTzoH9lH3x3NqOI9zzcSqfFYjFYpOjhj4+PvxFshGeW9sxdm6VSmU7toYsYWo/xd8j2+L5tR1D55Z9lOPjtZTj4Pm1LX3nl/2U42M/5Rh4bm3Hnvd8DMkTEREREREREZFdMDBFRERERERERER2wcBUO5HL5ViyZAnkcrm9m+JyeG5th+e24+BnbVs8v7bDc9tx8LO2LZ5f2+L57Rj4OdsOz63tOMK5ZfJzIiIiIiIiIiKyC46YIiIiIiIiIiIiu2BgioiIiIiIiIiI7IKBKSIiIiIiIiIisgsGptrBRx99hOjoaLi7uyM+Ph4pKSn2bpJDWbp0KUQikdYjODhY87ogCFi6dClCQ0OhUCiQmJiI06dPa9VRXV2N3//+91CpVPD09MSUKVOQlZWlVaa4uBhz5syBUqmEUqnEnDlzUFJS0h5vsd3s3bsXkydPRmhoKEQiETZv3qz1enuey8zMTEyePBmenp5QqVR44YUXUFNTY4u3TVbAfsow9lPWw36KLMV+qu2s0ZfRHe3Vn5FzYB9lGK+lrMvlrqcEsqm1a9cKUqlU+Pzzz4UzZ84I8+fPFzw9PYVr167Zu2kOY8mSJULv3r2FnJwczSM/P1/z+j/+8Q/B29tb2LBhg3Dy5Elh1qxZQkhIiFBaWqop8+yzzwphYWHCjh07hKNHjwpjxowR+vXrJ9TV1WnKTJgwQYiLixMOHDggHDhwQIiLixMmTZrUru/V1rZt2ya8/PLLwoYNGwQAwqZNm7Reb69zWVdXJ8TFxQljxowRjh49KuzYsUMIDQ0V5s2bZ/NzQOZjP2Uc+ynrYT9FlmA/ZR3W6Mvojvbqz8jxsY8yjtdS1uVq11MMTNnY4MGDhWeffVZrW0xMjPDnP//ZTi1yPEuWLBH69eun87WGhgYhODhY+Mc//qHZVlVVJSiVSuGTTz4RBEEQSkpKBKlUKqxdu1ZTJjs7WxCLxcL27dsFQRCEM2fOCACEgwcPasqkpqYKAIRz587Z4F3ZX8sOqj3P5bZt2wSxWCxkZ2drynz77beCXC4X1Gq1Td4vWY79lHHsp2yD/RSZiv2UdbS1LyP9bNWfkXNgH2Ucr6VsxxWupziVz4ZqamqQlpaGpKQkre1JSUk4cOCAnVrlmC5evIjQ0FBER0dj9uzZuHLlCgAgIyMDubm5WudQLpdj9OjRmnOYlpaG2tparTKhoaGIi4vTlElNTYVSqcSQIUM0ZYYOHQqlUtlhPov2PJepqamIi4tDaGiopsz48eNRXV2NtLQ0m75PMg/7KdOxn7I99lOkC/sp62pLX0ams1Z/Ro6PfZTpeC3VPpzxeoqBKRsqLCxEfX09goKCtLYHBQUhNzfXTq1yPEOGDMGXX36Jn376CZ9//jlyc3MxbNgwFBUVac6ToXOYm5sLmUwGPz8/g2UCAwNbHTswMLDDfBbteS5zc3NbHcfPzw8ymazDnG9nwX7KNOyn2gf7KdKF/ZT1tLUvI9NZqz8jx8c+yjS8lmo/zng95WZySbKYSCTSei4IQqttHdnEiRM1/+/Tpw8SEhLQtWtXrFmzBkOHDgVg2TlsWUZX+Y74WbTXueT5di7spwxjP9W+2E+RLuyn2s5WfRnpZ43+jJwDf3cM47VU+3Om6ymOmLIhlUoFiUTSKlKYn5/fKqpId3h6eqJPnz64ePGiZqUGQ+cwODgYNTU1KC4uNlgmLy+v1bEKCgo6zGfRnucyODi41XGKi4tRW1vbYc63s2A/ZRn2U7bBfop0YT9lO+b2ZWQ6a/Vn5PjYR1mG11K244zXUwxM2ZBMJkN8fDx27NihtX3Hjh0YNmyYnVrl+Kqrq3H27FmEhIQgOjoawcHBWuewpqYGe/bs0ZzD+Ph4SKVSrTI5OTk4deqUpkxCQgLUajV+/fVXTZlDhw5BrVZ3mM+iPc9lQkICTp06hZycHE2Zn3/+GXK5HPHx8TZ9n2Qe9lOWYT9lG+ynSBf2U7Zjbl9GprNWf0aOj32UZXgtZTtOeT1lcpp0skjT0qErV64Uzpw5IyxYsEDw9PQUrl69au+mOYxFixYJycnJwpUrV4SDBw8KkyZNEry9vTXn6B//+IegVCqFjRs3CidPnhQeeughnUtdhoeHCzt37hSOHj0q3H333TqXuuzbt6+QmpoqpKamCn369HG5pUPLysqEY8eOCceOHRMACO+8845w7NgxzVK17XUum5YNHTt2rHD06FFh586dQnh4OJdhd1Dsp4xjP2U97KfIEuynrMMafRnd0V79GTk+9lHG8VrKulzteoqBqXbw4YcfCpGRkYJMJhMGDhwo7Nmzx95NciizZs0SQkJCBKlUKoSGhgrTpk0TTp8+rXm9oaFBWLJkiRAcHCzI5XJh1KhRwsmTJ7XquHXrljBv3jzB399fUCgUwqRJk4TMzEytMkVFRcIjjzwieHt7C97e3sIjjzwiFBcXt8dbbDe7d+8WALR6PP7444IgtO+5vHbtmnDfffcJCoVC8Pf3F+bNmydUVVXZ8u1TG7CfMoz9lPWwnyJLsZ9qO2v0ZXRHe/Vn5BzYRxnGaynrcrXrKZEgCILp46uIiIiIiIiIiIisgzmmiIiIiIiIiIjILhiYIiIiIiIiIiIiu2BgioiIiIiIiIiI7IKBKSIiIiIiIiIisgsGpoiIiIiIiIiIyC4YmCIiIiIiIiIiIrtgYIqIiIiIiIiIiOyCgSkiIiIiIiIiIrILBqaowxOJRNi8ebO9m0FEpBP7KCIiIqK24fWUY2NgitpFfn4+nnnmGXTu3BlyuRzBwcEYP348UlNT7d00IiL2UUTUIZjS1x07dgwzZsxAUFAQ3N3d0aNHD8ydOxcXLlywY8uJyBnweoos5WbvBlDHMH36dNTW1mLNmjXo0qUL8vLysGvXLty8edPeTSMiYh9FRB2Csb7uxx9/xPTp0zF+/Hh888036Nq1K/Lz87F+/Xq88sorWLdunZ3fARE5Ml5PkcUEIhsrLi4WAAjJycl6yyxfvlyIi4sTPDw8hPDwcOF3v/udUFZWpnl91apVglKpFLZs2SL06NFDUCgUwvTp04Xy8nJh9erVQmRkpODr6yvMmzdPqKur0+wXGRkp/O1vfxMeeughwdPTUwgJCRHee+89rWMDEDZt2qR5npWVJcycOVPw9fUV/P39hSlTpggZGRlWOx9E5FjYRxFRR2Csr6uoqBBUKpVw//33692fiEgfXk9RW3AqH9mcl5cXvLy8sHnzZlRXV+ssIxaL8d577+HUqVNYs2YNfvnlF/zpT3/SKlNZWYn33nsPa9euxfbt25GcnIxp06Zh27Zt2LZtG7766it89tln+O6777T2++c//4m+ffvi6NGjWLx4Mf7whz9gx44dOttRWVmJMWPGwMvLC3v37sW+ffvg5eWFCRMmoKamxjonhIgcCvsoIuoIjPV1P/30EwoLC1v1bU18fX1t3EIicma8nqI2sXdkjDqG7777TvDz8xPc3d2FYcOGCYsXLxaOHz+ut/x///tfISAgQPN81apVAgDh0qVLmm3PPPOM4OHhoRVlHz9+vPDMM89onkdGRgoTJkzQqnvWrFnCxIkTNc/RLHq+cuVKoWfPnkJDQ4Pm9erqakGhUAg//fST+W+ciJwC+ygi6ggM9XVvvfWWAEC4efOmnVtJRM6K11NkKY6YonYxffp03LhxAz/88APGjx+P5ORkDBw4EKtXrwYA7N69G/fccw/CwsLg7e2Nxx57DEVFRaioqNDU4eHhga5du2qeBwUFISoqCl5eXlrb8vPztY6dkJDQ6vnZs2d1tjMtLQ2XLl2Ct7e3Jurv7++PqqoqXL58ua2ngYgcFPsoIuoIDPV1giDYu3lE5OR4PUWWYmCK2o27uzvuuecevPrqqzhw4ACeeOIJLFmyBNeuXcO9996LuLg4bNiwAWlpafjwww8BALW1tZr9pVKpVn0ikUjntoaGBqNtEYlEOrc3NDQgPj4e6enpWo8LFy7g4YcfNvctE5ETYR9FRB2Bvr6uR48eAIBz587ZuYVE5Mx4PUWW4Kp8ZDe9evXC5s2bceTIEdTV1WH58uUQixtjpf/973+tdpyDBw+2eh4TE6Oz7MCBA7Fu3ToEBgbCx8fHam0gIufDPoqIOoKmvi4pKQkqlQpvv/02Nm3a1KpcSUkJ80wRkdl4PUWm4IgpsrmioiLcfffd+Prrr3HixAlkZGRg/fr1ePvttzF16lR07doVdXV1eP/993HlyhV89dVX+OSTT6x2/P379+Ptt9/GhQsX8OGHH2L9+vWYP3++zrKPPPIIVCoVpk6dipSUFGRkZGDPnj2YP38+srKyrNYmInIc7KOIqCMw1td5enriP//5D7Zu3YopU6Zg586duHr1Ko4cOYI//elPePbZZ+39FojIgfF6itqCI6bI5ry8vDBkyBD8+9//xuXLl1FbW4uIiAjMnTsXf/nLX6BQKPDOO+/grbfewuLFizFq1CgsW7YMjz32mFWOv2jRIqSlpeG1116Dt7c3li9fjvHjx+ss6+Hhgb179+Kll17CtGnTUFZWhrCwMIwdO5bRdCIXxT6KiDoCY30dAEydOhUHDhzAsmXL8PDDD6O0tBQRERG4++678cYbb9j5HRCRI+P1FLWFSGCmQ3JhUVFRWLBgARYsWGDvphARtcI+ioiIiKhteD3l/DiVj4iIiIiIiIiI7IKBKSIiIiIiIiIisgtO5SMiIiIiIiIiIrvgiCkiIiIiIiIiIrILBqaIiIiIiIiIiMguGJgiIiIiIiIiIiK7YGCKiIiIiIiIiIjsgoEpIiIiIiIiIiKyCwamiIiIiIiIiIjILhiYIiIiIiIiIiIiu2BgioiIiIiIiIiI7IKBKSIiIiIiIiIisov/Bz7a/FgY65jKAAAAAElFTkSuQmCC", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "# Plot the results\n", + "fig, ax = plt.subplots(1, 4, figsize=(12, 4))\n", + "ax[0].plot(target_signal)\n", + "ax[0].set_title('Target')\n", + "ax[0].set_xlabel('Sample')\n", + "ax[0].set_ylabel('Amplitude')\n", + "ax[1].plot(data_matrix)\n", + "ax[1].set_title('Target + Noise')\n", + "ax[1].set_ylabel('Amplitude')\n", + "ax[1].set_xlabel('Sample')\n", + "ax[2].plot(variance, 'o-k')\n", + "ax[2].set_xlabel('SC')\n", + "ax[2].set_ylabel('Variance')\n", + "ax[3].plot(x_recovered)\n", + "ax[3].set_title('Recovered')\n", + "ax[3].set_ylabel('Amplitude')\n", + "ax[3].set_xlabel('Sample')\n", + "plt.tight_layout()" + ] + } + ], + "metadata": { + "kernelspec": { + "display_name": "eelbrain", + "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": "3.8.16" + }, + "orig_nbformat": 4 + }, + "nbformat": 4, + "nbformat_minor": 2 +} diff --git a/examples/example_mcca_2.py b/examples/example_mcca_2.py new file mode 100644 index 00000000..585a3851 --- /dev/null +++ b/examples/example_mcca_2.py @@ -0,0 +1,96 @@ + +# # Example 1 - sinusoidal target in separable noise. +# reproduced from de Cheveigné et al. (2018). +# +# Synthetic data for this example consisted of 10 data matrices, each of dimensions 10000 samples +# 10 channels. Each was obtained by multiplying 9 Gaussian noise time series (independent and uncorrelated) by a 9 x 10 mixing matrix with random Gaussian coefficients. To this background of noise was added a “target” consisting of a sinusoidal time series multiplied by a 1 x 10 mixing matrix with random coefficients. The target was the same for all data matrices, but the mixing matrices differed, as did the noise matrices. The SNR was set to 10−20, i.e. a very unfavorable SNR. The noise is of rank 9 and the signal of rank 1, so signal and noise are in principle linearly separable. + +# %% +import numpy as np +import matplotlib.pyplot as plt +from scipy import signal + +# %% +# Set the seed for the random number generator for reproducibility +np.random.seed(5) + +# Constants +num_matrices = 10 +num_samples = 10000 +num_channels = 10 +noise_rank = 9 +signal_rank = 1 +unfavorable_SNR_dB = -20 # SNR in decibels + +# Generate noise matrices and mixing matrices +noise_matrices = [np.random.normal(size=(num_samples, noise_rank)) for _ in range(num_matrices)] +mixing_matrices = [np.random.normal(size=(noise_rank, num_channels)) for _ in range(num_matrices)] + +# Generate sinusoidal target +t = np.linspace(0, 1, num_samples) +target_signal = np.sin(2 * np.pi * t) # 1 Hz sinusoidal signal + +# Generate signal mixing matrix +signal_mixing_matrix = np.random.normal(size=(signal_rank, num_channels)) + +# Prepare data matrices +data_matrices = [] +for i in range(num_matrices): + # Create noise for current data matrix + noise = np.matmul(noise_matrices[i], mixing_matrices[i]) + + # Create signal for current data matrix + signal = np.matmul(target_signal.reshape(-1, 1), signal_mixing_matrix) + + # Adjust the power of signal to achieve the desired SNR + noise_power = np.mean(noise**2) + signal_power = 10**(unfavorable_SNR_dB / 10) * noise_power + signal = np.sqrt(signal_power / np.mean(signal**2)) * signal + + # Add signal and noise + data_matrix = signal + noise + + data_matrices.append(data_matrix) + +# %% +# Concatenate data matrices +x = np.concatenate(data_matrices, axis=-1) + +# %% +from meegkit import cca + +# Compute Covariance matrix +C = np.dot(x.T, x) + +# Compute mCCA +A, score, AA = cca.mcca(C, 10) + +# %% +# Compute the recovered signal using first SC +x_recovered = x.dot(A)[:,0] +# Normalize the recovered signal +x_recovered = x_recovered / x_recovered.std() +# Compute variance across SCs +variance = np.var(x.dot(A), axis=0) + +# %% +# Plot the results +fig, ax = plt.subplots(1, 4, figsize=(12, 4)) +ax[0].plot(target_signal) +ax[0].set_title('Target') +ax[0].set_xlabel('Sample') +ax[0].set_ylabel('Amplitude') +ax[1].plot(data_matrix) +ax[1].set_title('Target + Noise') +ax[1].set_ylabel('Amplitude') +ax[1].set_xlabel('Sample') +ax[2].plot(variance, 'o-k') +ax[2].set_xlabel('SC') +ax[2].set_ylabel('Variance') +ax[3].plot(x_recovered) +ax[3].set_title('Recovered') +ax[3].set_ylabel('Amplitude') +ax[3].set_xlabel('Sample') +plt.tight_layout() + + From 284dc7b6d3d09c1a41be86fb2b5fbc2bd3bec32a Mon Sep 17 00:00:00 2001 From: John Kyle Cooper Date: Sun, 11 Jun 2023 15:24:09 +0200 Subject: [PATCH 2/7] included doc string in example .py file --- examples/example_mcca_2.py | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) diff --git a/examples/example_mcca_2.py b/examples/example_mcca_2.py index 585a3851..d093057d 100644 --- a/examples/example_mcca_2.py +++ b/examples/example_mcca_2.py @@ -1,9 +1,12 @@ +""" +Example 1 - sinusoidal target in separable noise. +================================================= +reproduced from de Cheveigné et al. (2018). -# # Example 1 - sinusoidal target in separable noise. -# reproduced from de Cheveigné et al. (2018). -# -# Synthetic data for this example consisted of 10 data matrices, each of dimensions 10000 samples -# 10 channels. Each was obtained by multiplying 9 Gaussian noise time series (independent and uncorrelated) by a 9 x 10 mixing matrix with random Gaussian coefficients. To this background of noise was added a “target” consisting of a sinusoidal time series multiplied by a 1 x 10 mixing matrix with random coefficients. The target was the same for all data matrices, but the mixing matrices differed, as did the noise matrices. The SNR was set to 10−20, i.e. a very unfavorable SNR. The noise is of rank 9 and the signal of rank 1, so signal and noise are in principle linearly separable. +Synthetic data for this example consisted of 10 data matrices, each of dimensions 10000 samples +10 channels. Each was obtained by multiplying 9 Gaussian noise time series (independent and uncorrelated) by a 9 x 10 mixing matrix with random Gaussian coefficients. To this background of noise was added a “target” consisting of a sinusoidal time series multiplied by a 1 x 10 mixing matrix with random coefficients. The target was the same for all data matrices, but the mixing matrices differed, as did the noise matrices. The SNR was set to 10−20, i.e. a very unfavorable SNR. The noise is of rank 9 and the signal of rank 1, so signal and noise are in principle linearly separable. + +""" # %% import numpy as np From 503756bf60b9ebf4151bc8ec8de73a939f1ca562 Mon Sep 17 00:00:00 2001 From: John Kyle Cooper Date: Sun, 11 Jun 2023 15:52:40 +0200 Subject: [PATCH 3/7] fixed formatting according to flake8 --- examples/example_mcca_2.py | 53 ++++++++++++++++++++++---------------- 1 file changed, 31 insertions(+), 22 deletions(-) diff --git a/examples/example_mcca_2.py b/examples/example_mcca_2.py index d093057d..565df25f 100644 --- a/examples/example_mcca_2.py +++ b/examples/example_mcca_2.py @@ -3,15 +3,26 @@ ================================================= reproduced from de Cheveigné et al. (2018). -Synthetic data for this example consisted of 10 data matrices, each of dimensions 10000 samples -10 channels. Each was obtained by multiplying 9 Gaussian noise time series (independent and uncorrelated) by a 9 x 10 mixing matrix with random Gaussian coefficients. To this background of noise was added a “target” consisting of a sinusoidal time series multiplied by a 1 x 10 mixing matrix with random coefficients. The target was the same for all data matrices, but the mixing matrices differed, as did the noise matrices. The SNR was set to 10−20, i.e. a very unfavorable SNR. The noise is of rank 9 and the signal of rank 1, so signal and noise are in principle linearly separable. +Synthetic data for this example consisted of 10 data matrices, +each of dimensions 10000 samples x 10 channels. Each was obtained +by multiplying 9 Gaussian noise time series (independent and uncorrelated) +by a 9 x 10 mixing matrix with random Gaussian coefficients. +To this background of noise was added a “target” consisting of a sinusoidal +time series multiplied by a 1 x 10 mixing matrix with random coefficients. +The target was the same for all data matrices, +but the mixing matrices differed, +as did the noise matrices. The SNR was set to 10−20, +i.e. a very unfavorable SNR. +The noise is of rank 9 and the signal of rank 1, +so signal and noise are +in principle linearly separable. """ # %% +from meegkit import cca import numpy as np import matplotlib.pyplot as plt -from scipy import signal # %% # Set the seed for the random number generator for reproducibility @@ -26,8 +37,10 @@ unfavorable_SNR_dB = -20 # SNR in decibels # Generate noise matrices and mixing matrices -noise_matrices = [np.random.normal(size=(num_samples, noise_rank)) for _ in range(num_matrices)] -mixing_matrices = [np.random.normal(size=(noise_rank, num_channels)) for _ in range(num_matrices)] +noise_matrices = [np.random.normal(size=(num_samples, noise_rank)) + for _ in range(num_matrices)] +mixing_matrices = [np.random.normal(size=(noise_rank, num_channels)) + for _ in range(num_matrices)] # Generate sinusoidal target t = np.linspace(0, 1, num_samples) @@ -60,8 +73,6 @@ x = np.concatenate(data_matrices, axis=-1) # %% -from meegkit import cca - # Compute Covariance matrix C = np.dot(x.T, x) @@ -70,7 +81,7 @@ # %% # Compute the recovered signal using first SC -x_recovered = x.dot(A)[:,0] +x_recovered = x.dot(A)[:, 0] # Normalize the recovered signal x_recovered = x_recovered / x_recovered.std() # Compute variance across SCs @@ -80,20 +91,18 @@ # Plot the results fig, ax = plt.subplots(1, 4, figsize=(12, 4)) ax[0].plot(target_signal) -ax[0].set_title('Target') -ax[0].set_xlabel('Sample') -ax[0].set_ylabel('Amplitude') +ax[0].set_title("Target") +ax[0].set_xlabel("Sample") +ax[0].set_ylabel("Amplitude") ax[1].plot(data_matrix) -ax[1].set_title('Target + Noise') -ax[1].set_ylabel('Amplitude') -ax[1].set_xlabel('Sample') -ax[2].plot(variance, 'o-k') -ax[2].set_xlabel('SC') -ax[2].set_ylabel('Variance') +ax[1].set_title("Target + Noise") +ax[1].set_ylabel("Amplitude") +ax[1].set_xlabel("Sample") +ax[2].plot(variance, "o-k") +ax[2].set_xlabel("SC") +ax[2].set_ylabel("Variance") ax[3].plot(x_recovered) -ax[3].set_title('Recovered') -ax[3].set_ylabel('Amplitude') -ax[3].set_xlabel('Sample') +ax[3].set_title("Recovered") +ax[3].set_ylabel("Amplitude") +ax[3].set_xlabel("Sample") plt.tight_layout() - - From 4db03166174224e7484115433b838d935b65b6da Mon Sep 17 00:00:00 2001 From: John Kyle Cooper Date: Sun, 11 Jun 2023 15:55:37 +0200 Subject: [PATCH 4/7] Removed unecessary import --- examples/example_mcca_2.ipynb | 31 +++++++++---------------------- 1 file changed, 9 insertions(+), 22 deletions(-) diff --git a/examples/example_mcca_2.ipynb b/examples/example_mcca_2.ipynb index 3983364c..7de34f75 100644 --- a/examples/example_mcca_2.ipynb +++ b/examples/example_mcca_2.ipynb @@ -8,24 +8,22 @@ "# Example 1 - sinusoidal target in separable noise.\n", "reproduced from de Cheveigné et al. (2018).\n", "\n", - "Synthetic data for this example consisted of 10 data matrices, each of dimensions 10000 samples \n", - " 10 channels. Each was obtained by multiplying 9 Gaussian noise time series (independent and uncorrelated) by a 9 x 10 mixing matrix with random Gaussian coefficients. To this background of noise was added a “target” consisting of a sinusoidal time series multiplied by a 1 x 10 mixing matrix with random coefficients. The target was the same for all data matrices, but the mixing matrices differed, as did the noise matrices. The SNR was set to 10−20, i.e. a very unfavorable SNR. The noise is of rank 9 and the signal of rank 1, so signal and noise are in principle linearly separable." + "Synthetic data for this example consisted of 10 data matrices, each of dimensions 10000 samples x 10 channels. Each was obtained by multiplying 9 Gaussian noise time series (independent and uncorrelated) by a 9 x 10 mixing matrix with random Gaussian coefficients. To this background of noise was added a “target” consisting of a sinusoidal time series multiplied by a 1 x 10 mixing matrix with random coefficients. The target was the same for all data matrices, but the mixing matrices differed, as did the noise matrices. The SNR was set to 10−20, i.e. a very unfavorable SNR. The noise is of rank 9 and the signal of rank 1, so signal and noise are in principle linearly separable." ] }, { "cell_type": "code", - "execution_count": 171, + "execution_count": null, "metadata": {}, "outputs": [], "source": [ "import numpy as np\n", - "import matplotlib.pyplot as plt\n", - "from scipy import signal" + "import matplotlib.pyplot as plt" ] }, { "cell_type": "code", - "execution_count": 172, + "execution_count": null, "metadata": {}, "outputs": [], "source": [ @@ -73,7 +71,7 @@ }, { "cell_type": "code", - "execution_count": 173, + "execution_count": null, "metadata": {}, "outputs": [], "source": [ @@ -83,7 +81,7 @@ }, { "cell_type": "code", - "execution_count": 174, + "execution_count": null, "metadata": {}, "outputs": [], "source": [ @@ -98,7 +96,7 @@ }, { "cell_type": "code", - "execution_count": 175, + "execution_count": null, "metadata": {}, "outputs": [], "source": [ @@ -112,20 +110,9 @@ }, { "cell_type": "code", - "execution_count": 176, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "data": { - "image/png": "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", - "text/plain": [ - "
" - ] - }, - "metadata": {}, - "output_type": "display_data" - } - ], + "outputs": [], "source": [ "# Plot the results\n", "fig, ax = plt.subplots(1, 4, figsize=(12, 4))\n", From 022bff26e9535b5c38a00757d082805f405baeac Mon Sep 17 00:00:00 2001 From: John Kyle Cooper Date: Sun, 11 Jun 2023 16:11:32 +0200 Subject: [PATCH 5/7] fixed syntax to comply with flake8 --- examples/example_mcca_2.ipynb | 18 +++++----- examples/example_mcca_2.py | 62 +++++++++++++++++------------------ 2 files changed, 40 insertions(+), 40 deletions(-) diff --git a/examples/example_mcca_2.ipynb b/examples/example_mcca_2.ipynb index 7de34f75..556a8fd4 100644 --- a/examples/example_mcca_2.ipynb +++ b/examples/example_mcca_2.ipynb @@ -17,8 +17,13 @@ "metadata": {}, "outputs": [], "source": [ + "import matplotlib.pyplot as plt\n", "import numpy as np\n", - "import matplotlib.pyplot as plt" + "\n", + "from meegkit import cca\n", + "\n", + "# Set the seed for the random number generator for reproducibility\n", + "rng = np.random.default_rng(5)" ] }, { @@ -27,9 +32,6 @@ "metadata": {}, "outputs": [], "source": [ - "# Set the seed for the random number generator for reproducibility\n", - "np.random.seed(5)\n", - "\n", "# Constants\n", "num_matrices = 10\n", "num_samples = 10000\n", @@ -39,15 +41,15 @@ "unfavorable_SNR_dB = -20 # SNR in decibels\n", "\n", "# Generate noise matrices and mixing matrices\n", - "noise_matrices = [np.random.normal(size=(num_samples, noise_rank)) for _ in range(num_matrices)]\n", - "mixing_matrices = [np.random.normal(size=(noise_rank, num_channels)) for _ in range(num_matrices)]\n", + "noise_matrices = [rng.normal(size=(num_samples, noise_rank)) for _ in range(num_matrices)]\n", + "mixing_matrices = [rng.normal(size=(noise_rank, num_channels)) for _ in range(num_matrices)]\n", "\n", "# Generate sinusoidal target\n", "t = np.linspace(0, 1, num_samples)\n", "target_signal = np.sin(2 * np.pi * t) # 1 Hz sinusoidal signal\n", "\n", "# Generate signal mixing matrix\n", - "signal_mixing_matrix = np.random.normal(size=(signal_rank, num_channels))\n", + "signal_mixing_matrix = rng.normal(size=(signal_rank, num_channels))\n", "\n", "# Prepare data matrices\n", "data_matrices = []\n", @@ -85,8 +87,6 @@ "metadata": {}, "outputs": [], "source": [ - "from meegkit import cca\n", - "\n", "# Compute Covariance matrix\n", "C = np.dot(x.T, x)\n", "\n", diff --git a/examples/example_mcca_2.py b/examples/example_mcca_2.py index 565df25f..fd72407e 100644 --- a/examples/example_mcca_2.py +++ b/examples/example_mcca_2.py @@ -1,33 +1,33 @@ """ Example 1 - sinusoidal target in separable noise. ================================================= + reproduced from de Cheveigné et al. (2018). Synthetic data for this example consisted of 10 data matrices, -each of dimensions 10000 samples x 10 channels. Each was obtained -by multiplying 9 Gaussian noise time series (independent and uncorrelated) -by a 9 x 10 mixing matrix with random Gaussian coefficients. -To this background of noise was added a “target” consisting of a sinusoidal -time series multiplied by a 1 x 10 mixing matrix with random coefficients. -The target was the same for all data matrices, -but the mixing matrices differed, -as did the noise matrices. The SNR was set to 10−20, -i.e. a very unfavorable SNR. -The noise is of rank 9 and the signal of rank 1, -so signal and noise are -in principle linearly separable. +each of dimensions 10000 samples x 10 channels. Each was +obtained by multiplying 9 Gaussian noise time series +(independent and uncorrelated) by a 9 x 10 mixing matrix with +random Gaussian coefficients. To this background of noise was +added a “target” consisting of a sinusoidal time series multiplied +by a 1 x 10 mixing matrix with random coefficients. The target was +the same for all data matrices, but the mixing matrices differed, +as did the noise matrices. The SNR was set to 10−20, i.e. a very +unfavorable SNR. The noise is of rank 9 and the signal of rank 1, +so signal and noise are in principle linearly separable. """ # %% -from meegkit import cca -import numpy as np import matplotlib.pyplot as plt +import numpy as np + +from meegkit import cca -# %% # Set the seed for the random number generator for reproducibility -np.random.seed(5) +rng = np.random.default_rng(5) +# %% # Constants num_matrices = 10 num_samples = 10000 @@ -37,9 +37,9 @@ unfavorable_SNR_dB = -20 # SNR in decibels # Generate noise matrices and mixing matrices -noise_matrices = [np.random.normal(size=(num_samples, noise_rank)) +noise_matrices = [rng.normal(size=(num_samples, noise_rank)) for _ in range(num_matrices)] -mixing_matrices = [np.random.normal(size=(noise_rank, num_channels)) +mixing_matrices = [rng.normal(size=(noise_rank, num_channels)) for _ in range(num_matrices)] # Generate sinusoidal target @@ -47,7 +47,7 @@ target_signal = np.sin(2 * np.pi * t) # 1 Hz sinusoidal signal # Generate signal mixing matrix -signal_mixing_matrix = np.random.normal(size=(signal_rank, num_channels)) +signal_mixing_matrix = rng.normal(size=(signal_rank, num_channels)) # Prepare data matrices data_matrices = [] @@ -91,18 +91,18 @@ # Plot the results fig, ax = plt.subplots(1, 4, figsize=(12, 4)) ax[0].plot(target_signal) -ax[0].set_title("Target") -ax[0].set_xlabel("Sample") -ax[0].set_ylabel("Amplitude") +ax[0].set_title('Target') +ax[0].set_xlabel('Sample') +ax[0].set_ylabel('Amplitude') ax[1].plot(data_matrix) -ax[1].set_title("Target + Noise") -ax[1].set_ylabel("Amplitude") -ax[1].set_xlabel("Sample") -ax[2].plot(variance, "o-k") -ax[2].set_xlabel("SC") -ax[2].set_ylabel("Variance") +ax[1].set_title('Target + Noise') +ax[1].set_ylabel('Amplitude') +ax[1].set_xlabel('Sample') +ax[2].plot(variance, 'o-k') +ax[2].set_xlabel('SC') +ax[2].set_ylabel('Variance') ax[3].plot(x_recovered) -ax[3].set_title("Recovered") -ax[3].set_ylabel("Amplitude") -ax[3].set_xlabel("Sample") +ax[3].set_title('Recovered') +ax[3].set_ylabel('Amplitude') +ax[3].set_xlabel('Sample') plt.tight_layout() From b36c56480ea1e22edd8f9ff81aa44a3df5f39698 Mon Sep 17 00:00:00 2001 From: John Kyle Cooper Date: Sun, 11 Jun 2023 16:16:47 +0200 Subject: [PATCH 6/7] changed single quotes to double for flake8 --- examples/example_mcca_2.py | 24 ++++++++++++------------ 1 file changed, 12 insertions(+), 12 deletions(-) diff --git a/examples/example_mcca_2.py b/examples/example_mcca_2.py index fd72407e..743d408b 100644 --- a/examples/example_mcca_2.py +++ b/examples/example_mcca_2.py @@ -91,18 +91,18 @@ # Plot the results fig, ax = plt.subplots(1, 4, figsize=(12, 4)) ax[0].plot(target_signal) -ax[0].set_title('Target') -ax[0].set_xlabel('Sample') -ax[0].set_ylabel('Amplitude') +ax[0].set_title("Target") +ax[0].set_xlabel("Sample") +ax[0].set_ylabel("Amplitude") ax[1].plot(data_matrix) -ax[1].set_title('Target + Noise') -ax[1].set_ylabel('Amplitude') -ax[1].set_xlabel('Sample') -ax[2].plot(variance, 'o-k') -ax[2].set_xlabel('SC') -ax[2].set_ylabel('Variance') +ax[1].set_title("Target + Noise") +ax[1].set_ylabel("Amplitude") +ax[1].set_xlabel("Sample") +ax[2].plot(variance, "o-k") +ax[2].set_xlabel("SC") +ax[2].set_ylabel("Variance") ax[3].plot(x_recovered) -ax[3].set_title('Recovered') -ax[3].set_ylabel('Amplitude') -ax[3].set_xlabel('Sample') +ax[3].set_title("Recovered") +ax[3].set_ylabel("Amplitude") +ax[3].set_xlabel("Sample") plt.tight_layout() From 7411f0ad80537ddd8655f0b1242bb639d25f65a9 Mon Sep 17 00:00:00 2001 From: Nicolas Barascud <10333715+nbara@users.noreply.github.com> Date: Mon, 12 Jun 2023 09:32:03 +0200 Subject: [PATCH 7/7] rebuild examples --- examples/example_asr.ipynb | 32 +++--- examples/example_dering.ipynb | 24 ++-- examples/example_detrend.ipynb | 72 ++++++------ examples/example_dss.ipynb | 40 +++---- examples/example_dss_line.ipynb | 56 +++++----- examples/example_mcca.ipynb | 90 +++++++-------- examples/example_mcca.py | 4 +- examples/example_mcca_2.ipynb | 188 +++++++++++++++++++++++--------- examples/example_mcca_2.py | 53 +++++---- examples/example_ress.ipynb | 42 +++---- examples/example_star.ipynb | 40 +++---- examples/example_star_dss.ipynb | 56 +++++----- examples/example_trca.ipynb | 50 ++++----- 13 files changed, 418 insertions(+), 329 deletions(-) diff --git a/examples/example_asr.ipynb b/examples/example_asr.ipynb index de462d2f..bb869fa5 100644 --- a/examples/example_asr.ipynb +++ b/examples/example_asr.ipynb @@ -6,10 +6,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:56.242976Z", - "iopub.status.busy": "2023-05-11T09:18:56.242363Z", - "iopub.status.idle": "2023-05-11T09:18:56.537008Z", - "shell.execute_reply": "2023-05-11T09:18:56.506069Z" + "iopub.execute_input": "2023-06-12T07:30:41.759200Z", + "iopub.status.busy": "2023-06-12T07:30:41.759042Z", + "iopub.status.idle": "2023-06-12T07:30:42.038050Z", + "shell.execute_reply": "2023-06-12T07:30:42.037069Z" } }, "outputs": [], @@ -35,10 +35,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:56.539575Z", - "iopub.status.busy": "2023-05-11T09:18:56.539423Z", - "iopub.status.idle": "2023-05-11T09:18:57.468309Z", - "shell.execute_reply": "2023-05-11T09:18:57.467945Z" + "iopub.execute_input": "2023-06-12T07:30:42.040609Z", + "iopub.status.busy": "2023-06-12T07:30:42.040439Z", + "iopub.status.idle": "2023-06-12T07:30:43.010249Z", + "shell.execute_reply": "2023-06-12T07:30:43.009872Z" } }, "outputs": [], @@ -70,10 +70,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:57.471431Z", - "iopub.status.busy": "2023-05-11T09:18:57.471281Z", - "iopub.status.idle": "2023-05-11T09:18:57.766418Z", - "shell.execute_reply": "2023-05-11T09:18:57.739208Z" + "iopub.execute_input": "2023-06-12T07:30:43.012747Z", + "iopub.status.busy": "2023-06-12T07:30:43.012410Z", + "iopub.status.idle": "2023-06-12T07:30:43.287467Z", + "shell.execute_reply": "2023-06-12T07:30:43.249524Z" } }, "outputs": [], @@ -111,10 +111,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:57.789136Z", - "iopub.status.busy": "2023-05-11T09:18:57.788591Z", - "iopub.status.idle": "2023-05-11T09:18:58.362129Z", - "shell.execute_reply": "2023-05-11T09:18:58.361743Z" + "iopub.execute_input": "2023-06-12T07:30:43.299495Z", + "iopub.status.busy": "2023-06-12T07:30:43.297956Z", + "iopub.status.idle": "2023-06-12T07:30:43.854373Z", + "shell.execute_reply": "2023-06-12T07:30:43.854021Z" } }, "outputs": [ diff --git a/examples/example_dering.ipynb b/examples/example_dering.ipynb index a7823d6c..864322b7 100644 --- a/examples/example_dering.ipynb +++ b/examples/example_dering.ipynb @@ -6,10 +6,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:02.307919Z", - "iopub.status.busy": "2023-05-11T09:19:02.307506Z", - "iopub.status.idle": "2023-05-11T09:19:02.653295Z", - "shell.execute_reply": "2023-05-11T09:19:02.652942Z" + "iopub.execute_input": "2023-06-12T07:30:47.585526Z", + "iopub.status.busy": "2023-06-12T07:30:47.585274Z", + "iopub.status.idle": "2023-06-12T07:30:47.909765Z", + "shell.execute_reply": "2023-06-12T07:30:47.909375Z" } }, "outputs": [], @@ -34,10 +34,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:02.656577Z", - "iopub.status.busy": "2023-05-11T09:19:02.656159Z", - "iopub.status.idle": "2023-05-11T09:19:03.577456Z", - "shell.execute_reply": "2023-05-11T09:19:03.577083Z" + "iopub.execute_input": "2023-06-12T07:30:47.911933Z", + "iopub.status.busy": "2023-06-12T07:30:47.911765Z", + "iopub.status.idle": "2023-06-12T07:30:48.853449Z", + "shell.execute_reply": "2023-06-12T07:30:48.853036Z" } }, "outputs": [], @@ -76,10 +76,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:03.579843Z", - "iopub.status.busy": "2023-05-11T09:19:03.579610Z", - "iopub.status.idle": "2023-05-11T09:19:03.748268Z", - "shell.execute_reply": "2023-05-11T09:19:03.747953Z" + "iopub.execute_input": "2023-06-12T07:30:48.855823Z", + "iopub.status.busy": "2023-06-12T07:30:48.855536Z", + "iopub.status.idle": "2023-06-12T07:30:49.026251Z", + "shell.execute_reply": "2023-06-12T07:30:49.025759Z" } }, "outputs": [ diff --git a/examples/example_detrend.ipynb b/examples/example_detrend.ipynb index 02a7b040..43d3dcbb 100644 --- a/examples/example_detrend.ipynb +++ b/examples/example_detrend.ipynb @@ -6,10 +6,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:48.845013Z", - "iopub.status.busy": "2023-05-11T09:18:48.844870Z", - "iopub.status.idle": "2023-05-11T09:18:49.144585Z", - "shell.execute_reply": "2023-05-11T09:18:49.143789Z" + "iopub.execute_input": "2023-06-12T07:30:34.360752Z", + "iopub.status.busy": "2023-06-12T07:30:34.360601Z", + "iopub.status.idle": "2023-06-12T07:30:34.641867Z", + "shell.execute_reply": "2023-06-12T07:30:34.641520Z" } }, "outputs": [], @@ -40,10 +40,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:49.147055Z", - "iopub.status.busy": "2023-05-11T09:18:49.146714Z", - "iopub.status.idle": "2023-05-11T09:18:50.130695Z", - "shell.execute_reply": "2023-05-11T09:18:50.130306Z" + "iopub.execute_input": "2023-06-12T07:30:34.644192Z", + "iopub.status.busy": "2023-06-12T07:30:34.644035Z", + "iopub.status.idle": "2023-06-12T07:30:35.591147Z", + "shell.execute_reply": "2023-06-12T07:30:35.590761Z" } }, "outputs": [], @@ -82,10 +82,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:50.132979Z", - "iopub.status.busy": "2023-05-11T09:18:50.132782Z", - "iopub.status.idle": "2023-05-11T09:18:50.224531Z", - "shell.execute_reply": "2023-05-11T09:18:50.224146Z" + "iopub.execute_input": "2023-06-12T07:30:35.593795Z", + "iopub.status.busy": "2023-06-12T07:30:35.593445Z", + "iopub.status.idle": "2023-06-12T07:30:35.728529Z", + "shell.execute_reply": "2023-06-12T07:30:35.728208Z" } }, "outputs": [ @@ -131,17 +131,17 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:50.226719Z", - "iopub.status.busy": "2023-05-11T09:18:50.226596Z", - "iopub.status.idle": "2023-05-11T09:18:50.404693Z", - "shell.execute_reply": "2023-05-11T09:18:50.404372Z" + "iopub.execute_input": "2023-06-12T07:30:35.731278Z", + "iopub.status.busy": "2023-06-12T07:30:35.731155Z", + "iopub.status.idle": "2023-06-12T07:30:35.873102Z", + "shell.execute_reply": "2023-06-12T07:30:35.872781Z" } }, "outputs": [ { "data": { "text/plain": [ - "" + "" ] }, "execution_count": 4, @@ -193,17 +193,17 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:50.406750Z", - "iopub.status.busy": "2023-05-11T09:18:50.406630Z", - "iopub.status.idle": "2023-05-11T09:18:50.514174Z", - "shell.execute_reply": "2023-05-11T09:18:50.513838Z" + "iopub.execute_input": "2023-06-12T07:30:35.875698Z", + "iopub.status.busy": "2023-06-12T07:30:35.875025Z", + "iopub.status.idle": "2023-06-12T07:30:35.985532Z", + "shell.execute_reply": "2023-06-12T07:30:35.985137Z" } }, "outputs": [ { "data": { "text/plain": [ - "" + "" ] }, "execution_count": 5, @@ -255,17 +255,17 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:50.516367Z", - "iopub.status.busy": "2023-05-11T09:18:50.516236Z", - "iopub.status.idle": "2023-05-11T09:18:50.595676Z", - "shell.execute_reply": "2023-05-11T09:18:50.595350Z" + "iopub.execute_input": "2023-06-12T07:30:35.987627Z", + "iopub.status.busy": "2023-06-12T07:30:35.987524Z", + "iopub.status.idle": "2023-06-12T07:30:36.069910Z", + "shell.execute_reply": "2023-06-12T07:30:36.069420Z" } }, "outputs": [ { "data": { "text/plain": [ - "" + "" ] }, "execution_count": 6, @@ -308,17 +308,17 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:50.598189Z", - "iopub.status.busy": "2023-05-11T09:18:50.597965Z", - "iopub.status.idle": "2023-05-11T09:18:50.679911Z", - "shell.execute_reply": "2023-05-11T09:18:50.679522Z" + "iopub.execute_input": "2023-06-12T07:30:36.072018Z", + "iopub.status.busy": "2023-06-12T07:30:36.071796Z", + "iopub.status.idle": "2023-06-12T07:30:36.154808Z", + "shell.execute_reply": "2023-06-12T07:30:36.154511Z" } }, "outputs": [ { "data": { "text/plain": [ - "" + "" ] }, "execution_count": 7, @@ -367,10 +367,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:50.682253Z", - "iopub.status.busy": "2023-05-11T09:18:50.682031Z", - "iopub.status.idle": "2023-05-11T09:18:50.766453Z", - "shell.execute_reply": "2023-05-11T09:18:50.766146Z" + "iopub.execute_input": "2023-06-12T07:30:36.156821Z", + "iopub.status.busy": "2023-06-12T07:30:36.156693Z", + "iopub.status.idle": "2023-06-12T07:30:36.241441Z", + "shell.execute_reply": "2023-06-12T07:30:36.241050Z" } }, "outputs": [ diff --git a/examples/example_dss.ipynb b/examples/example_dss.ipynb index 62f8bdee..6a132670 100644 --- a/examples/example_dss.ipynb +++ b/examples/example_dss.ipynb @@ -6,10 +6,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:04.772654Z", - "iopub.status.busy": "2023-05-11T09:19:04.772529Z", - "iopub.status.idle": "2023-05-11T09:19:05.083106Z", - "shell.execute_reply": "2023-05-11T09:19:05.082141Z" + "iopub.execute_input": "2023-06-12T07:30:54.332335Z", + "iopub.status.busy": "2023-06-12T07:30:54.330325Z", + "iopub.status.idle": "2023-06-12T07:30:54.678871Z", + "shell.execute_reply": "2023-06-12T07:30:54.678070Z" } }, "outputs": [], @@ -36,10 +36,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:05.085983Z", - "iopub.status.busy": "2023-05-11T09:19:05.085469Z", - "iopub.status.idle": "2023-05-11T09:19:06.011756Z", - "shell.execute_reply": "2023-05-11T09:19:06.011374Z" + "iopub.execute_input": "2023-06-12T07:30:54.684828Z", + "iopub.status.busy": "2023-06-12T07:30:54.684572Z", + "iopub.status.idle": "2023-06-12T07:30:55.651879Z", + "shell.execute_reply": "2023-06-12T07:30:55.651424Z" } }, "outputs": [], @@ -67,10 +67,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:06.013969Z", - "iopub.status.busy": "2023-05-11T09:19:06.013832Z", - "iopub.status.idle": "2023-05-11T09:19:06.138154Z", - "shell.execute_reply": "2023-05-11T09:19:06.097649Z" + "iopub.execute_input": "2023-06-12T07:30:55.654326Z", + "iopub.status.busy": "2023-06-12T07:30:55.653987Z", + "iopub.status.idle": "2023-06-12T07:30:55.774200Z", + "shell.execute_reply": "2023-06-12T07:30:55.735384Z" } }, "outputs": [], @@ -115,10 +115,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:06.153859Z", - "iopub.status.busy": "2023-05-11T09:19:06.153625Z", - "iopub.status.idle": "2023-05-11T09:19:06.397817Z", - "shell.execute_reply": "2023-05-11T09:19:06.393715Z" + "iopub.execute_input": "2023-06-12T07:30:55.796947Z", + "iopub.status.busy": "2023-06-12T07:30:55.795907Z", + "iopub.status.idle": "2023-06-12T07:30:55.875945Z", + "shell.execute_reply": "2023-06-12T07:30:55.875448Z" } }, "outputs": [], @@ -152,10 +152,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:06.411141Z", - "iopub.status.busy": "2023-05-11T09:19:06.410550Z", - "iopub.status.idle": "2023-05-11T09:19:06.672263Z", - "shell.execute_reply": "2023-05-11T09:19:06.671833Z" + "iopub.execute_input": "2023-06-12T07:30:55.882569Z", + "iopub.status.busy": "2023-06-12T07:30:55.882428Z", + "iopub.status.idle": "2023-06-12T07:30:56.226688Z", + "shell.execute_reply": "2023-06-12T07:30:56.226347Z" } }, "outputs": [ diff --git a/examples/example_dss_line.ipynb b/examples/example_dss_line.ipynb index b700d1c8..a87b0369 100644 --- a/examples/example_dss_line.ipynb +++ b/examples/example_dss_line.ipynb @@ -6,10 +6,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:22.419329Z", - "iopub.status.busy": "2023-05-11T09:19:22.419190Z", - "iopub.status.idle": "2023-05-11T09:19:22.787580Z", - "shell.execute_reply": "2023-05-11T09:19:22.787160Z" + "iopub.execute_input": "2023-06-12T07:31:11.002441Z", + "iopub.status.busy": "2023-06-12T07:31:11.002273Z", + "iopub.status.idle": "2023-06-12T07:31:11.379347Z", + "shell.execute_reply": "2023-06-12T07:31:11.378999Z" } }, "outputs": [], @@ -39,10 +39,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:22.789848Z", - "iopub.status.busy": "2023-05-11T09:19:22.789680Z", - "iopub.status.idle": "2023-05-11T09:19:23.756972Z", - "shell.execute_reply": "2023-05-11T09:19:23.756570Z" + "iopub.execute_input": "2023-06-12T07:31:11.381889Z", + "iopub.status.busy": "2023-06-12T07:31:11.381685Z", + "iopub.status.idle": "2023-06-12T07:31:12.409941Z", + "shell.execute_reply": "2023-06-12T07:31:12.409443Z" } }, "outputs": [], @@ -82,10 +82,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:23.759546Z", - "iopub.status.busy": "2023-05-11T09:19:23.759388Z", - "iopub.status.idle": "2023-05-11T09:19:23.892423Z", - "shell.execute_reply": "2023-05-11T09:19:23.889212Z" + "iopub.execute_input": "2023-06-12T07:31:12.412514Z", + "iopub.status.busy": "2023-06-12T07:31:12.412252Z", + "iopub.status.idle": "2023-06-12T07:31:12.544467Z", + "shell.execute_reply": "2023-06-12T07:31:12.544066Z" } }, "outputs": [ @@ -125,10 +125,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:23.905574Z", - "iopub.status.busy": "2023-05-11T09:19:23.905034Z", - "iopub.status.idle": "2023-05-11T09:19:24.400704Z", - "shell.execute_reply": "2023-05-11T09:19:24.400360Z" + "iopub.execute_input": "2023-06-12T07:31:12.547079Z", + "iopub.status.busy": "2023-06-12T07:31:12.546969Z", + "iopub.status.idle": "2023-06-12T07:31:13.038615Z", + "shell.execute_reply": "2023-06-12T07:31:13.038295Z" } }, "outputs": [ @@ -172,10 +172,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:24.402887Z", - "iopub.status.busy": "2023-05-11T09:19:24.402759Z", - "iopub.status.idle": "2023-05-11T09:19:25.688193Z", - "shell.execute_reply": "2023-05-11T09:19:25.686663Z" + "iopub.execute_input": "2023-06-12T07:31:13.040420Z", + "iopub.status.busy": "2023-06-12T07:31:13.040304Z", + "iopub.status.idle": "2023-06-12T07:31:14.225897Z", + "shell.execute_reply": "2023-06-12T07:31:14.212064Z" } }, "outputs": [ @@ -213,10 +213,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:25.698616Z", - "iopub.status.busy": "2023-05-11T09:19:25.697827Z", - "iopub.status.idle": "2023-05-11T09:19:31.463723Z", - "shell.execute_reply": "2023-05-11T09:19:31.463389Z" + "iopub.execute_input": "2023-06-12T07:31:14.241888Z", + "iopub.status.busy": "2023-06-12T07:31:14.241743Z", + "iopub.status.idle": "2023-06-12T07:31:19.747380Z", + "shell.execute_reply": "2023-06-12T07:31:19.747020Z" } }, "outputs": [ @@ -255,10 +255,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:31.465676Z", - "iopub.status.busy": "2023-05-11T09:19:31.465559Z", - "iopub.status.idle": "2023-05-11T09:19:32.064153Z", - "shell.execute_reply": "2023-05-11T09:19:32.063695Z" + "iopub.execute_input": "2023-06-12T07:31:19.749476Z", + "iopub.status.busy": "2023-06-12T07:31:19.749338Z", + "iopub.status.idle": "2023-06-12T07:31:20.379273Z", + "shell.execute_reply": "2023-06-12T07:31:20.378930Z" } }, "outputs": [ diff --git a/examples/example_mcca.ipynb b/examples/example_mcca.ipynb index 18adadad..bce53608 100644 --- a/examples/example_mcca.ipynb +++ b/examples/example_mcca.ipynb @@ -6,10 +6,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:51.933643Z", - "iopub.status.busy": "2023-05-11T09:18:51.932524Z", - "iopub.status.idle": "2023-05-11T09:18:52.192672Z", - "shell.execute_reply": "2023-05-11T09:18:52.179977Z" + "iopub.execute_input": "2023-06-12T07:30:37.202665Z", + "iopub.status.busy": "2023-06-12T07:30:37.202350Z", + "iopub.status.idle": "2023-06-12T07:30:37.521199Z", + "shell.execute_reply": "2023-06-12T07:30:37.520860Z" } }, "outputs": [], @@ -22,7 +22,7 @@ "metadata": {}, "source": [ "\n", - "# Example multiway canonical correlation analysis (mCCA)\n", + "# Multiway canonical correlation analysis (mCCA)\n", "\n", "Find a set of components which are shared between different datasets.\n", "\n", @@ -35,10 +35,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:52.219471Z", - "iopub.status.busy": "2023-05-11T09:18:52.219272Z", - "iopub.status.idle": "2023-05-11T09:18:53.114029Z", - "shell.execute_reply": "2023-05-11T09:18:53.113675Z" + "iopub.execute_input": "2023-06-12T07:30:37.524256Z", + "iopub.status.busy": "2023-06-12T07:30:37.523825Z", + "iopub.status.idle": "2023-06-12T07:30:38.497958Z", + "shell.execute_reply": "2023-06-12T07:30:38.497527Z" } }, "outputs": [], @@ -75,10 +75,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:53.116238Z", - "iopub.status.busy": "2023-05-11T09:18:53.116107Z", - "iopub.status.idle": "2023-05-11T09:18:53.121346Z", - "shell.execute_reply": "2023-05-11T09:18:53.121034Z" + "iopub.execute_input": "2023-06-12T07:30:38.500425Z", + "iopub.status.busy": "2023-06-12T07:30:38.500115Z", + "iopub.status.idle": "2023-06-12T07:30:38.505708Z", + "shell.execute_reply": "2023-06-12T07:30:38.505382Z" } }, "outputs": [ @@ -113,10 +113,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:53.123197Z", - "iopub.status.busy": "2023-05-11T09:18:53.123069Z", - "iopub.status.idle": "2023-05-11T09:18:53.284637Z", - "shell.execute_reply": "2023-05-11T09:18:53.282147Z" + "iopub.execute_input": "2023-06-12T07:30:38.507742Z", + "iopub.status.busy": "2023-06-12T07:30:38.507607Z", + "iopub.status.idle": "2023-06-12T07:30:38.567463Z", + "shell.execute_reply": "2023-06-12T07:30:38.564582Z" } }, "outputs": [], @@ -139,10 +139,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:53.295323Z", - "iopub.status.busy": "2023-05-11T09:18:53.295161Z", - "iopub.status.idle": "2023-05-11T09:18:53.854474Z", - "shell.execute_reply": "2023-05-11T09:18:53.853764Z" + "iopub.execute_input": "2023-06-12T07:30:38.585985Z", + "iopub.status.busy": "2023-06-12T07:30:38.585289Z", + "iopub.status.idle": "2023-06-12T07:30:39.122202Z", + "shell.execute_reply": "2023-06-12T07:30:39.121853Z" } }, "outputs": [ @@ -194,10 +194,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:53.857101Z", - "iopub.status.busy": "2023-05-11T09:18:53.856957Z", - "iopub.status.idle": "2023-05-11T09:18:53.863578Z", - "shell.execute_reply": "2023-05-11T09:18:53.862634Z" + "iopub.execute_input": "2023-06-12T07:30:39.124200Z", + "iopub.status.busy": "2023-06-12T07:30:39.124086Z", + "iopub.status.idle": "2023-06-12T07:30:39.129109Z", + "shell.execute_reply": "2023-06-12T07:30:39.128679Z" } }, "outputs": [ @@ -233,10 +233,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:53.866759Z", - "iopub.status.busy": "2023-05-11T09:18:53.866604Z", - "iopub.status.idle": "2023-05-11T09:18:53.878602Z", - "shell.execute_reply": "2023-05-11T09:18:53.874520Z" + "iopub.execute_input": "2023-06-12T07:30:39.131150Z", + "iopub.status.busy": "2023-06-12T07:30:39.130991Z", + "iopub.status.idle": "2023-06-12T07:30:39.134892Z", + "shell.execute_reply": "2023-06-12T07:30:39.134455Z" } }, "outputs": [], @@ -258,10 +258,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:53.893058Z", - "iopub.status.busy": "2023-05-11T09:18:53.890208Z", - "iopub.status.idle": "2023-05-11T09:18:54.532028Z", - "shell.execute_reply": "2023-05-11T09:18:54.531514Z" + "iopub.execute_input": "2023-06-12T07:30:39.137030Z", + "iopub.status.busy": "2023-06-12T07:30:39.136797Z", + "iopub.status.idle": "2023-06-12T07:30:39.724039Z", + "shell.execute_reply": "2023-06-12T07:30:39.720438Z" } }, "outputs": [ @@ -315,10 +315,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:54.534968Z", - "iopub.status.busy": "2023-05-11T09:18:54.534812Z", - "iopub.status.idle": "2023-05-11T09:18:54.538868Z", - "shell.execute_reply": "2023-05-11T09:18:54.538542Z" + "iopub.execute_input": "2023-06-12T07:30:39.726298Z", + "iopub.status.busy": "2023-06-12T07:30:39.726175Z", + "iopub.status.idle": "2023-06-12T07:30:39.730220Z", + "shell.execute_reply": "2023-06-12T07:30:39.729681Z" } }, "outputs": [ @@ -351,10 +351,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:54.541076Z", - "iopub.status.busy": "2023-05-11T09:18:54.540740Z", - "iopub.status.idle": "2023-05-11T09:18:54.545714Z", - "shell.execute_reply": "2023-05-11T09:18:54.544911Z" + "iopub.execute_input": "2023-06-12T07:30:39.733527Z", + "iopub.status.busy": "2023-06-12T07:30:39.732294Z", + "iopub.status.idle": "2023-06-12T07:30:39.738899Z", + "shell.execute_reply": "2023-06-12T07:30:39.737424Z" } }, "outputs": [], @@ -376,10 +376,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:54.548504Z", - "iopub.status.busy": "2023-05-11T09:18:54.548060Z", - "iopub.status.idle": "2023-05-11T09:18:55.181303Z", - "shell.execute_reply": "2023-05-11T09:18:55.180946Z" + "iopub.execute_input": "2023-06-12T07:30:39.743056Z", + "iopub.status.busy": "2023-06-12T07:30:39.742731Z", + "iopub.status.idle": "2023-06-12T07:30:40.559172Z", + "shell.execute_reply": "2023-06-12T07:30:40.558771Z" } }, "outputs": [ diff --git a/examples/example_mcca.py b/examples/example_mcca.py index d2f13400..d51c6714 100644 --- a/examples/example_mcca.py +++ b/examples/example_mcca.py @@ -1,6 +1,6 @@ """ -Example multiway canonical correlation analysis (mCCA) -====================================================== +Multiway canonical correlation analysis (mCCA) +============================================== Find a set of components which are shared between different datasets. diff --git a/examples/example_mcca_2.ipynb b/examples/example_mcca_2.ipynb index 556a8fd4..6823c38b 100644 --- a/examples/example_mcca_2.ipynb +++ b/examples/example_mcca_2.ipynb @@ -1,20 +1,56 @@ { "cells": [ { - "attachments": {}, + "cell_type": "code", + "execution_count": 1, + "metadata": { + "collapsed": false, + "execution": { + "iopub.execute_input": "2023-06-12T07:30:50.212756Z", + "iopub.status.busy": "2023-06-12T07:30:50.212616Z", + "iopub.status.idle": "2023-06-12T07:30:50.596404Z", + "shell.execute_reply": "2023-06-12T07:30:50.596058Z" + } + }, + "outputs": [], + "source": [ + "%matplotlib inline" + ] + }, + { "cell_type": "markdown", "metadata": {}, "source": [ - "# Example 1 - sinusoidal target in separable noise.\n", - "reproduced from de Cheveigné et al. (2018).\n", "\n", - "Synthetic data for this example consisted of 10 data matrices, each of dimensions 10000 samples x 10 channels. Each was obtained by multiplying 9 Gaussian noise time series (independent and uncorrelated) by a 9 x 10 mixing matrix with random Gaussian coefficients. To this background of noise was added a “target” consisting of a sinusoidal time series multiplied by a 1 x 10 mixing matrix with random coefficients. The target was the same for all data matrices, but the mixing matrices differed, as did the noise matrices. The SNR was set to 10−20, i.e. a very unfavorable SNR. The noise is of rank 9 and the signal of rank 1, so signal and noise are in principle linearly separable." + "# mCCA example: Sinusoidal target in separable noise\n", + "\n", + "Reproduced from de Cheveigné et al. (2018).\n", + "\n", + "Synthetic data for this example consisted of 10 data matrices, each of\n", + "dimensions 10000 samples x 10 channels. Each was obtained by multiplying 9\n", + "Gaussian noise time series (independent and uncorrelated) by a 9 x 10 mixing\n", + "matrix with random Gaussian coefficients. To this background of noise was added\n", + "a \"target\" consisting of a sinusoidal time series multiplied by a 1 x 10 mixing\n", + "matrix with random coefficients. The target was the same for all data matrices,\n", + "but the mixing matrices differed, as were the noise matrices. The SNR was set\n", + "to 10−20, i.e. a very unfavorable SNR. The noise is of rank 9 and the signal of\n", + "rank 1, so signal and noise are in principle linearly separable.\n", + "\n", + "Uses meegkit.cca.mmca()\n" ] }, { "cell_type": "code", - "execution_count": null, - "metadata": {}, + "execution_count": 2, + "metadata": { + "collapsed": false, + "execution": { + "iopub.execute_input": "2023-06-12T07:30:50.598931Z", + "iopub.status.busy": "2023-06-12T07:30:50.598746Z", + "iopub.status.idle": "2023-06-12T07:30:51.676071Z", + "shell.execute_reply": "2023-06-12T07:30:51.675663Z" + } + }, "outputs": [], "source": [ "import matplotlib.pyplot as plt\n", @@ -27,12 +63,28 @@ ] }, { - "cell_type": "code", - "execution_count": null, + "cell_type": "markdown", "metadata": {}, + "source": [ + "## Generate toy data\n", + "Constants\n", + "\n" + ] + }, + { + "cell_type": "code", + "execution_count": 3, + "metadata": { + "collapsed": false, + "execution": { + "iopub.execute_input": "2023-06-12T07:30:51.678567Z", + "iopub.status.busy": "2023-06-12T07:30:51.678325Z", + "iopub.status.idle": "2023-06-12T07:30:51.789638Z", + "shell.execute_reply": "2023-06-12T07:30:51.774290Z" + } + }, "outputs": [], "source": [ - "# Constants\n", "num_matrices = 10\n", "num_samples = 10000\n", "num_channels = 10\n", @@ -41,8 +93,10 @@ "unfavorable_SNR_dB = -20 # SNR in decibels\n", "\n", "# Generate noise matrices and mixing matrices\n", - "noise_matrices = [rng.normal(size=(num_samples, noise_rank)) for _ in range(num_matrices)]\n", - "mixing_matrices = [rng.normal(size=(noise_rank, num_channels)) for _ in range(num_matrices)]\n", + "noise_matrices = [rng.normal(size=(num_samples, noise_rank))\n", + " for _ in range(num_matrices)]\n", + "mixing_matrices = [rng.normal(size=(noise_rank, num_channels))\n", + " for _ in range(num_matrices)]\n", "\n", "# Generate sinusoidal target\n", "t = np.linspace(0, 1, num_samples)\n", @@ -68,76 +122,107 @@ " # Add signal and noise\n", " data_matrix = signal + noise\n", "\n", - " data_matrices.append(data_matrix)" + " data_matrices.append(data_matrix)\n", + "\n", + "# Concatenate data matrices\n", + "x = np.concatenate(data_matrices, axis=-1)" ] }, { - "cell_type": "code", - "execution_count": null, + "cell_type": "markdown", "metadata": {}, - "outputs": [], "source": [ - "# Concatenate data matrices\n", - "x = np.concatenate(data_matrices, axis=-1)" + "## Use mCCA to recover signal in noise\n", + "\n" ] }, { "cell_type": "code", - "execution_count": null, - "metadata": {}, + "execution_count": 4, + "metadata": { + "collapsed": false, + "execution": { + "iopub.execute_input": "2023-06-12T07:30:51.800552Z", + "iopub.status.busy": "2023-06-12T07:30:51.800357Z", + "iopub.status.idle": "2023-06-12T07:30:52.632461Z", + "shell.execute_reply": "2023-06-12T07:30:52.629031Z" + } + }, "outputs": [], "source": [ "# Compute Covariance matrix\n", "C = np.dot(x.T, x)\n", "\n", - "# Compute mCCA\n", - "A, score, AA = cca.mcca(C, 10)" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ + "# Compute mCCA from covariance\n", + "A, score, AA = cca.mcca(C, 10)\n", + "\n", "# Compute the recovered signal using first SC\n", - "x_recovered = x.dot(A)[:,0]\n", + "x_recovered = x.dot(A)[:, 0]\n", + "\n", "# Normalize the recovered signal\n", "x_recovered = x_recovered / x_recovered.std()\n", + "\n", "# Compute variance across SCs\n", "variance = np.var(x.dot(A), axis=0)" ] }, { - "cell_type": "code", - "execution_count": null, + "cell_type": "markdown", "metadata": {}, - "outputs": [], "source": [ - "# Plot the results\n", + "## Plot the results\n", + "\n" + ] + }, + { + "cell_type": "code", + "execution_count": 5, + "metadata": { + "collapsed": false, + "execution": { + "iopub.execute_input": "2023-06-12T07:30:52.641248Z", + "iopub.status.busy": "2023-06-12T07:30:52.638178Z", + "iopub.status.idle": "2023-06-12T07:30:53.161325Z", + "shell.execute_reply": "2023-06-12T07:30:53.160942Z" + } + }, + "outputs": [ + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ "fig, ax = plt.subplots(1, 4, figsize=(12, 4))\n", "ax[0].plot(target_signal)\n", - "ax[0].set_title('Target')\n", - "ax[0].set_xlabel('Sample')\n", - "ax[0].set_ylabel('Amplitude')\n", + "ax[0].set_title(\"Target\")\n", + "ax[0].set_xlabel(\"Sample\")\n", + "ax[0].set_ylabel(\"Amplitude\")\n", "ax[1].plot(data_matrix)\n", - "ax[1].set_title('Target + Noise')\n", - "ax[1].set_ylabel('Amplitude')\n", - "ax[1].set_xlabel('Sample')\n", - "ax[2].plot(variance, 'o-k')\n", - "ax[2].set_xlabel('SC')\n", - "ax[2].set_ylabel('Variance')\n", + "ax[1].set_title(\"Target + Noise\")\n", + "ax[1].set_ylabel(\"Amplitude\")\n", + "ax[1].set_xlabel(\"Sample\")\n", + "ax[2].plot(variance, \"o-k\")\n", + "ax[2].set_xlabel(\"SC\")\n", + "ax[2].set_ylabel(\"Variance\")\n", "ax[3].plot(x_recovered)\n", - "ax[3].set_title('Recovered')\n", - "ax[3].set_ylabel('Amplitude')\n", - "ax[3].set_xlabel('Sample')\n", - "plt.tight_layout()" + "ax[3].set_title(\"Recovered\")\n", + "ax[3].set_ylabel(\"Amplitude\")\n", + "ax[3].set_xlabel(\"Sample\")\n", + "plt.tight_layout()\n", + "plt.show()" ] } ], "metadata": { "kernelspec": { - "display_name": "eelbrain", + "display_name": "Python 3", "language": "python", "name": "python3" }, @@ -151,10 +236,9 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.8.16" - }, - "orig_nbformat": 4 + "version": "3.10.6" + } }, "nbformat": 4, - "nbformat_minor": 2 + "nbformat_minor": 0 } diff --git a/examples/example_mcca_2.py b/examples/example_mcca_2.py index 743d408b..696c766e 100644 --- a/examples/example_mcca_2.py +++ b/examples/example_mcca_2.py @@ -1,24 +1,22 @@ """ -Example 1 - sinusoidal target in separable noise. -================================================= - -reproduced from de Cheveigné et al. (2018). - -Synthetic data for this example consisted of 10 data matrices, -each of dimensions 10000 samples x 10 channels. Each was -obtained by multiplying 9 Gaussian noise time series -(independent and uncorrelated) by a 9 x 10 mixing matrix with -random Gaussian coefficients. To this background of noise was -added a “target” consisting of a sinusoidal time series multiplied -by a 1 x 10 mixing matrix with random coefficients. The target was -the same for all data matrices, but the mixing matrices differed, -as did the noise matrices. The SNR was set to 10−20, i.e. a very -unfavorable SNR. The noise is of rank 9 and the signal of rank 1, -so signal and noise are in principle linearly separable. +mCCA example: Sinusoidal target in separable noise +================================================== -""" +Reproduced from de Cheveigné et al. (2018). + +Synthetic data for this example consisted of 10 data matrices, each of +dimensions 10000 samples x 10 channels. Each was obtained by multiplying 9 +Gaussian noise time series (independent and uncorrelated) by a 9 x 10 mixing +matrix with random Gaussian coefficients. To this background of noise was added +a "target" consisting of a sinusoidal time series multiplied by a 1 x 10 mixing +matrix with random coefficients. The target was the same for all data matrices, +but the mixing matrices differed, as were the noise matrices. The SNR was set +to 10−20, i.e. a very unfavorable SNR. The noise is of rank 9 and the signal of +rank 1, so signal and noise are in principle linearly separable. -# %% +Uses meegkit.cca.mmca() + +""" import matplotlib.pyplot as plt import numpy as np @@ -27,7 +25,9 @@ # Set the seed for the random number generator for reproducibility rng = np.random.default_rng(5) -# %% +############################################################################### +# Generate toy data +# ----------------------------------------------------------------------------- # Constants num_matrices = 10 num_samples = 10000 @@ -68,27 +68,31 @@ data_matrices.append(data_matrix) -# %% # Concatenate data matrices x = np.concatenate(data_matrices, axis=-1) -# %% +############################################################################### +# Use mCCA to recover signal in noise +# ----------------------------------------------------------------------------- + # Compute Covariance matrix C = np.dot(x.T, x) -# Compute mCCA +# Compute mCCA from covariance A, score, AA = cca.mcca(C, 10) -# %% # Compute the recovered signal using first SC x_recovered = x.dot(A)[:, 0] + # Normalize the recovered signal x_recovered = x_recovered / x_recovered.std() + # Compute variance across SCs variance = np.var(x.dot(A), axis=0) -# %% +############################################################################### # Plot the results +# ----------------------------------------------------------------------------- fig, ax = plt.subplots(1, 4, figsize=(12, 4)) ax[0].plot(target_signal) ax[0].set_title("Target") @@ -106,3 +110,4 @@ ax[3].set_ylabel("Amplitude") ax[3].set_xlabel("Sample") plt.tight_layout() +plt.show() diff --git a/examples/example_ress.ipynb b/examples/example_ress.ipynb index 227be3cd..5e11c08b 100644 --- a/examples/example_ress.ipynb +++ b/examples/example_ress.ipynb @@ -6,10 +6,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:45.537090Z", - "iopub.status.busy": "2023-05-11T09:18:45.536959Z", - "iopub.status.idle": "2023-05-11T09:18:45.847738Z", - "shell.execute_reply": "2023-05-11T09:18:45.847323Z" + "iopub.execute_input": "2023-06-12T07:30:31.146658Z", + "iopub.status.busy": "2023-06-12T07:30:31.146502Z", + "iopub.status.idle": "2023-06-12T07:30:31.470818Z", + "shell.execute_reply": "2023-06-12T07:30:31.470501Z" } }, "outputs": [], @@ -37,10 +37,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:45.850708Z", - "iopub.status.busy": "2023-05-11T09:18:45.850539Z", - "iopub.status.idle": "2023-05-11T09:18:46.816387Z", - "shell.execute_reply": "2023-05-11T09:18:46.815997Z" + "iopub.execute_input": "2023-06-12T07:30:31.473009Z", + "iopub.status.busy": "2023-06-12T07:30:31.472851Z", + "iopub.status.idle": "2023-06-12T07:30:32.423303Z", + "shell.execute_reply": "2023-06-12T07:30:32.422922Z" } }, "outputs": [], @@ -72,17 +72,17 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:46.818725Z", - "iopub.status.busy": "2023-05-11T09:18:46.818565Z", - "iopub.status.idle": "2023-05-11T09:18:47.183158Z", - "shell.execute_reply": "2023-05-11T09:18:47.182828Z" + "iopub.execute_input": "2023-06-12T07:30:32.425508Z", + "iopub.status.busy": "2023-06-12T07:30:32.425310Z", + "iopub.status.idle": "2023-06-12T07:30:32.804600Z", + "shell.execute_reply": "2023-06-12T07:30:32.804215Z" } }, "outputs": [ { "data": { "text/plain": [ - "" + "" ] }, "execution_count": 3, @@ -154,10 +154,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:47.185190Z", - "iopub.status.busy": "2023-05-11T09:18:47.185076Z", - "iopub.status.idle": "2023-05-11T09:18:47.287122Z", - "shell.execute_reply": "2023-05-11T09:18:47.286607Z" + "iopub.execute_input": "2023-06-12T07:30:32.807278Z", + "iopub.status.busy": "2023-06-12T07:30:32.807125Z", + "iopub.status.idle": "2023-06-12T07:30:32.912144Z", + "shell.execute_reply": "2023-06-12T07:30:32.911820Z" } }, "outputs": [ @@ -219,10 +219,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:47.289673Z", - "iopub.status.busy": "2023-05-11T09:18:47.289435Z", - "iopub.status.idle": "2023-05-11T09:18:47.812541Z", - "shell.execute_reply": "2023-05-11T09:18:47.811995Z" + "iopub.execute_input": "2023-06-12T07:30:32.914047Z", + "iopub.status.busy": "2023-06-12T07:30:32.913919Z", + "iopub.status.idle": "2023-06-12T07:30:33.384618Z", + "shell.execute_reply": "2023-06-12T07:30:33.384284Z" } }, "outputs": [ diff --git a/examples/example_star.ipynb b/examples/example_star.ipynb index 6bb612f0..0d814412 100644 --- a/examples/example_star.ipynb +++ b/examples/example_star.ipynb @@ -6,10 +6,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:07.690921Z", - "iopub.status.busy": "2023-05-11T09:19:07.690652Z", - "iopub.status.idle": "2023-05-11T09:19:08.099730Z", - "shell.execute_reply": "2023-05-11T09:19:08.098650Z" + "iopub.execute_input": "2023-06-12T07:30:57.401848Z", + "iopub.status.busy": "2023-06-12T07:30:57.400306Z", + "iopub.status.idle": "2023-06-12T07:30:57.689618Z", + "shell.execute_reply": "2023-06-12T07:30:57.689170Z" } }, "outputs": [], @@ -34,10 +34,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:08.103338Z", - "iopub.status.busy": "2023-05-11T09:19:08.103095Z", - "iopub.status.idle": "2023-05-11T09:19:09.131645Z", - "shell.execute_reply": "2023-05-11T09:19:09.131247Z" + "iopub.execute_input": "2023-06-12T07:30:57.691921Z", + "iopub.status.busy": "2023-06-12T07:30:57.691754Z", + "iopub.status.idle": "2023-06-12T07:30:58.680274Z", + "shell.execute_reply": "2023-06-12T07:30:58.679780Z" } }, "outputs": [], @@ -67,10 +67,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:09.134613Z", - "iopub.status.busy": "2023-05-11T09:19:09.134404Z", - "iopub.status.idle": "2023-05-11T09:19:09.138357Z", - "shell.execute_reply": "2023-05-11T09:19:09.137951Z" + "iopub.execute_input": "2023-06-12T07:30:58.682905Z", + "iopub.status.busy": "2023-06-12T07:30:58.682667Z", + "iopub.status.idle": "2023-06-12T07:30:58.687020Z", + "shell.execute_reply": "2023-06-12T07:30:58.686669Z" } }, "outputs": [], @@ -114,10 +114,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:09.140466Z", - "iopub.status.busy": "2023-05-11T09:19:09.140315Z", - "iopub.status.idle": "2023-05-11T09:19:09.155584Z", - "shell.execute_reply": "2023-05-11T09:19:09.155082Z" + "iopub.execute_input": "2023-06-12T07:30:58.689500Z", + "iopub.status.busy": "2023-06-12T07:30:58.689322Z", + "iopub.status.idle": "2023-06-12T07:30:58.702109Z", + "shell.execute_reply": "2023-06-12T07:30:58.701774Z" } }, "outputs": [ @@ -154,10 +154,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:09.157893Z", - "iopub.status.busy": "2023-05-11T09:19:09.157751Z", - "iopub.status.idle": "2023-05-11T09:19:09.445154Z", - "shell.execute_reply": "2023-05-11T09:19:09.444740Z" + "iopub.execute_input": "2023-06-12T07:30:58.704027Z", + "iopub.status.busy": "2023-06-12T07:30:58.703891Z", + "iopub.status.idle": "2023-06-12T07:30:58.969383Z", + "shell.execute_reply": "2023-06-12T07:30:58.968962Z" } }, "outputs": [ diff --git a/examples/example_star_dss.ipynb b/examples/example_star_dss.ipynb index 8740b958..53d8aaf6 100644 --- a/examples/example_star_dss.ipynb +++ b/examples/example_star_dss.ipynb @@ -6,10 +6,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:59.359670Z", - "iopub.status.busy": "2023-05-11T09:18:59.359514Z", - "iopub.status.idle": "2023-05-11T09:18:59.678204Z", - "shell.execute_reply": "2023-05-11T09:18:59.677848Z" + "iopub.execute_input": "2023-06-12T07:30:44.902668Z", + "iopub.status.busy": "2023-06-12T07:30:44.902522Z", + "iopub.status.idle": "2023-06-12T07:30:45.181341Z", + "shell.execute_reply": "2023-06-12T07:30:45.180590Z" } }, "outputs": [], @@ -41,10 +41,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:18:59.680439Z", - "iopub.status.busy": "2023-05-11T09:18:59.680272Z", - "iopub.status.idle": "2023-05-11T09:19:00.629037Z", - "shell.execute_reply": "2023-05-11T09:19:00.628682Z" + "iopub.execute_input": "2023-06-12T07:30:45.183867Z", + "iopub.status.busy": "2023-06-12T07:30:45.183677Z", + "iopub.status.idle": "2023-06-12T07:30:46.156686Z", + "shell.execute_reply": "2023-06-12T07:30:46.156280Z" } }, "outputs": [], @@ -77,10 +77,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:00.631283Z", - "iopub.status.busy": "2023-05-11T09:19:00.631158Z", - "iopub.status.idle": "2023-05-11T09:19:00.636200Z", - "shell.execute_reply": "2023-05-11T09:19:00.635859Z" + "iopub.execute_input": "2023-06-12T07:30:46.159017Z", + "iopub.status.busy": "2023-06-12T07:30:46.158817Z", + "iopub.status.idle": "2023-06-12T07:30:46.163992Z", + "shell.execute_reply": "2023-06-12T07:30:46.163675Z" } }, "outputs": [], @@ -135,10 +135,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:00.638129Z", - "iopub.status.busy": "2023-05-11T09:19:00.637970Z", - "iopub.status.idle": "2023-05-11T09:19:00.651068Z", - "shell.execute_reply": "2023-05-11T09:19:00.650720Z" + "iopub.execute_input": "2023-06-12T07:30:46.165866Z", + "iopub.status.busy": "2023-06-12T07:30:46.165748Z", + "iopub.status.idle": "2023-06-12T07:30:46.178321Z", + "shell.execute_reply": "2023-06-12T07:30:46.177993Z" } }, "outputs": [ @@ -175,10 +175,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:00.653041Z", - "iopub.status.busy": "2023-05-11T09:19:00.652921Z", - "iopub.status.idle": "2023-05-11T09:19:00.657103Z", - "shell.execute_reply": "2023-05-11T09:19:00.656797Z" + "iopub.execute_input": "2023-06-12T07:30:46.180262Z", + "iopub.status.busy": "2023-06-12T07:30:46.180095Z", + "iopub.status.idle": "2023-06-12T07:30:46.185174Z", + "shell.execute_reply": "2023-06-12T07:30:46.184894Z" } }, "outputs": [], @@ -204,10 +204,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:00.661814Z", - "iopub.status.busy": "2023-05-11T09:19:00.661649Z", - "iopub.status.idle": "2023-05-11T09:19:00.665687Z", - "shell.execute_reply": "2023-05-11T09:19:00.665404Z" + "iopub.execute_input": "2023-06-12T07:30:46.187037Z", + "iopub.status.busy": "2023-06-12T07:30:46.186945Z", + "iopub.status.idle": "2023-06-12T07:30:46.190735Z", + "shell.execute_reply": "2023-06-12T07:30:46.190435Z" } }, "outputs": [], @@ -232,10 +232,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:00.667650Z", - "iopub.status.busy": "2023-05-11T09:19:00.667522Z", - "iopub.status.idle": "2023-05-11T09:19:01.034238Z", - "shell.execute_reply": "2023-05-11T09:19:01.033700Z" + "iopub.execute_input": "2023-06-12T07:30:46.192440Z", + "iopub.status.busy": "2023-06-12T07:30:46.192333Z", + "iopub.status.idle": "2023-06-12T07:30:46.556903Z", + "shell.execute_reply": "2023-06-12T07:30:46.556552Z" } }, "outputs": [ diff --git a/examples/example_trca.ipynb b/examples/example_trca.ipynb index 8fdc51dd..344dacdb 100644 --- a/examples/example_trca.ipynb +++ b/examples/example_trca.ipynb @@ -6,10 +6,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:10.402060Z", - "iopub.status.busy": "2023-05-11T09:19:10.401908Z", - "iopub.status.idle": "2023-05-11T09:19:10.717107Z", - "shell.execute_reply": "2023-05-11T09:19:10.716778Z" + "iopub.execute_input": "2023-06-12T07:31:00.133822Z", + "iopub.status.busy": "2023-06-12T07:31:00.133436Z", + "iopub.status.idle": "2023-06-12T07:31:00.425668Z", + "shell.execute_reply": "2023-06-12T07:31:00.424495Z" } }, "outputs": [], @@ -52,10 +52,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:10.720417Z", - "iopub.status.busy": "2023-05-11T09:19:10.720209Z", - "iopub.status.idle": "2023-05-11T09:19:11.713046Z", - "shell.execute_reply": "2023-05-11T09:19:11.712573Z" + "iopub.execute_input": "2023-06-12T07:31:00.430365Z", + "iopub.status.busy": "2023-06-12T07:31:00.430188Z", + "iopub.status.idle": "2023-06-12T07:31:01.400476Z", + "shell.execute_reply": "2023-06-12T07:31:01.400136Z" } }, "outputs": [], @@ -89,10 +89,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:11.716988Z", - "iopub.status.busy": "2023-05-11T09:19:11.716750Z", - "iopub.status.idle": "2023-05-11T09:19:11.720779Z", - "shell.execute_reply": "2023-05-11T09:19:11.720144Z" + "iopub.execute_input": "2023-06-12T07:31:01.402645Z", + "iopub.status.busy": "2023-06-12T07:31:01.402442Z", + "iopub.status.idle": "2023-06-12T07:31:01.405425Z", + "shell.execute_reply": "2023-06-12T07:31:01.405150Z" } }, "outputs": [], @@ -133,10 +133,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:11.723444Z", - "iopub.status.busy": "2023-05-11T09:19:11.723203Z", - "iopub.status.idle": "2023-05-11T09:19:11.830561Z", - "shell.execute_reply": "2023-05-11T09:19:11.830130Z" + "iopub.execute_input": "2023-06-12T07:31:01.407479Z", + "iopub.status.busy": "2023-06-12T07:31:01.407304Z", + "iopub.status.idle": "2023-06-12T07:31:01.514788Z", + "shell.execute_reply": "2023-06-12T07:31:01.514380Z" } }, "outputs": [], @@ -176,10 +176,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:11.833176Z", - "iopub.status.busy": "2023-05-11T09:19:11.833022Z", - "iopub.status.idle": "2023-05-11T09:19:12.051286Z", - "shell.execute_reply": "2023-05-11T09:19:12.050909Z" + "iopub.execute_input": "2023-06-12T07:31:01.516929Z", + "iopub.status.busy": "2023-06-12T07:31:01.516807Z", + "iopub.status.idle": "2023-06-12T07:31:01.717448Z", + "shell.execute_reply": "2023-06-12T07:31:01.717062Z" } }, "outputs": [ @@ -243,10 +243,10 @@ "metadata": { "collapsed": false, "execution": { - "iopub.execute_input": "2023-05-11T09:19:12.053627Z", - "iopub.status.busy": "2023-05-11T09:19:12.053496Z", - "iopub.status.idle": "2023-05-11T09:19:21.237323Z", - "shell.execute_reply": "2023-05-11T09:19:21.235672Z" + "iopub.execute_input": "2023-06-12T07:31:01.719389Z", + "iopub.status.busy": "2023-06-12T07:31:01.719283Z", + "iopub.status.idle": "2023-06-12T07:31:09.975282Z", + "shell.execute_reply": "2023-06-12T07:31:09.974623Z" } }, "outputs": [ @@ -266,7 +266,7 @@ "Mean accuracy = 97.1%\t(95% CI: 97.0-97.1%)\n", "Mean ITR = 299.8\t(95% CI: 299.4-300.2)\n", "\n", - "Elapsed time: 9.5 seconds\n" + "Elapsed time: 8.6 seconds\n" ] } ],