{ "cells": [ { "cell_type": "markdown", "id": "6e9f7b0a", "metadata": {}, "source": [ "# Whole Dataset Clustering" ] }, { "cell_type": "markdown", "id": "fde8b8ae", "metadata": {}, "source": [ "### Data preperation and clustering" ] }, { "cell_type": "markdown", "id": "47e82d9d", "metadata": {}, "source": [ "**Libraries used**" ] }, { "cell_type": "code", "execution_count": 1, "id": "6381345c", "metadata": { "tags": [ "remove-output" ] }, "outputs": [], "source": [ "import numpy as np\n", "import pandas as pd\n", "\n", "import matplotlib.pyplot as plt\n", "import seaborn as sns\n", "\n", "from numpy import arange\n", "\n", "import sklearn\n", "from sklearn import preprocessing\n", "from sklearn.preprocessing import minmax_scale\n", "from sklearn.cluster import KMeans\n", "from mpl_toolkits.mplot3d import Axes3D\n", "\n", "import sklearn.metrics as sm\n", "from sklearn.decomposition import PCA\n", "from sklearn.preprocessing import MinMaxScaler\n", "from sklearn import datasets\n", "from sklearn.metrics import confusion_matrix, classification_report" ] }, { "cell_type": "code", "execution_count": 2, "id": "48981434", "metadata": { "tags": [ "remove-input" ] }, "outputs": [], "source": [ "clustering = pd.read_csv('Data\\whole_clustering.csv')\n", "cluster = KMeans(n_clusters = 4)\n", "cols = clustering.columns[:]\n", "clustering.drop(clustering.columns[[0]], axis = 1, inplace = True)" ] }, { "cell_type": "markdown", "id": "95fe6be5", "metadata": {}, "source": [ "So for my clustering analysis of the whole study I would like to see if the studies as a whole have interesting cluster patterns. To do this step I will have to map an integer value to the study so that clustering can take place. I am focusing on the profit/loss margins of the individual subjects and I am also going to look at their total zeros received. This tells us how often the subjects didn't lose money. The lower this figure is indicates that the subjects were losing money more regularly." ] }, { "cell_type": "code", "execution_count": 3, "id": "106d4bce", "metadata": { "tags": [ "remove-input" ] }, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
indexTotal WTotal LStudyMargincount_zeroscluster
0Subj_15800-4650Fridberg1150801
1Subj_27250-7925Fridberg-675713
2Subj_37100-7850Fridberg-750763
3Subj_47000-7525Fridberg-525763
4Subj_56450-6350Fridberg100761
\n", "
" ], "text/plain": [ " index Total W Total L Study Margin count_zeros cluster\n", "0 Subj_1 5800 -4650 Fridberg 1150 80 1\n", "1 Subj_2 7250 -7925 Fridberg -675 71 3\n", "2 Subj_3 7100 -7850 Fridberg -750 76 3\n", "3 Subj_4 7000 -7525 Fridberg -525 76 3\n", "4 Subj_5 6450 -6350 Fridberg 100 76 1" ] }, "execution_count": 3, "metadata": {}, "output_type": "execute_result" } ], "source": [ "y_predicted = cluster.fit_predict(clustering[[\"Margin\",\"count_zeros\"]])\n", "clustering[\"cluster\"] = y_predicted\n", "clustering.head()" ] }, { "cell_type": "markdown", "id": "c8e4087d", "metadata": {}, "source": [ "This is the results of our clustering based on the amount of zeros each subject chose and their respective margin of profit and loss." ] }, { "cell_type": "code", "execution_count": 4, "id": "6dc04421", "metadata": { "tags": [ "hide-input" ] }, "outputs": [ { "data": { "text/plain": [ "Text(0, 0.5, 'count_zeros')" ] }, "execution_count": 4, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "df1 = clustering[clustering.cluster==0]\n", "df2 = clustering[clustering.cluster==1]\n", "df3 = clustering[clustering.cluster==2]\n", "df4 = clustering[clustering.cluster==3]\n", "\n", "plt.scatter(df1.Margin, df1.count_zeros, color='green')\n", "plt.scatter(df2.Margin, df2.count_zeros, color='red')\n", "plt.scatter(df3.Margin, df3.count_zeros, color='black')\n", "plt.scatter(df4.Margin, df4.count_zeros, color='blue')\n", "\n", "plt.xlabel(\"Margin\")\n", "plt.ylabel(\"count_zeros\")" ] }, { "cell_type": "markdown", "id": "dfb1926f", "metadata": {}, "source": [ "### Normalization and refined clustering" ] }, { "cell_type": "markdown", "id": "a5ba0762", "metadata": {}, "source": [ "Below we can see the dataframe after normalization has taken place. The overall aim of normalization is to manipulate the values of the choosen columns in a particular dataset to a common scale. In machine learning normalization can improve learning rates and can also make weights easier to initialise.\n", "The main reason I am using normalization in my analysis is that I want to investigate whether it improves model accuracy dramatically or whether the results are very similar. I am also curious to see whether there is one variable that is steering the performance." ] }, { "cell_type": "code", "execution_count": 5, "id": "6001a0e8", "metadata": { "tags": [ "hide-input" ] }, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
indexTotal WTotal LStudyMargincount_zeroscluster
0Subj_15800-4650Fridberg0.6750000.4044940
1Subj_27250-7925Fridberg0.4468750.3033711
2Subj_37100-7850Fridberg0.4375000.3595511
3Subj_47000-7525Fridberg0.4656250.3595511
4Subj_56450-6350Fridberg0.5437500.3595511
\n", "
" ], "text/plain": [ " index Total W Total L Study Margin count_zeros cluster\n", "0 Subj_1 5800 -4650 Fridberg 0.675000 0.404494 0\n", "1 Subj_2 7250 -7925 Fridberg 0.446875 0.303371 1\n", "2 Subj_3 7100 -7850 Fridberg 0.437500 0.359551 1\n", "3 Subj_4 7000 -7525 Fridberg 0.465625 0.359551 1\n", "4 Subj_5 6450 -6350 Fridberg 0.543750 0.359551 1" ] }, "execution_count": 5, "metadata": {}, "output_type": "execute_result" } ], "source": [ "clustering[['Margin','count_zeros']] = minmax_scale(clustering[['Margin','count_zeros']])\n", "km = KMeans(n_clusters=4)\n", "y_predicted = km.fit_predict(clustering[[\"Margin\", \"count_zeros\"]])\n", "clustering[\"cluster\"] = y_predicted\n", "clustering.head()" ] }, { "cell_type": "code", "execution_count": 6, "id": "1204346b", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "array([[0.69183361, 0.31542525],\n", " [0.50055668, 0.34221899],\n", " [0.53452744, 0.79761579],\n", " [0.31934307, 0.33429017]])" ] }, "execution_count": 6, "metadata": {}, "output_type": "execute_result" } ], "source": [ "km.cluster_centers_" ] }, { "cell_type": "markdown", "id": "a561642a", "metadata": {}, "source": [ "The below graph differs from the original in two mains ways. As I have mentioned already the data is now standarized, this should imrove the overall clustering of the dataset. I have also calculated the centroids of the clusters and added them to the graph giving us some added information." ] }, { "cell_type": "code", "execution_count": 7, "id": "e0f5e259", "metadata": { "tags": [ "hide-input" ] }, "outputs": [ { "data": { "text/plain": [ "" ] }, "execution_count": 7, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "df1 = clustering[clustering.cluster==0]\n", "df2 = clustering[clustering.cluster==1]\n", "df3 = clustering[clustering.cluster==2]\n", "df4 = clustering[clustering.cluster==3]\n", "\n", "plt.scatter(df1.Margin, df1.count_zeros, color='green')\n", "plt.scatter(df2.Margin, df2.count_zeros, color='red')\n", "plt.scatter(df3.Margin, df3.count_zeros, color='black')\n", "plt.scatter(df4.Margin, df4.count_zeros, color='blue')\n", "\n", "plt.scatter(km.cluster_centers_[:,0], km.cluster_centers_[:,1], color=\"orange\", marker=\"*\", label=\"centroid\")\n", "\n", "plt.xlabel(\"Margin\")\n", "plt.ylabel(\"count_zeros\")\n", "plt.legend()" ] }, { "cell_type": "markdown", "id": "494c97be", "metadata": {}, "source": [ "### Further analysis" ] }, { "cell_type": "code", "execution_count": 8, "id": "79f6b23d", "metadata": { "tags": [ "remove-output", "hide-input" ] }, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "C:\\Users\\User\\anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:881: UserWarning: KMeans is known to have a memory leak on Windows with MKL, when there are less chunks than available threads. You can avoid it by setting the environment variable OMP_NUM_THREADS=3.\n", " warnings.warn(\n" ] } ], "source": [ "k_rng = range(1,10)\n", "sse = []\n", "for k in k_rng:\n", " km = KMeans(n_clusters=k)\n", " km.fit(clustering[[\"Margin\", \"count_zeros\"]])\n", " sse.append(km.inertia_)" ] }, { "cell_type": "markdown", "id": "694e3b35", "metadata": {}, "source": [ "Below is a diagram of the elbow method. This enable us to find the optimum number of clusters in the dataset. It is the most popular method when dealing with k means clustering to calculate this. From the graph below you can the elbow starts to bend at 3 indicating that the optimum number of clusters would be three and that the results may improve with a correction." ] }, { "cell_type": "code", "execution_count": 9, "id": "069a9c29", "metadata": { "tags": [ "hide-input" ] }, "outputs": [ { "data": { "text/plain": [ "[]" ] }, "execution_count": 9, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "plt.xlabel(\"K\")\n", "plt.ylabel(\"Sum of squared error\")\n", "plt.plot(k_rng, sse)" ] }, { "cell_type": "markdown", "id": "3b9f6733", "metadata": {}, "source": [ "Below is the silhouette score for these clusters. The silhouette score is a metric used to calculate the efficiency of a certain clustering technique. The closer the the silhouette scores are to 1 means that they are further apart from eachother. The scores below are not below 0 which is good and tells us that there arent any overlapping clusters. The closer to 1 indicates the more dense clusters, which we don't seem to have in our case. " ] }, { "cell_type": "code", "execution_count": 10, "id": "e20ac3d7", "metadata": { "tags": [ "hide-input" ] }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " Silhouette Score: 0.577\n", " Silhouette Score: 0.428\n", " Silhouette Score: 0.356\n", " Silhouette Score: 0.375\n", " Silhouette Score: 0.372\n", " Silhouette Score: 0.338\n", " Silhouette Score: 0.350\n" ] } ], "source": [ "from sklearn.metrics import silhouette_score\n", "for n in range(2, 9):\n", " km = KMeans(n_clusters=n)\n", " km.fit_predict(clustering[[\"Margin\", \"count_zeros\"]])\n", " value = silhouette_score(clustering[[\"Margin\", \"count_zeros\"]], km.labels_, metric='euclidean')\n", " print(' Silhouette Score: %.3f' % value)" ] }, { "cell_type": "markdown", "id": "a8bb9ffe", "metadata": {}, "source": [ "Here we have an informative scatterplot of the different studies and the participants. You can clearly see that the subjects from both Steingrover and Wetzels more often than not did not lose money and still gained a respective amount of money. Another interesting observation is that some of the participants that were not receiving many zeros, so as a result were losing money in some cases still gained a large amount of money. This tells us that these participants found the more beneficial cards but reaped the downside to those cards also." ] }, { "cell_type": "code", "execution_count": 11, "id": "0e000cc7", "metadata": { "tags": [ "hide-input" ] }, "outputs": [ { "data": { "text/plain": [ "" ] }, "execution_count": 11, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgUAAAEGCAYAAAD14OY+AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAADSRklEQVR4nOydd3xUVfr/33d6n8mk94SQQiD0roiIBRULKquufV1dbOuudd11m67+rF8Ve1l17djFioqKBVE6oaST3uv0fn9/TBIyJECQhIDc9+vlS+6Zc889M5m59znPeZ7PI4iiiISEhISEhISEbKQnICEhISEhIXFoIBkFEhISEhISEoBkFEhISEhISEh0IxkFEhISEhISEoBkFEhISEhISEh0oxjpCfwSYmJixIyMjJGehoSEhMRhw/r161tFUYwd6XlIHNoclkZBRkYG69atG+lpSEhISBw2CIJQNdJzkDj0kbYPJCQkJCQkJADJKJCQkJCQkJDoRjIKJCQkJCQkJADJKJCQkJCQkJDoRjIKJCQkJCQkJIBhzj4QBOF5YCHQLIriuAFeF4BHgFMAF3CpKIobhnNOI4XbXYPNvpVQ0I3ekIPRkI8gSDaZhEQPQVFkq91NsdODXi5jvFFLqlY90tOSkDiiGO6UxBeBx4CX9vD6yUB2938zgCe7//+rwuWqYvPm3+FyVwIgCEomTnwRa9TMkZ2YhMQhxA8dDs7fUk6wu3Brnl7DSwWZpEmGgYTEQWNYl6qiKH4LtO+lyxnAS2KYNYBFEITE4ZzTSNDVtaHXIAAQRT8VFQ8TCLhGblISEocQXf4Ad5bX9RoEAEVODxtt0m9EQuJgMtL+62Sgps9xbXdbPwRBuFIQhHWCIKxraWk5KJMbKry+/vP1eGoIBqUbnoQEgDsUotbj79fe7g+MwGwkJI5cRtooEAZoEwdoQxTFZ0RRnCqK4tTY2MNLqdNsmtCvLSnpPNTqmBGYjYTEoUecSskFSdH92scatCMwGwmJI5eRNgpqgdQ+xylA/QjNZdgwmSYwduxS1OoEZDINaWm/JzHx7JGeloTEIYNMELgkKZorUmLQyARS1EqeH5vBBKNupKcmIXFEMdK1D5YD1wqC8AbhAMMuURQbRnhOQ45criEh/lSiLDMQRT9qdRyCIB/paUlIHFKkatX8MyuZJalxaGQC0SrlSE9JQuKIY7hTEl8HjgViBEGoBf4JKAFEUXwK+IRwOmIZ4ZTEy4ZzPiONtF0gIbF3FDKBZI1qpKchIXHEMqxGgSiK5+/jdRG4ZjjnICEhISEhITE4Rnr7QOIIp8XnZ22Xk002F2MMWmaa9SRKK0UJCQmJEUEyCiRGDHcwyP9VNvFCXWtv24IYEw/npWFRSl9NCQkJiYONdOeVGDF2un282McgAPis1Uapy8s0s/TVPBJxOstwOiuQK3QY9Lmo1QeWfiyKIk5nCS7XThQKMwZDLiqVdYhmKyHx60O680qMGN5QaEBRCm8odNDnIjHydHZuYOOmiwmF3ABERc0mf8z9aDQJv3jM9o7VbN78e0TRB0Bc7Mnk5v4LlUoK+pWQGIiR1imQ+BXjDgbZZndRaHfhCAR7232+drpsW0iS2ZmyWx56qkZFlk7Suj/SCASclFfc12sQAHR0rMZm3/KLx/T5Oigu/kevQQDQ3PIpNvv2A5qrhMSvGclTIDEs1Ht83LOzgTcbOwA4OcbEv0cnExWoYOu2P+FylSEIau4e9zZvdUTzeZud2RYDf0iNJVEtBRoeaQSDDhyO4n7tXm/jAYxpx92n5kgPfl9r/84SEhKA5CmQGCZWtdt7DQKAT1ttfNzSwc7Kx3G5ygAQRS+thadxY1wHK6bkcH9uKmN+pbK2vlCIeo8vwmMykohiEI+nEb+/66Bd0++34fE0Egr1r2egVFqJizu1X7tel/2Lr6dSxRIdfVy/dp0u8xePKSHxa0cyCiSGha/abf3aPm6x4Qv0b3e7S4lSKVDKBiqFcfhT6vRwQ1E1c34u4oItFazrco7ofNzuWkpL7+GnnxewYcP5tLV9N+CDeqgQxRDt7T+yceOFrPnpJEpL78TlqoroI5MpSUu7vPchLpfrycn5NybT+F98Xblcy+jRt2KxTAdAoTCRP+ZBDIb8X/5mJCR+5UjbBxLDwnSzng9bIlehc6IMKGz9v3JqdfzBmtZBx+YPcnNxDWu6DYGfupyct7mcFVNzyNJpDvp8QqEA1TX/pbb2JQAcgWI2b7mcqVPewWQqGJZrOhzFbNp8GaIYroJYW/cKgaCDvNz/h1y+a6tIr8ukYNwjuD11yGUatNrUPQ05aAz60UwY/yweTwNyhQ6tZsAirBISEt1IngKJYeH4aDOTjLu2ArJ1as6Kt5KRcQ0KhbG3PTnptxh/xSu3Wq+v1yDowREMUe7yjsh8vL5m6uuXRbSJYhCHo2TYrul0lfUaBD00Ni7H6+1f5kQu12HQZw/aIBBFkQ1dTm4rqeHKrTv5qs2Gc7ctGoXCgMGQLRkEEhKDQPIUSAwLmTo1L48fRYnTSwiRbJ2GeLUSmMK0qR/gdlehUBjR67NRKAwjPd1hQyOToZEJeEKRyZdG+cgUxJLL1KhUMXg8dRHtfQ21oUYh7//3VSqjkMsP3FOyxe5m0aYyvN2f7/KWLv47LoNTYy0HPLaExJGIZBRIDBsxKiUxA1S60+nS0enSR2BGB58MrYrbRiXyz7JdFcEXxJjI1Q9v2qXLVY3dvpVQyIPekIPRkI8gyFCposnO/juFhVdBt0qEXp+H0TR22OZiMIzBZJqEzbaxty0n+/Yh2Tb6rsPeaxD0sLSyiYlCMYpgCwZjPnrdqAO+joTEkYJkFEhIDCMyQeC3idGMM2gpc3pJ0CiZYNRiHcaywC5XJRs3XYrHUwOAICiZNPF/REXNACAmei5TJr+Jw1mMUmnBZCwYVte6RpPAuLGPYLdvxedvQ6/PxmQcmviFkNhf/ipEiLr6V7G3foxKGc2kSa9gMOQMyfUkJH7tSEaBxIB4vM24nBXIZAp0+tGolJaRntJhi1Eh56goI0dFRbro3YEgpW4vHb4AaVo1mfsQbXK5a/C4q5ArTOh1o1EodAP26+xc22sQAIiin4qdDzPR9F/kch0ymQqLZTIWy+QDf3ODRKtNRqsdWsMjEHAw2ySgFAT8fYyD38d6cFauBMDnb6OtfZVkFEhIDBLJKJDoh8NRSuHWq3G5KgCwWueQl3vXkN/Uj2TsgSBPVjfzf1VNABjkMl4qyGR21MB7+11dG9m0+fcEAp0ApKZcSmbmdSgHMNa8vpZ+bR5PLcGgB7l8YEPicMPrbaOi4kHc7at4cdQ9vG+LoTOk4pwoF9HND+ANeXr7ul01exlJQkKiL1L2gUQEohiirv6NXoMAoL39Ozo6fhzBWf362O5w9xoEEM5IuKGohlafv19fv7+L4pJ/9xoEADW1L2K3bxtwbIt5Sr+2pKTfDnshoEDAgdNZjtfbtO/OB4jdsZX6hmX4vI1QdikXOG/gH7q3mCIU4u36PqJvTGx/ASMJCYmBkYwCiQiCQTcdHav7tdtsv1yDXqI/Dd7+D/9Kj48Of3/Fw0DAht1e2K99Tw9fk2kCY8cuRa1ORC7Xk5F+NYkJiw580nvB4Shh85YrWfPTify89nSaWz4nFOr/HocKj6c+4tjpLKG29hUslhlkjboJhcKEShXHmLx7sZinDts8JCR+bRyR2wd+vx25XI1MJmns745criMm5niczsi8dYtlaG+sR/rfIFXT/33n6zXEqCJ/kq5gEFFuJSpqVj9vjWYPwYFyuYaE+FMxmI8iIAYxa6IQhOGz/wMBByUld9DZ+RMAPl8rhYXXMG3qe5hM44blmjptWr82q/UotNokMjKuIiFhEYIgP+DSy7921q9fH6dQKJ4DxiEtEo8EQsDWQCDw+ylTpjQP1OGIMgo8ngaamj+mof4tdLpRpKf/AbN54khP65BCEASSEs+mq2sdnZ0/A5CYuLg3cv1A8XgaaW7+hPr6ZWh1o8hIvxKzedKQjH04ka/XcHd2Mv8qq8cniiSqlTyYl0qUMvyTtPmDfNVu44nqZrRyGdemPoSZv2HvWIkgKBiVeQNG48BphIGQyJouB49WtdLsC3BFCpwUYyJ6mDIevN5mOjp3314K4XJXDptRYDCOIzPjOiqrnkAUg+h0WWSNugG5PCyYdSDllo8kFArFcwkJCWNiY2M7ZDLZQJXMJX5FhEIhoaWlJb+xsfE54PSB+gjiACk9hzpTp04V161bt1/niGKQsvIHqK5+prdNJtMybeq7UmTyAPj9NlzuSmSCEp0uo/dmeyCIYoiKioeorHqit00m03T/DXIPePzDjaAoUuHy0hUIkqJRktCnOuSHzZ1csa0yov/bE9IpkFcilxvR6dKRyQZ+yK/vcnL6xlKCfX7a9+ekcFFyzHC8DTyeJtauOxOfL3LhMWniy1its4flmgChkA+Xq5Jg0IVWm4pKFT1s1/o1IAjCelEUI1x+mzdvrigoKJAMgiOIUCgkFBYWRk2YMGFAAY8jxlPg8TRQW/tiRFso5MbhLJGMggFQKk2Ylb+sGM0Oh5vvOhx0BQIcE2VkkkmHSibD42mguub5iL6hkAeHo3ifRoHNto32jh8QQz6s1qMwmcYjCAOrAjZ4fPzU5WSz3cVkk57pZh3xh2A55mq3j7VdTircXmZZDGhkMixKBd5giGdq+nv2Pmt1cnTOvr0qP3U5IwwCgMdrmlkYZ+n1ROwNt7uGjs6fcTnLMJunYrZMRqWM2mN/jSaevNw72VJ4FWHvJMTHnYbBkNfbp8Ll5YdOO9VuH7MtBqaY9ZgUB6bqKJOppN/ugSOTDIIji+6/9x63io4Yo0AQFMhkWkIhX0S7TDj0HhaHMzscbs7cWEZXt/78/1U28dr4UcyLNiGTKZDLdYT6pIsB+4wrsNkKWb/hfEIhNwA7K5cyaeIrREVN79e3yx/g9tI6Pm7tKcbUwm8TrfwnOxndCEkLD0Sdx8dlW3dS5Ax/Fo9VN/PvrCSuTI1FJoBlgIe3WTm4+evk/X/vJoUchbDvKpRebzNbt12Pzba5u+UZMjP/RGbG1Xs0wgCio49l2rT3cbsqUSqjMBjyerMdatxeLi4sp8wV/u09Wt3MPTnJXJos7fdLgFwun5Kdne3uOf7ggw/KcnNzI27Uc+fOHf3OO+/sjImJiYjEveGGG5IMBkPwjjvuaJo+fXruAw88UHPMMce4Dtbcf40cMYElGk0Co7Nu3q0tFaNxzAjN6NfJ6k5Hr0EAYSHdBysbcQaCqNXxjM66JaK/Wp2MwZBPIOCgs3M9TU2f0NW1mWBwl+HQ0vJFr0EA4a2gmtoXByz3W+by9jEIwrzW0H7QCxA5HGU0N6+gre17vN7+ugHbHe5eg6CHeysbqfH4UMpkLEmNjfhx6mQyTowxD+raM816zLutwm/OSMA4iJW5w1nSxyAIU1X1BG53NW5PPa2tX9Pc8gUuV2VEH5lMgck4lvj4U7FaZ0ekP25zeHoNgh7urmikzhPZJnFkolarQ0VFRdt7/utrEIRCIYLBIKtWrSrb3SAYSvz+4cuUOdw4YjwFAPHxC1Grk2jv+AGtNg1r1OwhKc8qsQtHoP/vtjMQJNitsx8XdzIqdTwd7avRaJOxRs1GpYqluuZZdu58pPecvNz/kJT0GwRBjt/f2W9Mv7+dHld1X7yh/m0AvtDB85B2dq5n46ZLeg0Zi2U6Y/MfRKNJ6u2zu14/gCcYItAd4zPDbOCDSaP5psOORiZjrtXIeOPghIfyDFremzSaVe12OvwBjrWamGwa3LmhUH/jKRTy4ffb2L7jZlyuciBc0GjSxJcwGvdd4XKgv4k7GIpQIZQ4PHhlTZV16crS5Ba7VxVrVPv+OD+77sKZ6e1DeY3i4mLVySefnD179mz7+vXrDR988EHZvHnzctetW7cjMTExcOuttyYsW7YsJikpyRcdHe2fNGlSr2fgxRdfjL7++uvTHA6H/Jlnntk5b948l81mk11++eVpO3bs0AaDQeFvf/tb/YUXXti5dOnS6E8//dTs9XplLpdL9sUXX5Sde+65GWVlZZrs7GxPTU2N6rHHHqs+0jwPR5RRoFAYiYmZS0zM3JGeyiGP32/D7a5CEJQEg24UCh06XeY+Xf2zowzIdkY+rpekxmJShL9qCoWBmOhjiIk+pvd1m20rO3cujRinpPROLJYZ6PWjiItbQF39qxGvp6RcMuBcsnQaRmlVVLh3rUInGLVkaoe3AFEPgYCT8or7IzwbnZ0/02XbEmEU5Ok1GOUy7MFdn9R5CVaSu2MfFDKBaRYD0yy/rIJkvkFLvmFXcGgg4MBmq0QQZN2BowMbCXrdaJTKKPz+jt62hPizsNk29RoEAH5/B7V1r5KXe8detxUgbKTo5DJcfd7rxUnRve9V4vDglTVV1js/2p7uDYRkAM12r+rOj7anAxyIYeD1emV5eXn5AKmpqd7HHnusprKyUvPss89WvvLKK9V9+3733Xe69957z1pYWLjd7/czceLE/L5Ggcvlkm3cuLHo008/NVx55ZWZpaWl2/76178mzps3z/bWW29Vtra2yqdOnTrm9NNPtwFs2LDBsGXLlm3x8fHBf/zjH/EWiyVYUlKyfe3atZpZs2YNX5WwQ5gjyiiQGBwuVyVFxX/HGjWbpqYPcTiLEQQ5aWlXkp52Ocq9BJ1NNOp5Y0IWD1c20h4IcmVKLCfFmPZ6vfADKHLVGAp5exX8zObJTBj/HDsrHycU8pCediXWqKMGHCtereSFgkyer23l2w47x0ebuDgpBqvq4HzVAwEHDkdpv3aftzHieLRew5sTs3iiupltDg+LE6I4Jz4K9QDxAAeKy11DScm/aGv7BoCEhEVkjbppwLQ9nS6dSRNfpKrqv9jsW4iPX0hS4rlUVj3er6/NtoVQyLtP6eRcvYa3J2bxWFUTxU4v5yVaOSs+CqVs3zEOEocOS1eWJvcYBD14AyHZ0pWlyQdiFPRsH/QcFxcXqxITE33z58937t7366+/NpxyyimdRqMxBHDiiSd29n39t7/9bTvAySef7HA4HLLW1lb5N998Y1qxYoVl6dKlCQBer1coKytTAcyZM8cWHx8fBFi9erXh+uuvbwaYNm2aJycn54jyEPQgGQUSEYiiSEPjewQDDrpsm3E4i7vbg1RVPYnFMo2Y6P6eli5/AEcwRJxKwTFWI9PNOvwhMA4iOE6rTUEu1xMM7roHqNUJveI8crmGmJh5REXNQBRFFAr9XsfL1Wu5KzsFRzCIUSFHPogAu6FCpYomPm4hdfWvRLT3jcTvYZJJz+P56biDIUQRXKEQHl8nguhFpYpDGKJ5Nzd/3GsQADQ2vkdU1CySEs8esL/ROI78/PsIBl0oFCYEQSDaegz19W9E9EtIOHPQtRQmm/Q8mZ+BJxTCPIgMCIlDjxa7d0DXzp7aDwSdTjfwPiDs9Xex+2uCICCKIm+//XbZhAkTIvbGvv/+e33f6xyO6fnDwRETaCgxOIJBFy0tn2M0jaOrq78WhNNZFnEsiiI/dthZvLmcOT8VcXNxLRUuLxq5fFAGAYBOl8n4gqd73es63WgKxj2OWh0f0U8u1+3TIOhBIROwKBUH1SCAcMBdWtplxMScCIBcbiA3506MeygVLCCwzubitA2lHPNzEXeU17Oh5lN2Vj42JDUEQiEfLS2f92tvb/tuH+9DiVJp7r3JWizTyMz8EzKZCkGQk5R0HnFxJ+/XXNRymWQQHMbEGtUDRobuqX04OO644xwff/yxxeFwCB0dHbIvvvjC0vf1119/PQpgxYoVBqPRGIyOjg7OmzfP9uCDD8aHumNbfvjhhwFFV2bPnu144403ogDWr1+vKSkpOXBxlsMQ6RcqEYFcriUqaiZORwkGw5h+dRB2l5ctdno4b0tFb+DcG43tdPgDPDU2He1+pABarbOYOuU9/IEOVMpYVCrLAb+XkUKny2Dc2P/D46lHJlOj1abssW+h3cVFWyp6YzCeb/RB/FQW2m5GIdeRlnb5Ac1FJlMRFTW7X0aB2dK/aNLeUKmsZGZcTUL86YhiEK025YiVqD5S+eP87Lq+MQUAaoUs9Mf52XUHaw5HH320a9GiRe3jxo0bm5yc7J0+fbqj7+tRUVHBSZMm5fUEGgLcc8899VdeeWVaXl5eviiKQkpKivfrr78u233sm2++ueU3v/lNRk5OTv64ceNcubm57qioqGHLeDhUOWIUDSUGj8NRwpbCq0hJvoCq6qfx+VqBsCBNdvbfIvTkB1LeA/h+eh6j9Zr9v7azlKamj7HZNhIft5Do6GP6eQwOJpUuL5+1dfFVm4350SYWxJhJ38+gRY+nkda2r2lp+QyLeSpxcaeg12cB8Fp9GzcUR5b21ctlvJr4A6HWt5k29T2Uyr3HZOwLp7OcwsJrcLrCsQ5m8xTyxzyATte/fsC+CAZ9dHWto65+GTJBQVLSbzCbJ+9RXXE4KHK4ea+5gy12N4viLBxrNRGnPnjXP1zZg6Jh5YQJE1oHO8bByD4YKQKBAD6fT9DpdOK2bdvUJ554Yk55eflWjUZz+D0k98HmzZtjJkyYkDHQa5KnQKIfBkMOUya/htO5k/EFT+H321Aojeh1WSiVkbnyxgEC48wKORr5/rvt3e46Nm26HK83vPBob/+e1JTLGD361oP60Omh0x/ghuIaVneGFyPfdjhY2WbjmbEZA4oLDUQw6GVn5aO9+/Ht7d/T2Pg+kya9gkaTOKCqX4JSjuBvQKsZmtW4Xp/FpEkv43SWIQhy9PrRv7iMclfXOjZuuqj3uLHpAyZPep2oqGkHPM/BUOXycu7mcpp8YY2Kr9vt/Ck9jpszEw/6VtGRyIUz09t/LUbA7tjtdtmcOXNy/X6/IIoiDz30UNWv0SDYF8NqFAiCsAB4BJADz4mieM9ur5uBV4C07rk8IIriC8M5J4nBoVbHD2qFnm/QMifKwHcdu7x4/x6dRIpm/1MAnc6SXoOgh9q6l0lOuQC9LnO/xztQylyeXoOgh287HFS4vUwepFHgdldTX/9mRJvLXYnTWYZGk8gEo5Yxeg07uoWMDHIZ/xltpdRxFkpzHFq3SN5eshK93jbsjq14PQ1otWkYjeMG9Cyo1bEHXDFQFEVqal/evbU7cPHgGAXbne5eg6CHJ2taOC8hmgzdwUk7lfh1EhUVFdq6deuOkZ7HSDNsRoEQTl5+HDgBqAXWCoKwXBTF7X26XQNsF0XxNEEQYoFiQRBeFUVRkjo7TIhTK1mal8Zmh4tWX4DROg0TDEMcn3OIbXHt/3QGOiHclqpV81JBJpvtbjr9HjI0Ki7fXkdXAKAZvbyVdyaOZuIA4kOBgJ3yivtpaHirty0z889kpP/hoHpWxAHfn4SExOHIcGYfTAfKRFGs6H7IvwGcsVsfETAK4RBnA9AO9NeulTikSdSoWBBj4cKkGGZaDGj3o9CNKAZxOEppa/sWlSoGtTo54vWU5IsOmupktdvLqnYb67uc2PxBsnQaZpkjsx3mWAxk7ceKVKtNIynpN7u1ZaDXZ/cep2rVLIyzcGFyAt93eSNkop3BEK82tA04ttNZFmEQAFRWPtpPgnioEASB1JSLdm8lMWHRsFxvIPL1WuJ205xYkhpLqlYKepSQGAqGc/sgGegbQVULzNitz2PAcqAeMALniqI4YH6qIAhXAlcCpKXtf4CUxKGHKIo0N69g2/YbEEU/gqCmYNxj2Gybsdk3ER93KtHRcw/KqnezzcVvt5TT5g8/kM9NiOL2UYnclR7g83ZY7dBytMHNCVEBzIrB713L5Wri405DqYzCZtuMTpeJ1ToHhcIyYP+dLk+/tnKXh5AoItttzzwQcPTrK4oBgsHh01wxm6cyaeLL1NW/gSAoSU76DWbzvis3DhXpOjXLJmTxflMnm+0uzooPBxpK8QQSEkPDcBoFA/1Kd/czngRsAo4DsoAvBEH4ThRFW78TRfEZ4BkIZx8M7VQlRgK3u4odRX9BFMPFSETRS+HWJcyY/glZWX8e0mt5PPUEgy7U6gQUishNemcwyH/K63sNAoBljR2cGq1Fse0sJisMHK3NwNW4k8Y6J2nTPurNHtgTwaAbj6eeUChAadnduN3VGPQ5tLaspK7uNaZOfQ+zqX9p6kXxVj5oiSzodEFidIRB4Pd34fU1o1bH95Mk1utzhtWzIpersFpnY7XO3mffVp+fNn+AaKWCGNXQGXZjDFrGDPUWlYSEBDC82we1QN+7Uwphj0BfLgPeFcOUATuB/tJvEr9KfP72CBVDCG8neDwNQ3aNYNBHY+Nyfvr5NNb8dBJbCq/qJ8DU5Q+y3t5/dV3n9SOKAfz+drpsG/D7OwiFfL0pmnvC5apk67YbWPPTiazf8Buio+diNOTRZduA1xcWJPLtQZhopkXPA7kpxKsUxCgV3JmdzDyrsfd1u30HGzddxk8/LWBL4RLG5j+ExTIDmUxDTMwJjBv7yC/OLBhK1nQ6WLi+lLk/F3PGhlLWdvb3akhIAOh0ughX09KlS6MvvvjiA3YHFxcXq5566qmR/zEcZgynUbAWyBYEIVMQBBVwHuGtgr5UA/MBBEGIB3KBimGck8QhhFoV38+NLpOp0GgSh+waDsd2tm3/c28dhY6O1ZSV3UcwuKtgkVWpYJ61f4h/plaNXB6ptSCXG/aalREKBWhr+w4x5O0uJuWgqupJoiOkoWVoNAMLGlmUCi5MiuGLabmsnJbLFSmxWLtX2T5/J9t33IrdHhYicrur2LzlSnJy/sWsmV8wbuxDGAw5g/lY9ok/FKLLH9hv6VdbIECN28ulhTup7C6NXO72cdnWSmo9B7d8tcSRwZ7KHpeWlqqXLVsmGQX7ybBtH4iiGBAE4VpgBeGUxOdFUdwmCMKS7tefAu4EXhQEoZDwdsOtoigOWkhD4vBGq01m3LhH2LbtT/j9HcjlBvLH3ItON2rIruFy7ezX1tq2Eq+3GZ0uHQCNXMbNGYmUu7wUOb3IBfhzejwTTWYCY5eybfsNBAI2FAoTY/MfQqfLGPBafn8XLS1fUF//OjK5lqysm2hoeAenswR/wI5MpgZk5OXeuc/th7gB3O1ebyMOx7aINlH04XbtxBh30uA+kEFQaHfxRHUzm+wuFsVFcW6idZ+CTTVuL+80dfBmYwe5eg1/zojnwcombN1Bk63+ADUe/y9KVZU4hFj7Xyur7k3G0azCEOdj7q11TLt82HQLSkpKVJdccklGW1ubIjo6OvDSSy9VZmdn+84+++yMqKioQGFhoW78+PGuM888s/PGG29Mg3BA7OrVq4v+9re/JVdUVGjy8vLyzz///NaoqKjg8uXLLaFQSCguLtZec801jT6fT7Zs2bJolUoV+vzzz0vj4+ODDz74YMwLL7wQ6/f7hYyMDO/bb7+902g0hs4+++wMo9EY3Lx5s76lpUV555131l522WUdH330kfGOO+5Islqt/uLiYm1BQYHr/fff3ymTHZ5VBIZVp0AUxU+AT3Zre6rPv+uBE4dzDhKHNtHWo5k27QN83maUSmvvg3qoUKqi+7VpNGkoFMaItjyDlncmjqbK40MvkzFKp0Ypk0HMPKZPW47P14pKFbPX/fq2tq/ZUXRr77HNtpnRWbdSVn4PRkM+kye9ikJhQafL+EXFjhQKIwqFpdfr0fse91K1cn+p7BYHau+Or/i/qiZKXV6Wjkndo2y1LxTikeomXqkPPxsq3F6+77BzRWos/1cZ3iZRCBC1H1kpEocga/9rZcVt6QS84aedo0nFitvCP9gDMAz6lk4G6Orqkp9wwgldAEuWLEn77W9/23bddde1Pfzww9FXXXVV6pdfflkOUF5ervnhhx9KFAoFxx133OilS5dWnXjiic6uri6ZTqcL3XXXXXUPPvhgfI+k8dKlS6NLSkq0mzdv3u52u2W5ubnj/v73v9ft2LFj++WXX5769NNPR//jH/9ovuCCCzpuvPHGVoA//vGPSUuXLo3529/+1gzQ1NSkXLduXdGmTZs0ixYtGn3ZZZd1AOzYsUO7adOmioyMDP+UKVPyvvjiC8NJJ510WO6ZHZ6mjMRhj8/fQUvLF5SU3kV7+/coldFDbhAAmIz5xMYu6D0WBCV5uXcMuO8erVIy2aQn16ANGwTdaLWpmM2T9moQBIMuqqt3190ScbrKiI1dgMUyBbN5Enp95i+ufqjVJJOXeyd9f7bJSb/FYMjd4zmbbS4eqWzkgZ0NrOtyEAjtfTugxOXpNQh6+LClk2r3nqVDaj0+Xm+IfCbYg6GISON/ZCXtVyqnxCHIqnuTew2CHgJeGavuTd7DGYOip3Ryz3+33XZbb+zZxo0b9VdeeWU7wFVXXdW+fv363n2+s846q0OhCK9rZ86c6bjppptS//Of/8S1trbKlcqBA1tnz55tj4qKCiUlJQUMBkNw8eLFnQAFBQWuyspKNcD69eu1U6ZMyc3Jycl/5513ordt29a7h3j66ad3yuVypkyZ4mlra+u9SEFBgTMrK8svl8sZO3asq7y8/LDNkZVkjiUOOqIYpK72VSp2PtTbZjSOZXzBs2g0Q1vnQKWKIS/3TlKSLyAQ6EKry8SgH5p990gU/SSgAdTqJLJGXYBaHTMkV4mNPZ5pU9/H7a5CpYrGYMgb8LoAm2xOFm0sw91tCDxU1cQ7E0cz07JniUS1rL/BopYJKAdo70EpyNDIZDiDkdnE2ToNz4xNJ1mtYoxeE2FoSRyGOJoHftDtqX2YMRgMvV+4u+++u/HMM8/s+uCDD8yzZ88e89lnn5UMdI5Kpeq1imUyGT0yxjKZjEAgIABceeWVmW+//XbZrFmz3EuXLo1etWpVr1uxr+xx33gbtVrdeyCXy3vHOhyRfqUSBx2Xu4bKqscj2uz2bTidA/6ODxiVyorVOpu4uJMxGvIQhMF/7QMBF51dG2hs+oiuro0EAs4B+8nlKtLTr6LvT0ou1xMXe8KQGQQQDsQ0mcYSH38KUVEz9mgQAHzU0tVrEAAERXi2poXgXoIH8/RaJhsj0/2uT4/fa0xBikbJrZkJu42jYbpZz+lxUUwx69FJWweHP4a4gd1Fe2ofAiZNmuR87rnnogCefvpp69SpUwd0yW/btk09ffp091133dVYUFDg3Lp1q8ZsNgcdDsd+f/FcLpcsLS3N7/V6hTfeeOOIC1SUPAUSBx8xQCjUP2I4FDq0otNDIR+1dS9TXn5fb9vo0beRmnLJgIJKFstUpkx+g/aOH5DLdVijZmM05vfrd7Doq4zYQ0cgQEgU9yj2E69W8uTYDH7sdFDi9DDDYmCaWbdXcSBBEPhNgpVRWg0/dtoZpdMw22IgSXPYelAlBmLurXURMQUACnWIubcOW+nkJ598svqSSy7JeOSRRxJ6Ag0H6nfffffFrV692iSTycScnBz3Oeec0yWTyVAoFGJubm7+b3/729bBlkH+y1/+Uj99+vQxycnJvjFjxrh+iWFxOCOVTpY46ASDHnbsuI2m5l0ZqgqFhWlT39ljZP9IYLcX8fPa04BdbnFBkDNt2ocY97KPD9Dm81Pl9qGVyxilVaPuribpdtfg9bWgUsb0K10cDHpwuSsJBb1odemolJYDmv93HXYWbyqPaPvv2AxOjTuwcSUOT4aidPLBzj6QGB6k0skShxRyuYasrBvQ60fR2LQck3E8aWm/O6QMAqA7yj9yn1wUgwT8nXs9r9jhZsn2KnY4PciAa9LiuDotjmDX992aCTbkcgNj8x8gJuZ4BEHA622hsvJxauteAURMxonk59+3z9TFvTHVpOeVgkyWVjfhDYpcnR7HMVbjvk+UkNgT0y5vl4yAXzeSUSAxImi1qWRmXkdKyiXI5TpkskPvq6jRpPRLAVQqrXsUHgLwBkM8XNXUWwo5BDxa3cxMkwr59usJBsNbosGgg63b/sT06R+i142iq2sDrW1fkZi7FFGmx9PwLLV1r5KT/TfCBUfBHxJp9vnRyWREqfb9eWnlMo6PMXN0lIEQoNtDSuFgaPX58YVElLJwbEKCWtoakJD4NXLo3YkljiiUStNIT2GPaLUpTBj/NNt33IbbXYFOl8WYvP+HVrvnDKyOQICv2+392stcHnLEyDiKUMiD19OIXjcKZzBAXdpLXFflxB4Icnniv1ior8QfsKNSWqh0eXmyppk3G9tJ1aj49+hkjrEaB1UISHMAxoAnGGJlm41/l9fT5g9wZpwFvVxGjErJ+YlWYoewpoGEhMTIIxkFEhJ7wWKZytQpb+D3d6BUWvdZV8CskDPNrOeLtsiaXuHo/cgHuCCoUKliAdgpH88fi3aVSH641oE+PZN8uT4sDlTVxOuNYa9ticvLhYUVfDI5hwkm3RC8yz2zye7i8m2VvcevNrRzSVI0L9e3YVXKuTBp6DIrJCQkRh7JKJA4JHC5dtLcvAKFwojLXYndvp2E+NOIjpmHQhnHOpuTl+vbcAVDXJIUzQyLfo8Ke0ONShWNagBlxIHQyuXckpnAZruLZl8AgMXxUUwymRHz/h/bd9zSXSZawZi8/6DTZQKwpqt/yeTXmr2cn2DHJeh5sylyGzcoQonTM6BRIIoiNttm6uvfxOtrIinpXKxRs/qpOA6Gzbb+haI+a+1ifrSJl+raOCc+6oA8ERISEocWklEgMeJ4va0Ubv0T1qgZ1Na9gtcbrpLY2bmGFOcldMX9mbM2lveG/H3a2sWr40cxP/rQ3HooMOr4ZEoOO11edHIZ2ToNJqWcUNwpGAxj8HobUavj0OlG9cZSJKj6P1hT1TI0ChUhUUacSkmDN3L7waQc+GFst29jw8bzCYXC6eNtbd8wNv8hEhJO3+/3Ej1A7EKCWkmrL8AonRrFfmg+SEhIHPpIv2iJEcflKsfh2Ipcoe81CHpoa1vFR80du+UAwDM1zfhDu7cOHYGAnfb2H6ite5WW1nABpf0hRaNijtXIFLO+9+EtkykwGLKJjp6DwZAboXUwy6wlWb3rIa8SBK5PNWNQGYhTK7krOzli82GKSUeBIVJkqIfOzrW9BkEPOysfxx+wDdh/b0wx6Rml3RVUqBDgrPgofuy0c0VKLIq9KB1KSAyGvqWTly1bZk5PTx9XWlqquu+++2Ife+yxwbnoJIYMyVMgcQiw9wfLQM8dmSAg7OO8X4ooBqmre42yPqJFcXGnkpd7B8oD1A7YE2Msibw+VmCrw4dHFBmjUzHeHNf7+vHRJj6anE2J04NFKWe8UbdHcaCBFBt/6aeVqVPz+oQsttjddAUCxKuUeEMhPpicQ/4ejBIJiV/CBx98YLzppptSP/vss9Ls7GzfLbfc0jLSczoSkYwCiRFHr8/CZBpPIOBAo0nC4+mth0JszHxOjY3i2bp2gn10tq4cxlWqy1VFecXDEW3NzR+TmnIRFsu0YbkmQI45gZw9qBarZDKmmPVMMev3OY7ZMhWZTEMotCtOITPzOhSKX7bdkq5V77N0ssSRwbLiZdanNj+V3OZuU0Vro31LJiypOzf33APWLfjss88M11xzTcaHH35YOnbsWC/ADTfckGQwGIJ33HFH0+rVq7VXXXVVutvtlqWnp3tfe+21ytjY2OD06dNzCwoKXJs3b9a1t7crXnjhhZ133XVXYnFxsfaMM85oX7p0aT3AE088YX3yySfj/X6/MHnyZOdLL71U1VNMSSISaftAYsRRqaIZm/8QanUC6elXk572B6KtxzJmzH2kpV3OZLOB9yaO5reJVs6Ms/DmhKy9FvU5UIIhN6LYX859T3UPDjVMxrFMnvQaKSkXExt7EuPHP0t09NyRnpbEYc6y4mXW+9bel97qblWJiLS6W1X3rb0vfVnxsgOqD+Dz+YRzzz139DvvvFM2adKk/hG3wKWXXpp5991315aUlGwfO3as+9Zbb03qeU2lUoXWrVtXfNlll7UsXrx49LPPPltdVFS0bdmyZTGNjY3yDRs2aN5++23runXrioqKirbLZDLxqaeekrYl9oBkKkkcEuh0GaSnXb7H16dbDEwfRkOgL1pNCkbjBOz2zb1tCoXpkFNc3Btm8wTM5gkjPQ2JXxFPbX4q2Rf0RSwkfUGf7KnNTyUfiLdAqVSKkydPdjz11FMxM2bMqNn99ba2NrndbpefeuqpDoArrriibfHixaN6Xl+0aFEnwIQJE9yjR492p6en+wFSU1O9FRUVqm+++cawdetW3YQJE8YAeDweWVxcXOCXzvfXjuQpkJDYDaXSTH7+vcTFnYZcrsdimc7ECS8cVkaBhMRQ0+ZuGzCIZU/tg0UQBJYvX16xadMm/V/+8peEfZ8RSd/yx31LGPeUQxZFUVi8eHFbUVHR9qKiou2VlZVb/+///q9+zyMe2UhGgYTEABj02eSPuZeZM1YwYfxzmM0Th+U6waAPv79rWMb+NeIPhejwhSs9ShxcorXRA5ZI3lP7/mA0GkOfffZZ6dtvvx390EMPRShiRUdHB00mU/Czzz4zAPz3v/+NnjVr1oAllAdiwYIFto8++iiqrq5OAdDU1CQvKSmRdLr3gLR9ICGxB+RyNXJ54rCNb7NtobLySRzOYhISFpGYcCZabeqwXe9wZ5vdxVM1Lfzc5eSUWDMXJUUzSqcZ6WkdMSyZsKTuvrX3pffdQlDJVaElE5YMSenk+Pj44GeffVYyd+7cvNjY2ACEvQgAL7zwws6rrroq/Y9//KMsLS3N+/rrr1cOdtwpU6Z4br/99rr58+fnhEIhlEqluHTp0uqcnJwDNmZ+jUilkyUkRgCns5y1687qLZAEkJh4Drk5dyKXS4uY3al1ezllQ2mvSiTAMVEGnhubuUcRJ4lIhqJ08nBlHwzEJZdckjp58mTX9ddf37bv3hL7g1Q6WULiEMPpLI0wCAAaGt4lI/2q/YpdcLvr6Oxcg82+FbN5KlGWaajVcfs+8TCj1OWNMAgAvu1wUOXxUqAc3voPErs4N/fc9uEyAvpy/fXXJ23YsMFw7733Snv/BxnJKJCQGAFksv55/3K5FkEY/E/S52unqOivtHd8D0Bt7UskJZ1PTvbtyOW/Lre6WtY//EkugEpSVPxV8sgjj9QDkkEwAkiBhhISfQiFAtjt22hq+oi29h/w+YZnUWQw5KHXj4loG5X5JzSaPZdl3h2Xq7zXIOihvv4NXK7K/ZqLOxhkfZeT95s6+LHTgc0f3K/zDwa5eg1zoiJTUpekxpEpiSpJSAwpkqdAQqIPbW3fsKXwKuiuthAXdyq5Of/aZ8nk/UWjSWR8wRN0dv6My12FxTIVs2lSb2DVYAiF/AO0iojiQO17GEMUebOxg1tLanvbrkuL48/p8egUh85efbRKwUN5afzU6aDY6WGyWcc0kx7VAB4ECQmJX45kFEhIdOPxNlNU/HfoU36pufljkpLOJdp61JBfT6dLQ6dL++Xn60eh0aTh8VT3tlksM9Bq0wc9RoXbyz/LIoPHH61u5tRYCxMHKMs8kqRoVKQkDK1xJiEhEYlkFEhIdBMMOvH5+ldDDPg7RmA2+0ajTmDC+Kepq3uNjo4fiYk9gaTEc1AqB1/jwBYI4gn1z0Dq8EuCbxISRyKS701Cohu1Kp6oqN09AjK02oy9ntfs9VPn8THc6b3+UIg6t4/2PlH4BkMOOTl/Z+rUd8gadeN+qy4mq1Wk7lZtUS+Xka6V0iIlDg6CIEw588wzM3uO/X4/UVFRE+bNmzd6b+d9++23uksvvVQS9hhi9mkUCIJwnyAIJkEQlIIgrBQEoVUQhAsPxuQkJA4mCoWOnOy/ExU1GwCVKpaCgscxGHIH7O8IBHizoZ3j1xVzzM9FPFDZSKN38Pv5+0Oly8tfSmqZ9dMOFm4oYWWbjUD3Cl8Q5CgUhv2KR+ghXq3k2bHpFBjC2QoZGhUvFWRKokASBw2tVhsqLi7WOhwOAeC9994zxcfH7/OHdMwxx7hefPHFfrUSJA6MwXgKThRF0QYsBGqBHODmYZ2VhMQIYTBkM77gKWbO+IJpUz8gLvZEZDLlgH3X2Vz8saiaZl8AZzDEg5VNfNTcOeRz8oVCPFTVyKsN7fhEkQq3j4sLK9jqcA3J+BNNet6aOJrvpufx0ZRsjooyDsm4fTkcRdIk+tP++hvW0jnHFOwYkz+ldM4xBe2vvzEkQR7z58/veuuttywAr7/+uvXss8/uTfv5+uuvdZMmTcobM2ZM/qRJk/I2b96sBvjoo4+MPd6EPfWR2H8GYxT03BFPAV4XRXHYhSskJEYShUKPXj8KjSZ+r/1Wtdv7tb1c34Z9iFP6Gr1+3m6KjGsIilDq9A7ZNSxKBdl6DTGqgQ2gX4rTWUZ5xUNs2Hg+NbUv4/FIqeeHK+2vv2Ftvuee9EBLiwpRJNDSomq+5570oTAMLrroovZly5ZFuVwuYceOHbpZs2b11imfMGGC5+effy7asWPH9n/+8591t9xyS8ru5w+mj8TgGEyg4YeCIBQBbuBqQRBigQFrXu+OIAgLgEcAOfCcKIr3DNDnWOBhwsZHqyiKUuF3icOCFE3/ffcMnWrIBXW0MhnxKiX1u21NmA9xeV+Pp4HNW67E7a4CoLNzLTZbIXm5d/zqxJWOBNqeeCJZ9HojFpKi1ytre+KJZOv55x3QYnHGjBnu2tpa9bPPPms9/vjjIyqEtbe3y88999zMyspKjSAIot/v7/cDG0wficGxT0+BKIp/AWYBU8VwArQTOGNf5wmCIAceB04G8oHzBUHI362PBXgCOF0UxbHA4v19AxISI8UxUUYS+6ysNTKB69LiUcv3L363yu3lg6YO/lfbyrouJ75QKOL1WLWSu7KT6XuXm2rSMc6gPZDpDztOZ2mvQdBDY+O7uN3VezhD4lAm0No6YPTpntr3lwULFnT+85//TL344osjDIxbb701ee7cufbS0tJtH374YZnP5+v3AxtMH4nBsU9PgSAISuAi4JjuQKZVwFODGHs6UCaKYkX3OG8QNia29+nzW+BdURSrAURR7J8PJiFxiJKt1/DupNEUOlz4QiJj9FrGGvfvQV3t9nLhlgpKXeGtAAF4sSCTk2LMEf3mR5v4eHI2pS4PZoWc8UYdSQN4Kg4pBgx8FJCSng5PFDExvkBLS78vnSImZkiqDV511VWtZrM5OH36dPdHH33UG9his9nkKSkpPoCnn346ZqBzB9NHYnAM5tf5JDCF8Ir+CWByd9u+SAb6RobWdrf1JQeIEgThG0EQ1guCcPGeBhME4UpBENYJgrCupaVlEJeXkBh+MnVqTo+L4pwE634bBACb7e5egwBABP5VVheRdgigksmYbNZzbmI0C2Ith75BABj0Oeh0kVllyUnnSuWhD1Oir766TlCrI9xYglodir766iEpnZyVleX/+9//3m9heOuttzb+61//Spk8eXJeMDhwvM5g+kgMjn2WThYEYbMoihP21TbAeYuBk0RR/H338UXAdFEUr+vT5zFgKjAf0AI/AqeKoliyt7Gl0skSvxZeq2/jhuLIrCq9XMZ30/MOiwf/vnA6K2hp/ZKurvXExhyP1ToHjSZhpKd1RDIUpZPbX3/D2vbEE8mB1laVIibGF3311XUHGk8gcfA50NLJQUEQskRRLAcQBGEUMBhTrBbouyRIoX/Vq1rCwYVOwCkIwrfABGCvRoGExK+FMQYNMvoKK8PFSdEkqAefBeDxNBEIdKJSxQ55jYbBYA8EqfP40MplpGlUEXoJev0o9PorD/qcJIYH6/nntUtGwK+bwWwf3AR83e3iXwV8Bdw4iPPWAtmCIGQKgqACzgOW79bnA2COIAgKQRB0wAxgx+CnLyFxeFNg0PHq+FHkGzREKeRclxbH5SmxyAYhRCSKIm1t37Nu3SJ++vkU1q1fTFfXxoMw612UOT1cvnUnx64t5ri1xbxU34YzILlvJSQOV/bqKejOIJgAZAO5hKOEikRR3GeCtCiKAUEQrgVWEE5JfF4UxW2CICzpfv0pURR3CILwGbCF8GLpOVEUtx7QO5KQOIxQyATmRZuYbNLhDoWIUykHZRAAuFw72VK4hFDIDYDbXUlh4TVMm/YeavXeNRb2F1cwiC8kYlHuumV4gyEeqWri2w4HAM5giFtLasnRaZi1W5ljCQmJw4O9GgWiKAYFQThdFMWHCD+49wtRFD8BPtmt7andju8H7t/fsSUkfk2YlQrM++4WgdtT22sQ9OD1NeHx1A+ZURAIifzY6eDBykaafH4uT47l9DgLcWolrf4AH7d09Tun3O2RjAIJicOUwcQUrO4OCFxGWKMAAFEUNwzbrCQkJPaJSmkl7LzbFSwsk2lQKCxDdo3NdhfnbSkn2H2J28vqCIoif0iLwyiXk6fXsMEeKbccN8SqiBISEgePwcQUzAbGAncAD3b/98BwTkpCQmLf6PWjGTXqzxFtuTn/QqdLH7JrbLC5eg2CHp6pbaHV58eklPPP0UnoZLtuI/OjjYw36obs+hISEgeXfXoKRFGcdzAmIiEhsX/I5RpSUy7DGjUbr7cRjSYVvT4HQRg6cSCDov9YZoUcZXfcwwyLgc+mZlPm8mKUyxljGPr6CRK/fm699daEd955J1omk4kymYwnnnii6vvvv9f/+c9/bjUajaG9nXvuueem33LLLU1TpkwZlPz+SFBcXKxauHBhdmlp6ba99UtOTi7Q6/VBQRCIiYnxv/766zvT0tICeztnqBmMomE8cDeQJIriyd1SxbNEUfzvsM9OQuIIpMrtZbvDTUiEfIOWTN2eC74pFDrM5knDNpepJj0xSgWt/l33pdtGJWLuE3CYo9eSoz94ksutPj/bHG7a/UGytGrGGLQoh7jehMTB48svv9SvWLHCUlhYuF2r1YoNDQ0Kr9crXHTRRaOuuOKK9n0ZBcuWLava2+uDJRAIoFAMZkd9//D796+c+qpVq0oSExMD1157bfI//vGPxL7loUOhEKIoIpcPX92TwSwpXiScQZDUfVwC/GmY5iMhcURT7HRz5oYyLttayeXbKlm4oYRtdve+TxwmwlLOWdybk8KtmQm8P2k0xwxDaeXB0uoLcFtJLeduruCq7VUsWF/Cl239gx0lhofCVbXWF279vuDxJV9NeeHW7wsKV9UesDBGXV2d0mq1BrRarQiQmJgYeOWVV6Kam5uVc+fOzZkxY0YOwLvvvmuaOHFiXn5+/piTTz55VFdXlwxg+vTpud9++60OQKfTTbruuuuSc3Nz8ydMmJBXU1OjANi2bZt6woQJeePGjRvzpz/9KUmn002CcPnlGTNm5Jx22mmZubm5Y10ul3DOOedk5OTk5I8ZMyb/ww8/NAKMHz8+b926db1VvKZPn5773Xff6Ww2m2zx4sUZ48aNGzNmzJj8V155xQKwdOnS6JNPPnnUcccdN3rOnDk5fd/v0qVLo0888cSsOXPmZKenp49bsmTJgBUdjz32WPvOnTvVxcXFqlGjRo298MIL08aOHZtfXl6u+vvf/x4/bty4MTk5Ofl//vOfkyDsjcjMzBx71llnZeTk5OQvWLBglN1u32+34WBOiBFF8U269VVEUQwwOPEiCQmJ/eSTli4afLtWFm3+IMsa20ZwRmFPwCXJMfw5I4GZFsOgCj6JokiFy8uaTgeVrqEr8bzd4ebDPhkPIeDWkloaPXuW32/0+vi508EOh7tfsSmJwVO4qtb6w1tl6a4unwrA1eVT/fBWWfqBGgZnnnmmrb6+XpWRkTHuwgsvTPv4448Nt99+e3NcXJx/1apVJT/99FNJQ0OD4u6770789ttvS7Zv375j8uTJrjvvvLNfio3b7ZbNmjXLUVxcvH3WrFmORx99NBbg2muvTb366qubt27duiMpKSli6b5lyxb9/fffX1deXr7t3nvvjQMoKSnZ/tprr1VceeWVGS6XSzj77LPbX331VStAVVWVsrm5WTlnzhzXX//618R58+bZtm7duuO7774rvv3221NsNpsMYMOGDYbXX39955o1a/qJ8W3fvl33/vvvV+zYsWPb8uXLo8rKyvrtuS1fvtySn5/vBqisrNRcdtllbTt27Ni+detWTVlZmWbLli07duzYsX3Tpk26Tz/91NDTb8mSJS0lJSXbjUZj6P7774/d37/HYIwCpyAI0XSHOAuCMBOQTHMJiWFgq6O/V2CL3U1oH3LkhxIhUeTTli6OX1fMmRvLOH5dMStau9iXpPpgaPf3315t9gVwBAd+2G+xuTh1fSmnd8/jiepm7JK40i9i3SeVycFAKOKZEQyEZOs+qdy9ps1+YTabQ1u3bt3+2GOPVcXGxgYuueSSrKVLl0b37fPNN9/oy8vLNdOnT8/Ly8vLf+ONN6Krq6v76YArlUrxvPPO6wKYMmWKs6qqSgWwceNGw+9+97t2gN///vcRVvb48eOdeXl5PoDVq1cbLr744jaASZMmeZKSknyFhYWaiy++uGP58uVRAC+99FLUaaed1tE9L9NDDz2UmJeXl3/00Ufner1eoaysTAUwZ84cW3x8/IBftqOPPtoWHR0d1Ol04ujRoz3l5eW9e4Rz587NycvLy7fb7bI777yzESAxMdE3f/58J8Bnn31m+vbbb035+fn53Z4DTVFRkQYgISHBd+KJJzoBLrroorbVq1fvd27wYDZQbiSsRJglCMIPQCxSiWMJiWHh9FhLv9z/xQnWQQsaHQrsdHu5ZkcV7lDYCHAEQ1y9vYovp+buNT5iMGTp1P1koedGGUkcQBba5g/yt7Ja6rzhhWFQhHt2NjLdrGf2CG6BHK70eAgG274/KBQKFi5caF+4cKF9/Pjx7pdffjnCKBBFkaOPPtr24Ycf7tzHOKKsOxtGoVAQCAT2+cPR6XS9X6c9Ga6ZmZl+i8US+Omnn7Tvvvuu9emnn67q6f/222+XTZgwIcId9v333+v7jrs7KpWq90JyuVz0+/298+yJKeg5bmtrk+8+xz/96U8NN998c0TNiuLiYpWw231i9+PBsE9PgSiK64G5hFMT/wCMFUVx835fSUJCYp8cFWXg1swEtDIBtUzgurQ4jo8+vB5gDV5/r0HQgzMYotG3fwFXAzFGr+WFgkwSujMc5kYZuTM7Gb2if+BVRyDA2i5Xv/bavWw1SOwZnVk14Ae3p/bBsnnzZnVhYWGvtbhx40ZtSkqKT6/XB3viBo499ljnunXrDFu3blUD2O122ZYtWwZtYU6cONHx4osvRgE8//zze9zuOProox2vvPKKFWDLli3qhoYG1fjx4z0A55xzTvvdd9+dYLfb5dOnT3cDzJs3z/bggw/Gh7q3pX744Ydhj7g9+eSTbS+//HJMz2ezc+dOZV1dnQKgoaFB9eWXX+oBXnvtNevs2bMd+zv+Po0CQRDKgd+LorhNFMWtoij6BUH4aH8vJCEhsW9iVEquT49n1fQ8vp2Wx19GJRKvPsjVEkMH5l6PVylQ75YNoJUJxKkOPLJbIRM4KcbMiqk5rJ6Rx/MFGeToNQP2tSjkTBignPWvofrkSDD1lIw6uUIWsfqVK2ShqadkHFDpZJvNJr/44oszs7Kyxubk5OQXFRVp77333vpLLrmk9eSTT86eMWNGTlJSUuDpp5+uPO+880bl5OTkT5kyJa+wsHDgP/wAPProozWPPvpofEFBwZiGhgalwWAY8Et+yy23NAeDQSEnJyf/3HPPzXr66acrewIgL7zwwo6PP/7YesYZZ/QWhLrnnnvqA4GAkJeXl5+dnT329ttvP6CtlMFw1lln2RYvXtw+bdq0vJycnPxFixZldXZ2ygFGjRrlef7556NzcnLyOzo6FDfddFPL/o4/mNLJRcBmwAX8QRRFnyAIG0VRHL48qH0glU6WkBgGWstg6ztQ/iXkngr5Z4I1Y7+HCYoiHzR18KeiGnyiiFomsDQvjdPiLAd9G2R9l5OLCyto84efAdemxnFtelxEDYcjhaEonVy4qta67pPKZFeXT6Uzq3xTT8moK5ibcshXTbTb7TK9Xh+SyWQ888wzUcuWLbOuXLmyfKTnNZQMVgsBDrx0sksUxXMFQbgF+E4QhN/QV1dVQkLi8MfRDO9cDg2bwsc1P0PFKlj8Imj3ryqDXBA4Iz6KsUYtTd4ACWoFWTrNiMRFTDHrWTElh0qPD7NCzmitGu0AWw0Sg6Ngbkr74WAE7M4PP/ygu/7669NEUcRkMgVffPHFypGe06HKYIwCAUAUxfsEQVhPWLPg4Bdtl5AYCZwtULcB2sohejQkTwZ9zPBfVxShYTPUbwK5EpKnQFze8F2vtXSXQdBDxVfQXh5+z4PE7+/CZivE6SzFqk0hwzQBtTqu9/WuZhdNlTa8rgCxaUZi04zIB1BNHEpStGpStAcW4ChxeLNgwQJHcXHx9pGex3CSm5vrG4yXYF8Mxij4R88/RFFcKQjCScAlPW2CIIwVRfGAJyIhccjhdcLX98C653a1Tf8DHP8vUA2zvn/Nz/C/hRDsjuHSRsElH0HCuOG5nmwPD+b9kEwOhfzU1L7Mzp0P9bbFxp7MmLz/oFRa6Gp2sfyRTdjautVoBVh4zXjSxx0EI0tCQmJQDCb74MPdjqtEUbyjT9PLQz4rCYn9QRShpRiKP4PadeC1D824baWRBgHAz0+H2/dBV4uLqq1t1Jd24nHuZ9R9MABrnthlEAC4O6Bkxf6Nsz9E50D6UZFtY04Ha9agh3C7q6msfCyiraXlU5zO8OfVVGXfZRAAiLD63XI8rl2fj6+6Bse33+Jat45AZ+d+vw0JCYkDYyiibQ6fBGqJXyflX8Mb50Og+4Ez+48w52bQmg5sXJ9zD+3909z60lxlY/nSTXid4VTjURNjOeb8HPTmQbqwQwHorO7fbq8f3Pm/BH00nPEElH4Old/B6PmQNR80g0+HDAbdiGJ/AygYDAsyeV39X3PbfQT94YB2d2Eh1b+/glBXWKfBeOKJxN9+O8q4/RZlk5CQ+IUMxWaeFHQoMTTYGqBxazjobV/4nNC8A5q2w/JrdxkEAKuXQvPW/ueEQtBWAc3bwTuI9N3oLLDsVoY4KgOsmXs8xe8NsuaD8l6DAKBiUwtNlTYAnF1eWmvtODv3Iv2r1MC0K/q3jz4+/B52w9EZHtNlO0A5YWsGjDkN5t4KOaeAJRU6qqBpG7j3LWKq1aZhMk6MaFMoLGi14c8rNtXI7rGG4+amoDOpCLpcND/8SNggkMsxLliAOisLX8WhFSBu8wfZ4XBT5R466WYJiUOJIy8vR+LQQxSh/Cv44GqwN4YfvGc9A6kzBu7fVgFf/AOKPoRjbgHbAGnS9qbIY48NNrwEX/8H/G4YfQIsuAdiRu95XsYEOO81+OYeqPoeMubA3L+E2/eA1x2gubL/9oW91U19WQdfPr8De7sHQ5Sa+ZeOISV3DzG7OSfCKffD9w+BQgMTfwurHw8HBE65BDRmRFGkZkc7K/+3A1eXD1OslhMuyydh1P5lC/RS8Q28twTsDWDJgAV3wyc3ga0eUqbD6UshbsweT1cqTYzJv5eqqqdpbV2JyTSBrFE3oNOlAhCbbuTUaybw43tluGw+xs1NIf+oRARBIGS349myBWQy4m68AdtHH2P/7DNkr7xC/O1/w7RgATL1yAYLljg93FJcw5ouJ3q5jH+PTuKsuCh0UjbDASGXy6dkZ2e7g8GgMHr0aPebb75Zua/KiPvL2WefnbFw4cKuyy67rGMox/01MhSeAkke7AjB6XTi8w3dn1sMibjtPgJtNbDswrBBANBRCcsuCj+MdicUgg3/CxsEEI6Oj8vv3y8qI/K4bj18/rewQQBQ9gX89GR4/74vrvbI7YGEcXD2s3DVj3DWs5Awdq/vSWtQkjG+f+CcKVbLJ08WYm8PezQcHV4+fbKQrpY9bEXoYwhMuoSO05fiH3cO/PAIVH0HX/wdatcD0Nnk4tMnC3F1hf8mthY3nz5diKPjF1RVbCuDNy4IGwQAnZXw4fWQf0b4uPZn+OKf4HMRDAXp8HTgC/b/Lhj0oxmTdxczpn/K+ILHMZkKel+Ty2Wkj4tm0Y2TOff26Uw7NQNDVFh/Rm61YjjuOAxz5mD/ciWe7eFA8ZDdTsOtf8FTVDSot+H32/EHbPv//veBOxjk/1XUs6YrvKXkDIa4qbiWwgFqVUjsH2q1OlRUVLS9tLR0m1KpFB988MGI/aJAoH+9i8OJw23+g1E0XLm3NlEUZw71pCQOLTo7O/nmm2945plneOONN6iuHmC/ez/panax+r1y3rx7LSveaKVpzv9A02eF62iCrtr+J7o7YNu7u453LIfJF0FUt0tfpYczn+y/om0aYDth+/vg6tZtsdXD94/AM8fCK2fDzu92KfsptWBKDLv194FcIWPySenEjwrHM8jkAjNOz0RrUEZsKQD4PEHsbQO7oSs6K7h77T0s3ngvtwfrKD7+b6DoXik3bgHA3uYh4I9cULm6fDhKCqF9rxLx/emsAd9uWyrOFlD1qadSuoLKrp3cv+5+Fn+4mL989xd2tO3oN5RMpkKjiUcuHzhDQ61TojerI3TZZUolMVdege6o2bg3bOh3jn8f37lAwElT0yes3/Ab1q07h4bG9/H7h844aPUF+Lyt/3g7j7BthE1ffGJ96g8XFTx47sIpT/3hooJNX3wypOnpRx99tKOsrEy9e0njQCDAH/7wh5SecsH3339/DIRLH0+bNi33lFNOGZWRkTHu6quvTn7yySetBQUFY3JycvK3bdvWz710/fXXJ5199tkZ27dvV2VnZ/da+f/4xz/ib7jhhiQIl0a+/PLLU6dOnZo7atSosatWrdKdeOKJWenp6eP++Mc/JvWcc/zxx2eNHTt2zOjRo8c+8MADvasBnU436U9/+lPS+PHj81auXLnfRYlGkj1uHwiCoAF0QIwgCFHsCig0AUl7Ok/i10UwGOSHH35g7dq1AHR1dVFdXc0VV1xBfHy/yqWDwu8Lsvq9cio2hhU4HR1e6ss1LD75n1hW3xDupNSG0/B2R22AxIm7AvGCfvjyX3De66Czgtoc3vPfffPanBb+vz4G1CZor4D48aA2hrcvNrwEq+4JGxetxfDymfD7LyFp/4U7rYl6Fl47AXurB4VKhjlWS2eTC5lcIBTcFYIjyAS0xv6FfDo9nfz1+7+yrS2c6btWDFHn7eKBGVeQsOMTiMkGQGNUhn+VfaJ6FEoZmo5NsHJ12DhSDlKKXW0Ipx8qNGBK6jbIRIgbC5nHQtX3OKZeyn82/B8/Nf5Mgj6BLS1bWPLlEl475TWSjQeu7qrOykLQaul4+RX8NTURr8mjo/dwVpjOzp/Zuu263uPt22+koOBJ4mJPPOB5ARgVcnJ0GnY4PRHtsUeQMuKmLz6xfvO/Z9ODfr8MwNnZofrmf8+mA0w84ZQDFjTy+/2sWLHCdOKJJ9ogXNJ448aN2/Ly8nwPPPBAjNlsDm7dunWH2+0Wpk2blnfaaafZAIqKirRvv/12RVxcXCA9Pb1ArVa3FhYW7rjzzjvjHnzwwbjnn3++98u0ZMmSFJvNJn/rrbcqS0tL96p3rVKpQuvWrSu+88474xYvXjx67dq1O+Li4gIZGRkFf/3rX5sSEhKCr776amV8fHzQ4XAIkyZNyr/wwgs7EhISgm63WzZu3Dj3ww8/PIzRwcPD3r7RfwD+RNgAWM8uo8AGPD6805I4VLDZbKxfvz6iLRAI0Nzc/IuMAq/LT3OVvdcg6MHnDtARTMMC4Qf6KQ8MnA6nUMPRfwpHyLu7twcTJ4QflJa0PV84ZRqc9nB30FwnTLksHCOg0ocDHOvWwfx/hj0KumjQx4YDGffDKBADIbzVdjzb2xDUcsxjrKgS9ABY4nXM+U02q17fVVr9qLNHY4nrv5qusdewrW0b0epo7sz6O8kNFgQvKEdZEUMKhNYyaCnGmjCaGaeP4qcPKsInCnDMyRrMOx4FWw0c+1eIzSEYDNG800bF5lbkCoHMCTHEpZsiK6gFg3D6o2Fjqa0cCs6lMfEiakpcdIXuInW+AW1aiPqfr+ax2X9C6y0FZDhUWdTaa4bEKABQJSWR8O9/UfuHJYj+cLaCadGZaMbsOZYBoLHx/X5tdXWvERtzwi+qFLc7FqWCu7OT+e2Wit5iT6fFmikwDrNexSHEmrdfT+4xCHoI+v2yNW+/nnwgRoHX65Xl5eXlA8yYMcN+/fXXt3755ZeGviWNv/zyS1NRUZGup3yx3W6Xb9++XaNSqcSCggJnenq6HyAtLc178skndwFMmDDBvWrVqt70mXvuuSdx8uTJztdff71qMPNatGhRZ884o0ePdvdcIzU11VtRUaFKSEhw33vvvfEff/yxBaCxsVG5bds2TUJCglMul3PppZcelvELezQKRFF8BHhEEITrRFF89CDOSeIQQi6Xo1arcbsj906Vyv4r3MFQuq4Je5sXhVLWz/UtT8yFc18Or+pjx+xZUMfVDlN/BzI5CPJwtoJnH9Hx3i744l/g6QwfF74Fi/8HyZNArobUmWGPQw+6aFj0zH69N09FJ20vbOtduTu+rSX2D+NRpRiRyWXkzk4kLsOEo92DPkqDNVGHXNn/PaoVagQE/pP1dzI+0EAgHHfgXufCe+oJaFaeBj89heLSj5lwXBopaSLOylJMGjtRRf9A6KoGpS6shAg0lnfxwUMb6SlzsvGLahbdOJmEzD7bNQo1bHo9HFAJtGRdx2cvVPRmSRT/1MzMM0dx55TL6Nr5d+xieGtFEJSkjHt2vz6nfaGfOZPMd9/BW1WF3GxGk5OD3Lz34EmVqn/aolodPyQGQQ+zooysmJpDucuLWSEnT6/Bqvplv4PDEWdnx4Ar6z21D5aemILd23crFyw8+OCD1WeffXbEHs5HH31kVKvVvb4ymUyGRqMRe/4dDAZ7vwATJ050btmyRdfU1CSPj48PKhQKMdQnm8fj8UT8GPuOs/s1AoGA8NFHHxlXrVplXLduXZHRaAxNnz491+12yyDsZVAoDk8v0j5nLYrio4IgzAYy+vYXRfGlYZyXxEjRXhFOCwSIH4spOosTTjiB5cuX93aJiYkhIWHPEfh7wtHpZc0HFWj0SsYek8zmlbtcxAmjTERnxoP59L0P4rXDV3dBw8bI9pgcSCgY+ByA2rW7DIIevr4LMueCGIT1L0S+5moLxzUMEjEQxP5NbYQrX/SH8JR0oEoJL1aUSjlx6Sbi0veun5BmTGPJ+CUkVVt6DYLwgGAv0qM+/SmEjhJo2IQy00RCmhZW3xv2dvQw728QlYGt3cOmL2voW/csFBAp39ASaRT4Xb0GAUCLMxpnZ2RcQvW2drKSf0AUdxWYE0U/9vaVEDtn3x/SIBFkMtTZ2aizswd9TnzC6dTVv04oFHbvy2QqkpLOG7I59ZCj15KjH/bquIckekuUbyADQG+JGvZg8xNOOKHrySefjF24cKFdrVaLW7ZsUWdkZOyXKtiCBQtsJ510ku2kk07K/vrrr0tSUlIC7e3tisbGRrnZbA6tWLHCPH/+/EEHonR2dsrNZnPQaDSGNm7cqNm8ebN+/9/Zocc+jQJBEF4GsoBNQM/dQAQko+DXRtN2ePmMXToBumi4eDljx47FYrFQU1OD2WwmLS0Ni8Wy38OLIZGgP0RXs5v2BiczzhiFo8NLTLKetHHRgxP3EUPgH0BUKLCPgK+BBIf87rBBIIZ2ZSVEXmzf8+npKYLo61+NNTRA277QKDScP+Z8AtVNeImct+gXoPBtqPg83DD+vHBq5TkvQM2acIBhyhRImUZ7o4s1y8vxefpHP/vde7+Pi2J/D4aASCjY/7MPDtB2sDGbxjN1ypt0dq5DFINYLNMwGodJEvoIZeY559f1jSkAkCuVoZnnnH9ApZMHw5///OfWyspKdUFBwRhRFAWr1er/5JNP9lvE4ne/+12HzWaTLViwYPTKlStLb7zxxobp06ePSUlJ8Y4ePdqz7xF2cfbZZ3c988wzsTk5OflZWVmeCRMmjPwPYQgYTOnkHUC+uK+OB5FfW+nkgN1HsM2NoJKjiNEiU41Q3vNX/4Fv749sm/1HOPHOIRne1eWlens7tjYPtUUdNFV2MX1+Kjn5VtRGFQpVO6K7hc5gCl6/GoNVjSl6gFXZxlfDmgY9yOTwu88hZWr/vj2UrIBlF4QDE3uYe2v4/akNBDa8QXtzM74AWPy1GKo+g8UvhEWRdDGEorLobPHisXvRa9yYhTowJYN51166a3ML7a+HU+c0Y6NRJuhRpRhQJOlRmndlLoRCIdra2nA5HRiVItZQazi4z5wSMWVvRSctzxZG2CbWk1Xovjst0gi69BPI2E2iGNiwooqfllcw/bRRrHk/8v55xu8TSZnaZ5/e3oT4zh/oTFyIOxSFmHkMHz+5Hb93l1EzYX4yucfUsmXblRFjTZr4Mlbr7D1/9hKHBENROnnTF59Y17z9erKzs0Olt0T5Zp5zft1QBBlKHFwOtHTyViABaBjKSUmE8TU6aX9lO4FWDwhgmJWEcX4qcv0BbdP9MhoL+7c1bB6SodvqHax4dhsdDWFjOmd6PHMXFOD7ZCddPzWE3/s0I53xej59bTuhoIjGoOTkJQUkjbZEDpZ7Cpz1XFhnQB8Ps6/dd0BgVz2ccEdYJMnZBtnHh9X6fE7cvhBrtztZ+9EXIIqYYuM585rniX35LPB0EkiYQknu06x6p5ZQQEStU7DgTIGUTaeFtQtSpgCgzonCekEegXYPnqJ27CvDGRKqTBOW00ahSjISCATYunUrH330EYFAAI1Gw+K5Y8lad1U4WyB1eu+UVWkmYi4fh/3bOkRvAMN0M5otN/f3irgHvie31TsIBUWqCluZeWYWlYWtyOQCk2cIJLS9Cfyzt29IH0f5mEf56pVyAv4Q1qQKFlyRz/bVjXS1uBk9MZo4dznGUC7jxz9DdfV/EQQF6Wm/x2wefBVFicObiSec0i4ZAb9uBmMUxADbBUH4Gei9G4miuI/NX4l9IQZC2L+uDhsEACI4VtejzolCmze81amDTj+iN4jMoOz1TDin/ZVQ1FHoy1+HqFxC8bMQ0sayv34LMRAiaPchKGTIjSqCgRCbvqjuNQgA3F0+gj83Emzv895/tiPMj+pN2/M4/Kz83w7OvmUyOmOfrQVdFIxfHJbkFeSgGESwV2w2vHhqOAshehQ2n4xQ9iJMuhiatm5m7YfvEZs5mrQp0/HZbfz00cecFD8JZdXXdIy+iq/frOldsXtdAb74WMXiU/6I4YdHYN5tIII8JhtdQSy2r6vx7dy1NenbacOzowNVkpHW1laWL19OT4CTx+Ph3e+L+MOJdyDYbQSbajCq5chN8QgKOZrRUaiTFPDNvQi1XeCLzNpArtqVpeFsC2sN6ONApWXUhFhKfmqiobyLpiobSdkWxk1SkFb0F/zT/4q/pRVlbDi1urPJxZcvlxEKhN9ke72TFf/dwcITRURfCc6H30IYNw7VaXOJVadjjQrHEMjlI2C8HiCBgAO/vwOFwoxSeYD1MSQkfmUMxij413BP4kgl5PLjLe3s1x5odsEwGQWiKOKt6KLz/TICrW40eVaMJ6ZTXmHjx3dthMSJLLroHEJrGvEVOlGm6LBobKjTBnfz9Le5sX9VjWtjM3KjGssZWYhJeqq3RS4uklIN+Df3r3Gg8gcRBHoD42wtbtw2f6RR0MMgxIR2XXASnPEE7uIv2Wo5ja/WFRNYt55ZzQpSjBomn3UebaKCb3eUYDAYmDnlaNxCHMqqr3H4jP3CC1xdPlyKNAwTzoMXTwnHLEy5FI76M96yzn6X91Z0wvw0bDYbod3qF5jNZsq9Uaz8YjVu949MGT+W2RlqLBkTwZyM4G2Dnx8Nb5Mc/+/w+65eE5YiXvgQxOZCxSr4+EZoL4PchTD/7yTljGLWWVms+7iSUFAkLsZHjKKBpsqJdDx8FYooC/G33YZh3jycHZ5eg6AHnzuAs82H//GHMMyfT+wfr+uVGj4cjQEAm30bpaX/obPzZ4zGieTm/B2zeeJIT0tC4pBhMKWTVw3038GY3K8dQatANYBOvSJ2+KKbA80uWl/YSqDFDSJ4drTT+UE5tVta8Tj9FEyLx/95Lb6q8KreX+uk7cVtBDo8BIPd+8vBgYN+w56PGlzrmyEEwS4vbS9vR9blIznPEtG3sc6BMqO/oeFXyCMi5Y3Rml0CP8EABPYScLy7ZHFfVHqC48+nZuJNfPzdJtxuN36/n2+/+x6noKBTpmJrURGhUAibzcbn335Hhz4DAIPK0a8WqNaoRNv+M6x9LiyzHPRB/QaoWYMqo//fVNX9Xo1GY780ubFjx/LBJ5/jcDgIBoP8vHEL62rciCWfhTtooiB5alhh8Yt/hAWXjrkpLL+cOSdcNvrjGxGNKWHBp6IP4eOb0Co8TDoumfP+MZ3f/nMa009NxfPdFjreehf8fgSdHve2IrxVVeitGlLHRmG07jK0dBYVcfNnMerLL0i6/z5UaXvWgRCDuwwdfzC0x34jidfbTGHhEjo7fwbAbt/E5i2/x+0eQDlTQuIIZTDZB3Z2rZNUgBJwiqIo+d0OEJlSjml+Gv46B8GO8M6MblpCbwrbcOBvccNuK0J/pY2Uo5Mp3dyK1aTqnUsPIVeAlm0VfP3Z84yZNoXRugaMYjtMuADidwWrBR0+XBt3W/2LEGx1M/mkDBorbNi7t0q0Fg2mY5PoaHIRtIUj4fXTYrEZfb2eApVWwfxLxqDTBKBoJax/MRz4N/liGDUvXO4XwgWRij6BLcsgeQpMuihcs6Cblmo7276vw+Pw44vvr1tid/soKusfyNxi95IuyIgqe5JjFj3O9+83EAqJKNVyjj9VwLjmybD+wQl3Qs6CcLzC+hfRzpmJp8SAvyYsG6xMNaDND881NjaWhQsX8vHHHxMKhdDr9YT8/TMnNhVVMiM2AaO7E7QWOOU+eO3csPRw2ZfhLYOGLfDF3/GMu4mutuNxfV+JafblGDMdqLY9CdveQ9iyDNP038OoefjtCrre+wCApKXPIwbi8O200+lUU964gWZtMQnTU8hWpyK2yxiTasTz3k68sRrUsU10vPIElkVnYTj6qF7dAH+TE+eGZnwVXajGRVMZreLfq8qYkh7F4qkp5CYcOrcJt6cWjydSYM7v78DtrkarTdnDWRISRxaD0SmIeEIJgnAmMH3g3pEIgrAAeASQA8+JonjPHvpNA9YA54qi+PZgxv61oEo0EHfVBPytbmQqOYpYHTL18GUfyLT9/+SCRo7HF17dBQXC/qPdFntdHU00lBbRUFpEx9yjmSv7DvnW9+B3n4VL7AKCUo7coibYFpnZI9PIiUk2cPZNU+hsdqFQyYmK1+Hd2Ix2XAyCRo4gE/BUdpEw0cpvbhqDx6/GFKPBFKOD0s9h2W937SnsXAVn/xcKzgl7Dn54DNY/HzYYatfC1nfg8i/AmklHo5MPHt6I1xVArVOQltw/lVgmgNlkpLkl8uGsi82Ai95HoY8lP2oUSWlyXDu3YhSasKx9IKx7YIgLu++XXdgbAKiqyCX68kL8zjQQQZmoR2EJr8AVCgUTJ04kNTUVh92GOdRJXX3/GN4oswmV3hSWHYawsfP7L6FsZVjUqXYjfHIjvoJrqLntXgJ14awwb3ExnrmzSJyyAJmrNaw9UPU9nPkUsqyFKJKTMS36Dd4yLb7qemR5Jr7e8h2llWGjqLa2ltiYchbPPg3nm906BdV2/DoF+pnHUX/jjST+504s55xDoNNL64vbeo1IX40da14UUSo5z32/k48K63nrD7NJtR4aqn8KuR5BkEfoLAAoFMNnhEtIHG7sd5VEURTfB47bVz9BEOSE5ZBPBvKB8wVB6FfOrrvfvcCK/Z3LSCKKIk0VZWz6/BMKv1pBa82glDN78TU6cK5tpOuLKrzVdpSxurDy3TAaBADKBD3agkgdedMpmZR3xzZsWd+MYnpixOuyCQYK133Re7zpu9V0pS8EWy207KpeJ9crsZyWFeFqV6YZQBBw/NyA0u0nOSeK+AwT8kAI+1fVOFbXY/+qBtuX1fjKuvC3yLCIHVg7WxAKG/CUtRGq2wG7Z8SuewH8bvyN7TgVp+OYshzvca8hRmWFV9PNYYG01loHXld4W8HrCmBSxmM07noIqFQqUq1JnHD0PGR9FBQTExNJSsuAUXMhPh+5p5VooYzUhqewrLk1fA1BgJlXh936M66C4/8F5lSC0fm01DdSUuekqs2Go3YbbHuXQG0ZjlWr6Pjf/9Bu20aaCqLdFaRYVMTG7KqsKJfLmT97EuqNL0Ooz3ZJwNtdyrgBNv4PAJ83mkBdHZoJk0m4+0ni/vYsumMvwTfln0T8Ib57ALkiSNzNN6E/agG+6nB5Z0+qvNcg6KGltYUuwYkyZVcdl5ArgDwqPfyZPvU0gY4O/M3Ofl4lWVEH/5qayXUzMmjs8lLa3L+M9B4J+KD6J1jzFGx+Iyy3PASIIRFvtY1QkZb0pGsjXktNvRydbgA5bYmDxuWXX556xx13xPUcH3300dnnnntues/xFVdckfKvf/3rlxVaIax6OG/evL3USJfoy2C2D87qcygDpjI4VZfpQJkoihXd47wBnAHsLmd5HfAOMG0wEz5UaCgp4s07biPYXRZTrdPzm3/+P+IyRu3zXH+rm853y3pvzACmkzIwzE2OeDANB3K9Esvpo9FNTSDk8KOI0aJM0nNipoXWGjvBQAhNsgHN+BiCHV4CCj/frXiF6qItvWPIZDKEnq+AEDlfTXYUcddMxN/sQpDL8FZ20fZS+E8uqOTEXlmAKsWISLgg0O4IagVtn3jw1/dkKtRjOeUEDIbHdokqAciV+NsCtPxvJyF7dyyBzETMqUvRrDwjnJUAyHa7hlLUMStvAUGVE61BQZxfh+KdFhRqORceczY2gxeNQUdiYiLmHmldZyt8+GdInhh222fND8cQKLWw9X0YdQysXgoyBZx0NzXOcXz8ogPEMgCMViWnnQTeZc/S8eYuZUjz6QuIzy3HWvsdF0y/kYaZ8/EFReJDjcSvuAyOuYGIB7sg6/5P6NMkILdYiLrgBhyrXUBYhMlb6STmuAm7fuByFQgy1FMn4y5tBcKp6cLuwRLdhNp9aMdEE2z3EHL1xGqE/+aCUgEyGYJsgNgBARz1LhK2O/j91DTkwn58nyu+htfP3WUAWtLhovcg+sAe2t6dXbT+txBCYMiZwdipT+PRNqDRJxMdOwWF4tDwZBypHHXUUY633347CmgOBoN0dHQoHA5H7+po7dq1hvPPP79mL0NIDCGDyT44rc+/A0Al4Yf7vkgG+v4ha4EZfTsIgpAMLCLsedirUSAIwpXAlQBpewl4OhiEQkHWf/JBr0EA4HU5KV//8+CMgnpHhEEAYP+6Bk1uFKqk4a+yKTeq0OZGZjeYY7WYBwhwtLU001gbuWKbeuLxmMtfDtcn2K1EsSAXUKUYUaUYaX+3FNfPjb2vib4grk0tqFKMKIwq9Mel4PqhAd3EOMRgCEElBxl9DIIwXV93oZlxFYo1/+6+iADTr8Rb5SBk77OSDoFtkwZVwcXI4sMVUaNTDOjMKlxdPjR6Ja5OH1u+bkKplrPwhFRC39eFd0pcAVSfeUnKMmM5To539Te4oiyoc8Ygt5VC6WcQlxuOG7Dvcvc7TnuWFreMwNwniPXVoHH5+PE7WYTZbG/30ySOR7npf8Rcdx2i349MpcK2YgW+mXPR1qzC8tP9WDTPwKQLYeMr+KfcTsCVh6zWhyLRj1ynDNd8WPhwuKLhnJvguwdQyeuw/v5qXJsjAzADTR78viQUc28JexhSpoPGhBKosbow5FjwlXSirvBRkDuWwuJtvecmJyRhaJfjKK5HOz4W55oGZCYFgcawVyjmuutQmM0Igg9Fop5An1RTRUEMRVvb6KxxMn1GHDnxg/w+uzvhi39GeoQ6q6B+4wEZBSFfENvKqt7tMH+JD0rUKGePY0f7WuZdOrDT0+ttxuksRRRD6PWj0WgSB+x3pOFYU2+1raxJDtl9KplR5TPNT60zzEw6IN2C4447znHbbbelAqxfv16bm5vrbmpqUra0tMgNBkOovLxc09HRIR8zZkx+MBhkwoQJrpdeeqlKq9WKH3zwgfEvf/lL6u7tb7/9tunmm29OtVqtgYKCggHkTCX2xGBiCi77hWMPtATZ3cPwMHCrKIrBfRUuEUXxGeAZCCsa/sI5DQliMIS9rb8ImL2tZYDe/RlI+lb0BREDh17Utik2jkXXXkXFzz/Q1NjG6OxkUs0+ZNabwml+pj1X0Q529lcNDfRtS9ZgmJ1I5/KK3m+G8fg0lAk6/I27fseiJ0Ao71zwVUHADWMXQfrRhH7sX4Qs6BQRj7oJLGGlQYvezemLBXZWaHH7dAT9YUNOoZIhePv/HUI2H63PvU3Xe2+Fz//NmcRddGZYq2HNkzDnRnC1Qlc9HROX8N5326muDxs+BoOB80+egdvRP3DQ5/YTfdJJtD7+OIRCIAhEX/F7Qso++9leG6jNeI95kdbPNYieILANbUE05plBFMsWga/7ATxuMZzzPModH2GcM5OW//Wv0xDqaoXv7wvLVeef2dueGpNO5wmtGLJNBHY6mDNmBhkZGZRXVJAcnUCqGIO4qh1RJqBM0mGcG4Nca6drRQkpTz+FbmpYFE9uUhF9wRhc29pw7+wkGK+nrMVNQ7dGQ4pRQ6JlkJk0AS84+6eo7rPQ1b4IioRs/TNWBJ+Ao3PgInYuVyWFhdfhcIY9XBpNKhPGP4vBMPhaDL9GHGvqrZ0f7UwnEJIBhOw+VedHO9MBDsQwyMjI8CsUCrG0tFS1atUq/cyZM511dXXKr776yhAVFRXIyMjwXnPNNRmff/558fjx472LFi3KuP/++2Nvuummlj/84Q+ZA7Vfe+21GV988UXx2LFjvQsXLtz3Sk2il8FsH6QAjwJHEb51fw9cL4rivvJ4aoHUPscpwO61pacCb3QbBDHAKYIgBLrjFg5Z5EolRy++GFtFA36/mw3ff0xnUwPZ0wcn9aqM1yGoZIi+XUaAOseCImZkCq2IIZFAmwcxGERh0SDT9PlaONuI/vwPRLvawJgIW0rCrvOL3oeYvW/T6acl4i3pjGyb3Lt1iEqQ0/ZJZeSqemU1puPT8TfuitHQ5Ecjyoxw+iMA+EN+au21RKX0/7z0MxLo8JvRtHswWDXQuIXoj84k2hCHmHkcdTl/Rq3T4ehS4o9S9wuq0UyIwlY1FmXWRlQnn4Rv0jQ6hUSiznoVWfVX8ONjoNAg5p3GTpui1yAAcDgcrC2u5cSLj+ajp7cT7DbyBAFi4hS03fJU2CAAEEXann8BwwtPhuWW2yvA7yJkyaNrtRHRs8t4che2oUtTojjmlrCssijCT90VHBe/gCoQQjeVCK8McgGF0U/rzH8hhIJYNr+JPHo0CDKUnVXEqgTI1yJOiifg0WHo1JOyTkGgzIPYrZCoHm9B0/kciq0vgVyFYfwkmFYAul3BmsoYLfrZiWxr81D4SVVv5UuFUkZ86n4E8BniEOf9k0CXiCjIUVS+iaz2e0gcP/gxdiPkCRDo8mA+OYOuTysJtO6qb+E1esmddfSA57W2fdNrEAB4PDU0NL5D9ui//OK5DJau5ia8LifG6Bi0xkMncwPAtrImuccg6CUQktlW1iQfqLdgypQpjq+//lr/448/Gm6++eam6upq1Q8//KA3m83BhIQEn1qtDo0fP94LcOmll7Y9/vjjcSeccII9JSXFu3v78ccfb09JSfEWFBR4AS644IK25557rn8ZTYkBGcz2wQvAa8Di7uMLu9tO2Md5a4FsQRAygTrgPOC3fTuIopjZ829BEF4EPjrUDQIAf7MLzU8isgodgtLAwhOvx5sSIiFvcCsJdaqJ6Ivysa2sJtDiRjPGin5WYthFfJAJuvw4f2rAtrIaAiLq3Cgsp41CGdO9z+q1QWtJ+N+utl0nDqKCoHq0mahzsrF9VQMCmOanocq09L4uBGWIu5VPRgS5WYUqzUig3YMmNwqZXkWwwwNpJjrcHbyy4xWe3/Y82cbR3HfOv9F87ybkCqCbmciOZjfrXi9Fa1Ry3MVjSJfZEAB3ximUWU7n24f/jcdhJyoxhfS864g7JxvPd3WI7gDyiXFsrnZRti2Rufc9xHtffkzHF5+i0+lYOPcksmKORX1cHqH2CmqzL6J6e0W/91xbX88sWRu/uTKfj5aVodYIzJoTxNywii7/bivWQABx5zrYfm9Y6+DY2wgRi7+x/8o2SBxsXw7160FlgLm3gCwsICQoZJiOTUWmVeBa34Q8SoVxtp5vV3/Alm+/RSaXM3X+sUypKUT388PhbA65itC8O3B6j8b2jQ1Bq8B0fBrOtY0EO7woxpqQFyhQvPp/fZSk6sBzF+git54USjlj5yQRDIQoXdtEVIKemWeMIno/tsICdh+Ohtk41jSCCLpxd2C6WIci8Zct8vwtLjqXl4fFwRQyjMckE2h142t0wCQtHp2DtHETBjzX1rWxX1tn58+EQn5ksuH5jQYDfkp/+pEv//s4XqeT6NR0Tr7mBuIzD50gyJDdN6Bi1Z7a94dZs2Y5Vq9ebSgqKtJOmzbNPWrUKN/DDz8cbzAYghMnTnR9/fXX/SykvZXjGcqS2Ucag4kCihVF8QVRFAPd/70I7NPqEkUxAFxLOKtgB/CmKIrbBEFYIgjCkgOa9TAgBoIEHb4IEZaB+4Wwr6rBVxF2kYr+EP7vO4gzpKHSDH6lr8mOIvqiMcQumYDljNGok0cmLcpXY8e2oqpXu8Bb3IHzxwbEUPg4KI/Fc8Zb+HJ/A/I+N0Rr5kDDRSDXKtFPTSDumglYf1+AIs+KJ+TF6w271+VmFTJ95E1WUMoIeQLItAq0+dH4qu04vq1FbtXitXtxtNup6azhwrTz0Sq0LC66hO2ntGO5PJ/P1zSxbmXYgeW2+/ns6a10aCaBIMOWv4TWsnJU2rCx09FQyw9vPYZslAbLhbl4j0vh42/q2PpzExlTLXz0xUd0dITdyy6Xi3c+X06rMpdg804cBddxxfJW0hNj2J281GyCxTZ8n1dx1uVJnH50IWmrz0Xpq0FmiryvyfQ6lMqwlgHN2xG/e5AyWTSKnAEErZStYYMAwlLGX/wDUbYrU0Vh1WA+KYO4P04i9jQlhYWfs+XbbwEIBYP8/PlK6qqawjERgBidh980B3dJCNEfImTz0fl+OfJcE81nC9yufYjbSu7CndVnzz1rPk6VlU5X/wqL5lgdR52Tzbm3T2fhNeNJ3L1exT7wlnfhWN0Y3vsXwVXowF1viPzODRIxKGL/oW6XWmgghP2rGnST44i6LA/L1DRGz5iJRj+w0RIdfWy/tri4U4fNIABora7i40fvx+sMbw+11VTx+dOP4nE4hu2a+4vMqBqwtOae2veHuXPnOr788kuLxWIJKhQK4uPjgzabTb5x40bDkiVLWuvq6lRbt25VA7z00kvRc+bMsU+cONGzp/ba2lrVtm3b1ABvvPHG8GrG/8oYjKegVRCEC4HXu4/PB9r20r8XURQ/AT7Zre2pPfS9dDBjDge+ege2r2rwVdnQjLFiPDoZZdzAEclBpx/31v5vP9Dsgv28Ecr1KuQjXIHbX9//puPa0op+VhLuonacP9SBNgbZ5H8hjlpC3Lrrkc24HOILBjW+x+mnYlMrLQ3tBIxtFO7YiE6n47jj5qENhZAdZ0K2yk7I5kOmV2I8NgWUMgJtHjzFHQgqOZazs/G5/Dg/LEfR5eW2giUIMrB1zKdiTBtLi54kN8FMS3XkewkGQri8WvSL18N3NvLaJ5A7dzrVnh2s/vh1ps87C9/KBhwlNtRJeuYuSOfz98rRWqFzt22PYDBIu8KHqvVc/C/V8nR2Aka9i9mTclmzuZRQKER2Rha56lRCHW2EAN9P2+l8fyUxl/wfxvrnSbn7b9T/+wECLS3Io6NJuvFSVGW7pDuEhk0U1dShTslgss2Pv9oOCgHlzCTaBRexibNQNvzY21/srGFLTScTUi3h82UCCpMabyiBHVuK2J3qkhKyLWn4x9+MvT4bz4dOVIl6zKdmYvu8CtEfwvN1A1sWlvBj0xoA2rIuJqVsJWL8ONpm/52LntuEyx9kydwsTh6XgEW3a5EokwmDK3890PekqP9vyr2pBcPMRAT5/mXkBF0+PIX9x/M3uTDlRQ9wRiRW62ySky6grv51IERc3ELiYk/arznsL51Njf3Sbpt3luHoaEdjGP7g48Fgmp9a1zemAACFLGSan3rApZOnT5/u7uzsVJx11lm9f7i8vDy30+mUZ2Vl+Z966qnKxYsXZ/UEFN50000tWq1W3FP7o48+WrVw4cLRVqs1MGPGDMeOHTtGZm/2MGQwRsHvgMeAhwjv/q4Gfmnw4SFHoMND6wtbe6PYXT834m90EnvpWGQDuPNlajnKJAO+nZEBUHLL4akFrxigNLEqxYC/3oHto273eIeXYEMFgePTaT7hbRKy4yPS4vZG9fY2fvqggpR5AX748RsAOjs7efXV1zj12GMo+uwDssfNxKCzYkm3okmzIvpCaH4fhegMIGgUuDs92J7fBt2Fkpzf16GbGo+qBbLrTNxy+nXoWqtRqmMjSv0igFmtpOWF+l0qjut9ZEwYi3jmb7FWROFpCAeMyltl6Ow+ps5Nwu31oVarez0aPdgcNqwxFsRiO4r1zYRcZuZFbWLC+b/HVeVFUxMk+E34nibTKQg01+HdvIW6m7aS8cT/Q7/jXjKe+H8E64uRxSSi+vJacHcHu+ljICYPv1zHspJGxs/NxlFpxx8U2bK2mc4mF6f/5hZSGxb1zsejjuai//7EB9ceTWZMn31+o5Wk7DzadtPOiE1KJBQ6jvbCbPx14ZgFj92Hr8GBfkYCju/rEeLVFDnD20WZpkwM+Ytg1IlU+K2c9HQhgW4P0m3vFqJRylk0KZmhQJVixL05MnhXlWnab4MAQKZWoEzW74pnkYW/54OVD1er48nO/hspKRchikF0unTk8uF9pugtlv5tUVY0+hFeNfShJ25gqLMPICzq5XA4IvZt3nnnncqef59xxhn2M844Y/d09j22n3POObZzzjln2+7tEvtmMEbBncAloih2AAiCYAUeIGwsHPYEWt2RaW2Av9qOv92DeiCjQKPAvCCD1ue3InY/gNR5VpQj5P4/UFRpRtRZZrzlYSNH0CowHptKx1slkR1FUNu82LtURFXacG9rRZQJBLJVvNL1FnqVnrkpc8m15ob7h0IEm0rY8pWdmacns2LdB5HDiSKegJwzzr4VV1cnDR1lfPD4f7jo/scwJnWv5ixhiWKhztlrEPTg2tSCYXYijm/rGNuVgyxYw2kXRPPeCzW9C64ZC9MQOj39ZJ0DW21M/N1JtD27FVmSFvskFSVNO9EqNWSOMlP6rpMFpy1g+YfLe/ctp06dSlFRETH5M+i5TXuLuhDPORlr62q8rokEi7srI8pAN1lJ8//7X+9n4S6tJDD1Nupa46mtMZIkhkiZ8xCmr6+mbdYjVLamYnNpydNFcdcpFn58tZSmPpUWASqqdKRaR0F7BWLeQhBD3D5LTUWzg0xvMVSsok0zhaqmWNILjqV8/U+4ujoBiM9IJy0zCX+HEf/PkVkhIbsfQa1AUMnQzTaT2xnD53I1t8+8HYs1HEz68gfbeg2CHv63eicnj0tAozxwwS1NnhXnuiYCTeGsE7lZhX5yWK+mzuPju3Y7q7sczDIbmBNlJEW7ZyNcppJjOiGd1mo7iiw5wYlNtIU+xKMdTYxzHgb9vmN/5HL1Qc02iEnLYNLJp7Hx0w8BkMkVnHjldRis4d9CR8c6WttW4vO1EBNzHBbzTNTqg+8VN8xMah8KI0Di0GUwRsH4HoMAQBTFdkEQ9lG8/vBBUA6wEpGFg7f2hDrdRNy1Ewm0uBHUcpQJOuT6w9RTYNFgPS8Pf5ML0R9EEadDblQhMyihT7Q2gKiUE6eR0/LMll0ZAz8IzFw0iSs3Xsf/tv2Pl05+iSxLFtStR/j874wa8whWZxCtWksnnRHjqUIKut4rx3xiOnHV8cw8+VwUyl2GWGutnff+bwOnnJzO7sh0iu6UPRCdATq+9GOY6uC86xLprKxBnxBPVMsy7L7z+p+rURBy+kEh0DlZwZtfvtv72k9b13HBBRfh9Ng49thjCQaDKBQKysrKqKmpQT1jLnKrhmC7B0EhQ6aSI7M18EW8nfyzx5EhJKOprqL10bsJtvVxYY+eyA9rtZRvCXtai9dB2pg0jjr/J95/ogqPww24YU07x12chzlW288o0EVbYMINYK9HqN+EbvkVnGnNRcx/Bt65nM7cq1n+oQyXrQG5UkbBCX8mKtqGwVtOdGIihhVX4TvpNZB5+8lYy00i6vRaWq67mTmjMllw3xvEJewKcosx9v9+xxs1KAYQoPolKGN1xP5uXPh7GBJRxutQRGmwB4L8o7SOj1vDRuubjR2cEG3ksTHpmJV7vn2pU03E/XES9V1vULbzP+HGNqit+x+TJ7+GTtf/OzWSaPQGjvrNheTOOgaXrZOo+ESsKeHkrc7ODWwpvIJAIPx9aGx8j7zcu0lOPnckpyzxK2UwRoFMEISo3TwFgznvsEARp0MzNhrPtl03cMOcFJT7SA9UxupQxh6GSmiudmjcAvYmiMqAhALkxrAh4G904q9qJyB6MR2fROsLNuj2xivTjWgzTNDqDqcMNrtwb26BoEhitYlkQzJ1jjoKWwvDRsHWt5HV/MjoqWpsb9ZyzCkzefPL93tX3gaDgUSVFdN8HWJIxDgjGVN0BsqACndxeCGi6vJyysIMZCY1RKkJ9ZHUNcxMxP5dHZox1l4hKMcGB1ETkoiKMSDziPiTFmNrb0MTqyXYssvAMcxKxLm2Cd1JaXy2/fOIj8dsNuMXvfy4Zg3R0dFs3ry597UxY8bgEfyIJ5pRfSfDOC4Wr19E0J/FqXEyVjh+IFNmIuSox1+9y3WviI/HY0qhfEtlxLWUJhkNdifjFmkxGRIp/95OVWEHPy2v4OQlBZStbybU7SFRaeRkpLnDyop9qlSq2osR23dAewUtoVxc3cWlgv4Qm1d2oTEoOW9uE/ofX4AT70Dp24Zl0QmEbCHEkIggF/DWtNH5xpPYPw2rLQbXd5BQvhPBWQF+J/hcXJUSy2lnuCh2W/nbjyE6nH4un5OJ4he49/eE3KxGvltMQrnL22sQ9PBFm50yl5cp5r3fhoK6Liq3PhLR5vU14nBsP+SMAgiroibnjunXbrNv6TUIeqiqfhqr9Ri0WklUSWJoGczD/UFgtSAIbxNeH/4GuGtYZ3UQkeuUWE7PwjcxlkCLG2WiHlWaaa+egsMWjw1W3gHrX9jVtvBhmHIp3mo7rc8V9qYIyqNVxFyYiqdBQFQrUKjkdL26o9dDoM62oC2IwV3YiswPSlV4he/veWD5wm5gded2CBoxrfFxwbyzqXc2o1aoSU1IxmhX0vX5zt6tAWWKgWCeG5lGgW1F5a50xWg18uPTEZtdqAFtsp5gpxvjUUkE2jy7KjOKEGwJ4lxeGR4vUY8lNwb5VCWIIkGbH7lRhWdHG75qO6pJ1l3zBbRaLRMnTqS2tpba2lqMRiPz5s3D4/FgMplobGzkjbeWhT+2E09BXe/Gs6L72kYFJ50+HfHtt7BX7CT2hhsItLQg02pBEPDYIhUsR02PolnYzpr3SoFwCtVpp5xBZ5MWj9OP3qzm7BvG07CtGrlGTUKMnRhlNYT6l4cWut9DaIBkoqA/hCjXhEWfyr4kqBuLo6SBQGP3FoJcwHJqHK33R8QDE3J1wnfPQs1P3d1kpJ9wB+nrbmHymS9QZ5pGQXL/LImhJiAOnA0U2Es6Wg+iGCQU6h8YHxL7i1Ydyoih/imqoZCPcIKXhMTQMhhFw5cEQVhHWIpYAM4SRbFfYMfhjMKsRlFwBGhbNG+PNAgAVvyVUOYJ2L/ujNAMCLb5CNa1oI8XCZnywvUL+tyHvaWdmE5Ix721ldZRHiq3V6KSqRgX012yeOJ5BNRZiKZ4lCkC/lonmk88ZKk1KGK16DRynOsaI2IF/LUOdONj8exoj9QvsKpwymw0yp0kRJtou/psLP++C5vCgjxWhibLQKjcgTo/Gu/WXcFqmilxdOLApXOhNehQBv3wTVNvLIiokjF98jQ+/PQjAMaNG8eaNWuIj48nKyuLHTt2UFRUxPHHH8+6desYN24cxxxzDMFgkO9++oG4MSfuegzbA6hqlNjGzUFfMIWW225CZjAger0gCMTcV0D8/2/vvKPkKM69/dTkPDubc1RcRRQRKCGyQGQMCCOwAQMG7M8BfC+2wThgX4x9fTE2wWSwSSYJDAiEQCgjCaVVztqcd2d2ck/398esNmhW0iqsYj3n7NFMdXV31cyo+9dvvSHPSm153GKRNsCIJZpNbn4WiqKwatUqPv/yM66YeTVayIgjtBWHKUT6oCZ46+Z4XYWskTDoUtj4AVrqEKKl30cVHgyegRiSi0k1lmM0Z3dzthw10Yp980sw4luw/FkiY77bKQgAYhr+lT4c0y6g7dO4MDAW5GHO9sDSZZ39NBW+fgYGXkLqskdJvfHNeHnJPqbYamG0y8ZKb2eGyxEOKyW2g0c5WCxZ5Ofdxq7df+to0+sdOB2D+mSsfYXLNRydztRN4OTm3oTNlneAvSSSw6NXywDtIuCUEgKnJT2ljI0GIKIQrQ8mbgom0fphE/bRTfE1+H0xCEw35vNWw1NcVHgRNw6+kUHJ8QtuWC2l8WsdajCA85w8zIVuwjtaMecL9EOy0eqixFoS0wFriorS2uWmNcbN4vAmyt6P//zS09OZ/n9/4a1Fi2hoii8zjBg8jLO/NQYDFvxvxh0kddlWNlLOp/M+Q1VVLBYL55xzDulXpOFYFUad6OG9rz5GbzJwzjnnsG3bNjLS01m+fDltbW1MmDCBlJQUysvLsdvtDBkyhIULFxKLxbBYLEyZMgU13D3RR7QuxNJyO2azjbP+9iKBp/+MIS0N+9ln0/jIw4z99h1UZvfDq5gJ0cKCBQtQVRWr1cqUKVOYN28e5VW7KCsrwz06hYzPvg/5Z8G3XoX5f4gnDZpwN2rxdPxNpbTObQNVQ2erJ+Xa97E3f8m0b/djy8pG/M1h8oYkk1+kILJujZdPjgaJhU1A98891hzBPf1ClIpd2Ib3I2l0Bkb/Po6mAK0V8VLRWz+NW4LMfR8ql2wy8NfB+fy7ppnPGr1MS3ZxbaaHVNPBcwYIoSM399uYzZlUVb+Jwz6A3NybsNtProJ5bvdYhg97horKV4lE6snMvJLUlHOO97AkpyinoI1csl+Si8G4jx9EWik6pw376MTKpMZsJ+Z8F9HqNkwF+yQU0wks/TykDy3g4UkP84dJf2Bk+kiEEMT8EUJbW7GPzcSQbsM3r5xgWSOeSWHaspZw7rIZPBp9CtNwT8ex9EPdcG4qaooBS//2dr2gzhOkbGunHq2rq2NNeTl0qSa5ZuM66g0tKDrR8fQaGmbhk7lzUNvTCodCIRYtWsSOpj0ok9PYUL6N3RV72LFjBwsWLMBoNOKyOClpL2i1ZMkSNmzYgMPhwG63s2DBAmKxWMexVq5ciTnfjfH6fJwzCjFm2YlmOWiu8VO1w8+eKoExOxvPjdfT8vbbKHV1BP78G9Je/i/6Fzcyd+7cjrEFg0GWLFnCpEmTqKqqoqmpiQV7VJTcM2HP4rgPyC3/gaufh7zxRFIuoPULH7RHA6gBhaYPqgmkXcSc5zbjawxhcRhZN6+CxZ/6iaQMj9/IlRBGe0vC92wrNeOs/F8KrtCR7vkUc9n/gj0tMey0//mwayGMnBkPoTwYwVZo2gkhL6oao7W+rseaIT0SDUHTLmirp9hm4f7iLD4Y1Y//Lsmin93SrasSjdBSV4O/h1oGZnM6ubkzGT3qDQYP/gMuV+/ya5xI6HQ6UlImMWzoE4wc8Qp5uTdhteYe72EdNY5m6eTHH388ZdeuXR2KMScnZ1h1dfUp4wN3LJCi4HQitT/c+BakDoi/zxsPo26Ej+/HNsiIY0Im6EU8LPH8fAIr61Aagpj7ebAMScZckgSA3mMm5eZSjFnx4DyT3oS+Pbue4gsTWNOAf2k1voWVmHIc2M/MItYSRmkIUq20EYwFeX/3+zQP1GEcnkr00hQ+8C/h2UVv8Nay/+AtEdjGZqBPMlPTmlgkZ095Oenp6d3b9tTyxUe7MVxchM5jxqckWj68Xi9CCGpqWti1pzNFsaIo7Nixg+WrVnLe+KkM6Bf/fPR6PSOHjqChvqFbStWsrCyGDx/OP995nadnv8QXlStQz/YQNus7HAMr63S4zpuCtaSA7N8/gm3CBAAM6WkEnYmx562trXg8HrZvj1ek3FlRQ2DCz+Mbt3wCOiPoDWyp9bG73J+wf6wpgtoWt7A0lLexZ30TkVCMqm0thDwjYPJ9YLBg2vEknssz45kkdWAbm4FjfDoCFd2OTxEGC1zzHGx4D879VbzglRDQ/wLIGR2vjDn6O91EWY9UrIRXroDHR+Jb8A8W/esFXvjRHbx0392s/vQ/hPyJc+igYSu8ewf8dSQ8Ow22zIGYgkWfGPrYXF3FnCcf5/kffI9/PvAjtq1Y2q166V70ejPiUMo4n4DodEaMxhMnb8HR4uyzz25bunSpA+JJwpqbmw2bN2/u8PRevny5Y/LkyQdN7agoCq+++mrqnj17jn2++FMIqaBONwonwtUvwMb3oGYdfPpL0FQM0RDuq17ANj6L4Nr4TX1v/gbf53uwjckAs470H5yB3mVC7+gMUdMUtcMxM7LbR+vszlLLgW/qsJ+VHd/H2ExaTMWitxCKhZjfspTRmROYt3g2be3pXCurKnnrk3e59ZIppGaVk6vP4Ou13aeQn5dHeUX3elxOq4edFW3M/vc2hozLIK/YBou67+dyudBUjZbKKNlp+ZSXl5OXl4emaZSXl1OUmYdtd4xzgoOZfM4YLA4LhjoVZ5EDIUSHMCgtLeXzzz/vOO7qDWswGYyMcQzBmWLB1xgif5ATV+OL4HNByRRy/vYEakMDqreR8s2JNRNcLhe7du3qOEdBVh56NQtt8OUwYDqqBoFglP9+ex3fH5TJwH3213vMmKwqWSVuana0duRqyOyXhNluhcn3wxmz0AkddmcGltJcVCXuTyMMurhY9DeANQlsyaiZo9ApQRh2NcRiaAY7qBGEIx0MB7nmtpbDa9eBvx5syWxtsPD1B+8BEItG+fy5J3GlplM8qodq6ZEgzH0YNsXj9WnZA6/fALd9AdndaxUo0ShL3n6NTYu+BMDX2MD7j/2OG3/7JzLbhZ3k6LJ8+fLk+fPn57S1tZkcDkdkypQplWPHjj1upZNzcnKG3XDDDQ1ffPGF69Zbb60vKyuzzZo1q9hisagrVqzYCPDoo4+mz5kzx60oinjjjTd2DB8+PFRcXDx0yZIlm7Kzs5VYLEZRUdHQZcuWbcrKyjrtvTdPbuksOTwaN8NXf4w/ge717t7yMaKtCi0UzxPfNaGTGlAQRh2OcZmYsh0dgiBa56fl453UPbUG7xflKE1BIrsS/RZCm5pwXVKAseZ98le+wu9Kv4tZb+aflS8gUsIdgmAvbW1ttFaXYZx3F7lWPUO6hGmlp6UxsqSk25P7sEFDKXJnMGB4CtFwDL0AR4OeC6aej679idZisXD22WeTn1XI5kUNZHryueiiCxFCoNfrmT59OsXp+QTLGoltb4OPawm9tZu2z8txVGhceMGF6HQ69Hp9QqZDgLVbylDcKiMmZpFdbKMkaRPeUXexolrl+eef56O5c2k0GlFqa+H1Nzhv5MhuY5t2zjTWrVsHgCfJw/iCkYTKwtQGLmBjZSnv//kbFr2ymdtKs3lzZwPBSVkd/3uF1UDS9FzCS3yc7TFx2bf6k5Jlx+Y2MfYMQfXd38P77qvEandA7Xp4aQb6OXdi9K+h4+HZ4oKUYsLVzdT+8TF2f+cuGt6eQ9hnxr/TSv0/K2me4yNclWiBSaBpV1wQAJGcs1j3zcaELjvXfNPzvr5q2Pxh9zY1Bo2JPg7+5iY2LZrfvVHTaKw6WAFXyeGwfPny5Dlz5hS0tbWZANra2kxz5swpWL58+RFlUeqpdPKYMWP88+bNcyxYsMC2t3TyG2+8sX3Lli0bFEXhj3/8Y4dnuMViUVeuXLn5+9//ftPQoUMDL7/88o5NmzZtcDgcGkBqaqqyYcOGjd/97nfr//CHP2To9XquueaaxmeffTYZ4P3333cNHjw4KAVBHGkpOB2xJCW2OdLBZENY9PF1+X2y15n7eTAXdvoVxLxhGl7eQKwhbrKOVrQR2ePFNjJxrdmQYsHcLxld8SPQtJ3zzA76F55HfbiZ1LCr21M4xMPzLBY7yrgHCc2Gs3MGMWpSKTFUPDYnHv/X3HTZFdTtaEGPwLpDQf33boZMyiEQjlGIRuDdHRSmW7j53G8RsCpY7DZMu6MYV7Vx3XeTqFfreXd2Z46CnTt38u2rb8CdYulWYhdASdKzcdVGJk2ahNvtRunBPJ2UlITXGiY5N4mBgTcxrVrAvPyfsPCbeA2CyspKNmzYwC1XXIaybh0Zzz/PzKuvIWQ04NTrsTc7uG78DKKqgqPNiPZJA2KklYakUuZ/uHcd3otubT3nX57P/yur5I9X51Pk345IyaXx37vRgu3j2t7C9Ov70fzZe/h/8g+0YJDKr1eQ88u7cRUB9ZviVqKtn8ItH8aXBYBoTQ3ld95FtLw8Pu/aWgyZE2hbGH8QjOz2ESxrJO2uEZgOVAHR4o4vOWgaen81KWklNOzZ1a2LJyOz531NdnBkgK+me7vFk9DVaLbgTE3DW9e9Yuf+Ch1Jjoz58+fnKIrS7UFSURTd/Pnzc47UWnA4pZOBOoBZs2YlOpN0YebMmc0A48aNC8yePdsDcNdddzVcdtll/R588MG6559/PvWWW27ppbPLqY+0FJyOZAyFfhd0vhcCpj8GriyMqVZc5+d36+6YnIMpx96tHGm0LtAhCPYS2tiEzmFGn9zpCCaMOpyTcjDYTeDMgIKz0GUOpyi5P+OyxuEJ2zn7jDO7HWfiqAm4dSYCrnNR26Kom72YP2vB9pmX8AeVxGJGWrxt1IUbaYp5iRaY4pkp19Rz9rRclPawRLUuhP6DepxvNpO8R482r47ILi8WUxYrN+1K+FjWrF+LY2J2tyyX+mQLzYqP3bt3M3/+fFpbW2lubiYjo9PvSa/XM3LkSD5f8AV7tregDzXQMv7nLF27rfvnEwrREImSdNUMlC1bif7+9+h//Rus1dUYI/WY53uxz/WhLW2Kr/mX5rFq5T4piRUNe5tKUFFJM6pElVzC1TqcE3MQ5vY1dw2i6xtQ5n2IFuwUOKEmEwH9+bSOmINv4mf4xj5HuLwamncCEN6xo0MQADinX0Fgdff8ClpU7bGIVjdS+8PEn8Q/m6qVjB5TgrFLBVFXajoFI0b1vK8zI/5b7OrkWHIeZCY6CNrcbs79zp3dfAVyBw8j/QQqN3wqsddC0Nv2Q2Hf0slTp05tW758uWPp0qWOkSNHBg60r9PpPGBpW4vFogEYDAZNURQB0K9fv2hqaqoye/Zs56pVq+zXXnttD6FZpyfSUnA64syAy5+Ie7UHGuOOhxnx/ALCoMMxIRtTgZtYcwh9khljlh2dqftPRY30kABGgLDqSblpMNEqP1pUxZhpw1y4/yQ3ekUwuC6dvKlX4lMCOA02nDs1ooPcNG2tT/yB6nVUOAt49Z03O6wLdrudb029BNNKP1aDjqAQdEuqQPd7TMynYiDRac1oMaOGVTzXDiBaG0AIUMMxWrzeLscRrFixgksuuYTW1taONMhLlizBbDZjshhBn4eiK0WnS4zi1et1pF02AueU8SiNLZjycjFotbTY7PjPdpPiMBBqU6gPKNQuq0fXQ8bADLeZf10yhPA/NxNud2yM16zIjZfBBoRRj9bFomGfPA2dayxN73Y6blpKM2j5z6ukX7wb8/hLEfs48mmRCMLUQy6Cg+UnMFrh7B9A8VRorSAruYiZv55IQ0UleqOBtIJikvZnKQDofyHcNg8atoA1GbKGx3+zPVA4YhQzf/dnmqrKsdjspBeVdNQLkBxdHA5HpCcB4HA4jkrp5L/97W+Z+fn54a6lk7du3Wp96qmn9rz66qupZWVl5qFDh4b3lkjezxhjra2tvSrG8d3vfrf+tttuK7r66qsbDQZ5K9zLafNJRBuDKHUBhEmPMcPWzVHudENpCROtNSJ0YzAW2hJSy+osBizFbsDd3j9EsNyH0In4Z+cyowaiGLPtRKs6vcitw1IJ60O4MtIwZfXOhGvMsqFriJKe7CHLlYQWBeMYG+vK5qO1Rcj3FKM2d/o32C9K5dOv53dbbvD7/VTHmhh2bn8ie1pJuqIEYiBsBmKNQWK+CGq7ad1U5Ca8tZmRRUPYsHVTx3F0Oh39+xfyaW0LF6a78S+pQg3E93GO85CRnkFtXS2xWAyTyYTX62XRokVEIp3Xw6vOvYwMTxJq2xSUz5uZNGoCny/tXPN2Op1k2AWGmh04rBbwzobohbD2dWoLn2HV8jZyB3mwG3WkJ5lx59gpPSOND/6xHrV9Ocdo0VNUmoJuQyPinLy4Q2FMpW1xNTFfFJ3diBpUsJY64/4L7TimXkJgbfcn/NDGAK7bf8aO5gi2ciMp+QOwDBtGqN23wfvRe2T+7gp8c+s79hFWA6acxO9WaQ4RrQsg9AJDhg2D0w1Fkzq2pwKpBb18gjeYIGdU/O8g6PR6Mkv6kVlycuUeOBmZMmVK5Zw5cwq6LiEYDAZ1ypQpx6V0ck/HmTVrVsO9995bcN9993U4Gu6PG264ofWee+7Rf+9730uss30aI7RepAs90RgzZoy2YsWKXvcPV/hoeK6sY83VPMCD5+r+GA6z9vvJTLTWT8NL64k1xZ3lDBk2Ur49eL91HKK1fhpeXE+sve6AIdNGyqwhtH62E4PDDAKUhhDGTFv8damBlJxDi6GOVPhoensLSnV7hTyXCWWKhbee+CXnXvM9UvU56FsFIs+EIz/Ecx8tpKmp+xLmOZOncma/UbR+uKOjFoLOZsB5Th7BDU3YRqURawihRlX8i6vQpZjxT7CxtWkXQoXBpYOYW61yscsNH2zHOSmXWGsENRhFV5LErqCPAHXU1VQyZNgw6hsasTud1NXWEmwLMCC9GFeZgjIikwxnCw3/aoLBTpoKFLbW7CTVmcyA/oVkvTkBlHiWQ876Iez4EoZfx9e109ixzsuZ03IwLalCbV+a0XnMmC4pZu3KWuzJVopHpJHkMND85hYiu7yd85yWT7TGT8xiwJoH9sW3ECq8Be+6OtTWFlxX3Erz2/us0wOhKbnMeT++fFA6MYvRk1NRFn5GcNVqHBMnYh1/FrEWHcGyevRJFqxDUhIEX6TaT8OLZaitcYFkzLGTMnNwj2W5JccPIcRKTdPGdG1bs2bNrhEjRvR6Pb0vog+OF1999ZXtRz/6Ud7KlSs3H++xHGvWrFmTOmLEiMKetp3ylgI1EsP7WRcnLCC8pZlIue+0FAWBVXUdggBAqQ0Q3NCEcUrPosC/so5Ycxi924x1RBrCIIhWtmEdlkLzy5sRFj16t5nw1mb0U5NIcu+/0IzSEkb1R9E7jehdnZ99pNzXIQgAYt4Itlo3nsxs5r75FBaHE7snmYml3yaj3wTGj4/y8ccfdzt2QWEh0Rp/hyCAeNREeFsLxFSEEIQrvES2x2+kamMY+3yViRNHoIVjGNos3FzqouH5MmKBGN45u9E5jeisRgzFSWxf6uXimYMwOWPodr5DkXsQTQsdpPuT0BlTUFZ60fdPImozYAquw5hTQnSjj+TNgrNTiwENe04D0exxGPcsAE2DJX+FST+FzZ+QO+JcyhaFcbZFCXbx1VCbw4hdrUy7eUj8M2xuJrzd2yEI9s4zssdLeIyL2rpWivyrEahYTZVYp+XB4O8RCZvQJTWjdskiqUuzUtPlXBsWVlM0PI3CmTNh5szODzcTrIO6OJhHg9C8CxBoSYXx8NXWTotJtNJPaGsLjgOIgpii4q0PoqoazlQrJnOvLL6S48zYsWObTlYR0JUHHngg88UXX0x74YUXdh7vsZxonPqiIKQQrUh0jIo19SK06hRDUzXC+5TjBYjsaQUSn+5jEYXIzlaM2XYsg1NoW1CJFolhyLCR/K0BuK4pxv9lFVpUxTQ1FfMgNxZHomlZ0zRCW5ppfmsLalsUvdtE8nWDMBfHlycilYnfj1IR4NzvfJ+5zz+BpmqMnXEVuaXDARgyZAhqTGXpsqVYLBYmD5uAdYkfJTnR6hWtDWAucqM0BvFc3g/fV5WENjVhzLJjLU1Bi2m0La9Ci8TwXDegwyICoPqiqL4oNn+Ui87Lo/WFragBHfrkC0g+R8F9rgvvIj9KbQDjyHTEkBTSMzR0LzxA8rjf4fP0I7QrhkjS488L8sb//on+w8cw8YyRuFb9NR5qZ0+Hy/6P9HWzuejaqcQ29PBbLW+Lf3dbNlP98MO4r/5Zt+0izcKe3DY+efs9wuEwqSnJXHXOH8h+/1qIBmixOHgjsIvB54+k//o0tD1BjMUuWtLsrHu3e84Ef2tiuGU3Wsrhi0dg7WuAQLvgT0T2jEzoFjmAM2LAG2bN5+Ws/qwcVdUoGpHK2df0w30yVh2VnJQ88sgjNY888kii6Uxy6osCvd2IZUgKga+7f//GzNMvbEnoBNbhqUR2dxcG1sE9p6xtrqsimgNOVyreTztLASu1AVr/s5OUWaVYByajRCKY3Hb0+3HWURqCNL26saPIUaw1QuM/N5J+70gMSRbM/ZMIrOgeVmYdmkpqSRYzf/dniIDRaEIIHTF/BIfDweiSERRUOREhjdgnXqIxFetFroRzm4vdRPb4cJXmo3eacZ1fgH18Jv5l1SjNIfzLazDlOuMVH8saMeY4iO4jUvQpFnzzyrEMSSHWHCa8vYWmeWaSb8/G+S2BTgXVqMPqMIEShYEXY/zyTjw546k5/yG++vhN9syJZ2Da+PUKMrIuZLQtJb6MUDwVkoswJGWS/ckVtI1/n8im7nOwDk9FDfhp+2oZhrQs9LlOWNFpEQmPsjL7i7c7/CMaGpt4/6s13DzyVmzrX2OtLsYTG18CXmJM6mjO7j+BC7MvYuXLNR2+CgAIcKfv58Yc8sXHu+UTWPOv9kYN3eL/wTp0NtHq7hkKLf2SOl6rwSCxtjYMHg/CYKB6eyvfzNnTsX3nmgZS8xyMu7S453NLJJJjxikvCoQ+HhIXawwS3t4KBoHr3HyMeaefKGiLtFGV3YJ9uAPWxW98tnEZmPsn9di/paaKhtbtDE6alLAtvKOVWFsUY6oVIwdeO461hLtXPQRUf5RYSwRDkgVLcRKOiTm0La4EFSxDUuJJaP61EfuYDNq+qkQNKFhHpKLFNIwZNnQ2A+qK7uHJkV1eHJNzaVtUCTENc7EbvceMxWEkUh+EWBO+rypQ/VFso9IxFbjQu80E1zag+iIEV9fhnl6MP6rGnVLNelwXFCCMOkxZdkIbmzCm23BfWoz3s91Ura5n/qflDJ2Sw6Az273pDUY4+4fQvAsRbGblgs/Ys7F7SsZN63cycvhk9GNmQWq7g1z+mTByJpbWt7GPmIF/bdySZTsjHXOJm7bF9UQbBmHqPxSdy4bhzCSU5a0Q0/DpQ+zrG1RbV09j6TkYxhSwvnVrR/uKhpWsaFjJ+9Uf8uR1LzH/n1toqQ1gsho466oS0vOd3b88JQK7FsRLbicVgH+ftNO+GmzuTUSGDSC0rgF04Dg7B3NR3AoULCuj4YknCJatx3n++SR/9ztUbUm00m1bWceIc/MwW2WGWonkeHLKiwIAY5qN5JtKiTWHEAYdhhQr4hiUfT3RWFm7knu+uocz08Zz2eUXowGZuWbGuXtOCetISmbRulfILxqRsE2fakFn6d3PR2c3xotud30oNerQ2eP7650mXBcWYMxzInTgX1qDb34lSTOKaXlvG1o4LijavqrEMSmHwKo6HJNyEs4T2e3FPCCJ5OsGImwGNF0MEYjR+mkFrvPyaHptc8cYfF9U4Dwnj9COFqK7fZiL3Ai9jtaPd2Iblop1aAqaomHMsuP7opzw1hYAwm2tRCrbcEzOYXtLBH9LmGXv70CNaYy9pDCeyyF1AFz/ejwc7+v1bF7e3Sk2d+Ag9NMv7x5m5ymAab/AULuBJAw4Jpeg6U3ok820fVmBb148S1/MG6H11W3ELrTRMqoFg85ErIdsIw6Hg+jX39AUaGbodYle/P3dhaQV25nxgxF460OYHQbScp2JB6otg09/Hg8PDDRCyTTYs6RbF0NsF8nXXopybj5CJzAkWxAGHZHycvbcehtqazwEvOW114hWVZHy7V8knCaz2I3RJP0KJJLjzWmTvEhvMWDKcmBMs52WggDgwx3x9LFL65fxwOZf8fPNv+K5Dc8lPGVCPMTM3uDg4pF3YgjpcVzSpXa7QYfnyv7oHb17qjOm2XBf0sU0LCDpin7dvNPVgIL3s12Et7RgSLfiOCuL4OYmHGdmYz8zq6NfsKwBy8BkYg1BPNf07zymTuCcnItvbjnN/9nOnq1rCLVuoPWzChRvJF4aep9p+lfWYjsjXlgpsKYO55RcUDUCq+vxzSsHVSPmj3YIgr1o4RjCbmLN0s4lqXVfVnRfj7c4IWMwxWMmkJpX2NHsSk1nyDkXgjMDTdNYV7+O/13xv/x2yW/5umEdoYyh1EYKWba4gWXzK/FV+WlbUt194CrYow7mvvU0n7z+ODsWfsG4sZ1O5Xq9nouGDCH673/T8u5HDNOyGZs6vGN7kjmJ24Z8B7PBjCvFSlKGlcbyNj57YT3r5lfQ2l5GO7x1Kw3vzqdq6xi8Jb8iljQE3LmQ1CW5VepAGHgJOpMeU6adsEnP5hW1fPb8ejaUhTDfdGe3JBH++fPJzBSkF3Uu9dhcJoafk5uQk0ENKQQ3NdH07y20zt1NpPqgNXEkEskRclpYCiRxMmyJCWAy7BndMhUCxIJRmt/dRnhLp3leDPbg+V4pIqRhTLViOASnMGHUYR+fibnQRcwbRp9kwZjeXZwJg0AY9Rhy7ITWNhDeEX+6DAHmfklYBifHMybajER2ewltbMJ+VjZp945EbQ2jd5pQQlHsM3JRrFFsxgBJugANLiNKYwhdD0+hOqsBLRhPwhRrDhNYUx9fLjAZ0EIKwc1NmPq5EUZdwvJHqz9KwNvpcW9xGDEYEs/hycrh6gceprFiD6qqkpKbjys1nrZ9feN6bvnkFiJq/DhvbHmDt8d9xOd/3dax1q+FFPrbDMS6RM8AGJQQl110JWE1Ruqo0bgHDmZwTi6tW7fiCIZQH/0jMa8XQ3o6Dlc6j034DdtathJSoxR5BpCXErcORSMxls3ewaYlcYGzZVktmcVuLrg+l5pbb0Wpi4eDt378ORn33kLyN6/AqFnxhEI6I6QPildSBJRojBX/2cn6BVXxYwFp2f0Zd90sgq+/1P5bMGI3RLjkzmE0VvtRFRVPph1XauISVHBDI81vdtY8aFtURdqdwzFlnHqVAk93fvazn2W+/fbbKTqdTtPpdPz973/fvXDhQvuPfvSjhoNlLLzuuusK7r///trRo0eHDtTvRGTbtm3GG2+8sai+vt6o0+m4+eab63/5y1/WAdTW1uqvvPLK4srKSnNOTk74/fff35GWlharqanRX3755SXr1q2zX3PNNY0vv/xyh4POvffem/PWW2+leL1efSAQWHU4Y5Ki4DRietF03tzyJsH2ssJmvZlrBlyT0E+pD3YTBADhjc24zy3AVNyDibkX6Ix6TLlOoPv+0boA0Rp/3HowvSj+ZL6je8bR8LYWXOflE9rUhG1kGjFfBFdOPpoGql8Box6dy4ReBc2nYBY2jBk5CLsJ5zkthLeXoXeZ0LlMqHtv5AKcU3LwLeksnqPUB9EiKm1LylG9ESzDUvAvq8E+Pou2hZ35WUyFLsq7CAIEnHVlCZZ2y0m0IUC0OgCqijHLjiM9pccse/PL53cIAoACVwHbltejqhq5/d0MH5aKWdVwDE8lvLqO0Ob4d6JLMhLdsQrl8T9iGzMa75SR6LU2clJTif33A0R3dkZZZfzXf2HI7UcyMC450ZGvtTbQIQj2UrOjleaqJAznXkps4HiimLDHWmh+53kc374cU/ZI6H9+4rHqgqxfWNWtrb4qTGTiKCAuClLu/j6m/HyEXo/tACHBsbZoN+dWAC0YjySSouDUYu7cufY5c+YkrVu3boPVatWqq6sN4XBY3HTTTcW3335708FEwRtvvLH7QNt7i6IoHMvMhoqiYDQa+dOf/lQxceLEQHNzs+6MM84onT59unf06NGhhx56KGvq1Km+Rx55ZOsDDzyQ+eCDD2Y++eSTlTabTfv1r39dtWbNGmtZWVk3NX3FFVe0/PSnP60bPHjw0MMdlxQFpxGlqaW8cvErrKlfQ0yLMTJtJINTBid2VHtOaKXtp/1wiVS1Uf+PdR05JHQOI0lX9++xrz7ZguvCQjQguK6hI3RQmHS4Ly4iWunD+0nntcF6RjpJM4oxF7pJu2sEkQovnmv6odQGUYNRDCl6Gp/7HfaxE3FeN4WYN4opx4GwGtA7jaDXYS500fjaJoiquC4sQPVF0dkM6NOs9Mt2YM+0E/JHyShykdbuoBet9VP/3DpUbzwLo7DoSbttWLsg2mf+se7ZYQ06A6oCyVk2xvdPQpm3BxVoBewTsvB8a0A854ItTPOiVYR+dQ9fJjfx94XfZVzmOP4w6Q/kP/0UwdWriTY0YB02DOuwxJoBXVH3k7wspjex2jiRqi9DQBAhzJw7635EqQmKevjNEA893XeJBsBYXELyT36CZdAgrMOHJaRT7vlYKlqsh3vBUf4NSg6Niop/Ju/c9UROJFJvMpnSIkWF91Tm5t54RHkLKisrjcnJyYrVatUAsrKylN/+9rfpdXV1xilTpgzweDzKsmXLtrzzzjuuX//619mRSEQUFBSEX3/99V1ut1sdN27cwMcee6x88uTJAZvNdsatt95a9+mnn7otFov64YcfbsvLy1PWr19vnjlzZlEsFhPnnXde6zPPPJMRCARWffjhh87f/OY3Wenp6dENGzbY1q1bt2HWrFkFa9euten1eh599NHyGTNm+BRF4e67785dtGiRMxKJiNtvv73uvvvua7jkkkuKZ82a1Xjddde1Alx99dWFM2bMaPn2t7/d0lP/fc+3ffv29QUFBVEAj8ejlpSUBPfs2WMaPXp06JNPPkmaP3/+ZoA77rijccqUKQOBSpfLpV544YVtmzdvTlDV5557rn/ftkNFioLTjIHJAxmYPPCAfQyptoTQPGOBC2MPJt7eElNj7PbupiHYQIYtg3xXPoEVtd2SSqltUbSggjHX0S23hDHHgaZqBFbWYhmc3C2XgBZRCW1pRg13r8UQXFWHY1wm5iI35jwn5jwnwXXrqPjRd9GZzaiN8cym7osuwH5GZx7+jrDJ5jBqOIbrwkKaXtpAeFsLwqxHU1TS7xyOPc3WY1x9cENThyAA0EIx2r6uwZPjSFimmZo3lRc3vIjaXr56e8t2SsamoKXHUJZ0f+L2L6nGPjoDU66TxmAjd2+dw07vTnJ8OfxX/wcxR+w01vpILSzClN+9oNVelKYmIjt3IoxGTEVF6J1OkmwB8ocksWd9S0c/T6YNo8NC1Z5Oa6ymwbJlUfLOGY7R0HOKcFeqleKRaexY3ZmB1p1mIW1wDo6zbutxn65Ew2GaqiqIBIN4MrNwTs2j9YPOPArCqMPYQ4plybGhouKfyVu3/a5AVcM6gEikzrR12+8KAI5EGFxxxRXe3//+99mFhYVDJ06c6L3hhhuafvGLX9Q9+eSTGfPnz9+SlZWlVFdXGx555JGsr776aovL5VJ//vOfZ/7mN7/JeOyxx7o53ASDQd2ECRPa/vrXv1beeeeduX/961/THn300ep77rkn7/vf/37dHXfc0fToo4+mdd1n7dq19lWrVq0fNGhQ5KGHHsoA2LJly4ZVq1ZZpk+f3n/79u1lf//731PcbnesrKxsYzAYFGPHjh00Y8YM73XXXdf0xhtveK677rrWUCgkFi1a5HrppZd2/+Uvf0ntqf++5+s6js2bN5s2bNhgmzJlShtAY2OjYa9gKCgoiDY1NR2T+7UUBZIE9A4jydcPJNBusrYMSsY2Mh2d7fDCxRRV4ZOdn/DQ4oeIqBGsBitPnfcUuTWJP7/ghkaSrxtI28paIttaMJckYXCZCK5twD2jmFBZYppypTGEIS1RsKj7rMNbSkspePopml5+GaWpieQbb8R+1lmd/UMKoa0ttLy3DdUfjS8xTMsn+TultH1Zgc5uwnF2Nsac/S+hKPWJBd2Uan/8CVffXRQMTRvKsxc8yysbXsEf8XNj6Y3kpqag6sP4Fiamk987n6ASZKd3JwX2An7oepAt//KhKhpN5nIst7koHJaYdyK8YweVP/4J4U3xJAjOiy4i486ZmN6/lsnDf8q2jNHs2G4gd2ASAydkU787MclVW3OIaGT/llyTxcDZ1/Qjo8jF9m/qyB7gYfCETBwey3732UuwzcfX773Fig/fBU3DkZzCt372ezzfGoB/WTV6jwXHWdkHLtks6VN27noiZ68g2IuqhnU7dz2RcySiwO12q2VlZRs++eQT5+eff+68+eabSx588MGKrn2+/PJL+/bt2y3jxo0bBBCNRsXo0aMTPE+NRqN2/fXXtwKMHj3aP3fuXBfAqlWrHJ9++uk2gNtuu63xV7/6VUe2tuHDh/v33qAXL17suPfee+sAzjjjjFB2dnZk3bp1lrlz57o2bdpk21t62efz6Tds2GC55pprWu+///78YDAo3n77bfe4ceN8DodD219/k8mkdT3fXlpbW3VXXXVVyR/+8Ify5OTkAy6X9DVSFEh6xJhmw31+Ia5z8lCDCmpERVNUhOHQA1Z2te7il4t/iaJ23tR+Ov+nvHvGy0T28R+wDU/DmGbDc1ER0cYgjS9viIsRVaPxxfW4Ly5KOL51SArhfW9iBh2qw0g4EMXcLmaEXo9t9GisI0eiqSo6Y3eRE60P4P10V1wQAGjg+3wPqbcPI+22YSBEj5ErkZpGtBgETKCOcaPf3UqssdOaYRubgRpQUCMxDG4Tot0h0agzMjZzLCPTRtJWWQlKDJveCHkmAskWYk2dT+rCasDgiF8r0qxpXFhwIaOMZ7H5VV/Hsk40HGPuCxv41gNjuznuaa2VtLz+eocgAPB98gmO8Wdg73c97rWPMVpvYmTOSPRn/hyy7CiRGELELQR76TcmHbv7wIXEXKlWRl1YwIhpuegMugTryP6o3bGNFR+80/G+ramRua/8jcvv+wVpI9L2+9lLjh2RSH2PX/7+2g8Fg8HApZde6rv00kt9w4cPD77yyivdnHA0TWPixIneDz744IBpiQ0Gg6bT6TqOubdU8oGw2WwdN+H91QLSNE386U9/2nP11VcnqOUzzzzT984777jeeOMNzw033NB0oP4ffvihs+v5AMLhsLjkkktKrr322qabb765ZW97SkqKsnv3bmNBQUF09+7dxuTk5O5POX3EaROSKDl01GiMwLpGav+yito/raD5vW0ojYeeHrouWNchCPZSH6ynJSeC6/x8hFGHMOlxTy/CXJLUeX5fBKU2QGRnazwLowqhTc24LylCWA1gENjGxEP7XOcVYB4Q39eQZsUwvYg3n1jL7MdXU7ePYBB6fYIggHhYpFKfOL9YSwih1yXclGI+H/5VO2n8uobFq1fx92ee4unXnmPzSD+MSkIYdTin5aFPMlP7+DfUPraCpn9vJdrQeQ5/UxPfvP0mr/zip7z84H0s/ucLtDVWkvLtwZgK42F7hkwrqeeHMbw5DVb/C3NM4e4z7iaL3AQ/j3BA6YyKCDTB4idQP/oVbQsXJczLv3IdFW/sxJt3P6olBf2OOdAUN9en5Dq4+M5hODxmhIgLgnEzijEYe5dLQG/U91oQAHgb6hLaKjetJ+Tz9fjZS449JlNajyWS99feW9asWWNet25dx/r4qlWrrLm5uRG73R5rbW3VAUydOtW/YsUKR1lZmRnA5/Pp1q5d2+viNSNHjmx78cUXPQDPP/988v76TZw4se3VV19NBli7dq25urraNHz48ND555/f+uSTT6aFw2Gxd5vX69UBXH/99U0vvvhi6vLly51XXXWVF+BA/buiqirXX399wYABA0K/+tWvuqV1vfDCC1uefvrpFICnn3465aKLLmrp7XyPBGkpkOyXSEUbzW90FhALrKhFZ9bjvqT4kC7SGbYMDDpDN2HgMDrQOY04p+ViG5UBAgxJ3c3MOqcpvo7fxV8gvCteGtlSmowWUcEg0NuM6O0mTLmlhBqCrP6ykrWvbSEWVQl4I3z81Dqu/tkYHEkHvobobAYM6TaUuu5LAPqkns3fkT11tH7SQM14mP/Vwo72LxfNx3P5FQw5fzQxf5T6J1Z3bAuurkcYdXgu74cw6Chf+w0L336tY/vXcz7E6fYw4rKrSLm5FHXbcvQbX0T3+esQi8J7d4E7j4KiSdjyPWzQreyWqthsM2BztT+47V4Mn/0SrfRa3NdeTv2fn4Bop7+DKS8P35w5VK5bR+Hv78La8AtIiuej0Ot1FI1II6PIRTSiYnebei0IDgd3anpCW87goVichxftIjn6FBXeU9nVpwBApzOrRYX3HFHpZK/Xq//BD36Q7/V69Xq9XissLAy/9NJLu59//vnkiy++uH96enp02bJlW55++uld119/fXEkEhEADz30UOXw4cMPUqwjzl//+tfyG2+8sejxxx/PvOCCC1ocDkesp373339/3U033VQwYMCAUr1ez9NPP73LarVqP/rRjxp27dplHjZs2GBN00RycnL0o48+2g5w5ZVXeu+8886i8847r8VisWgAB+rflc8++8zx3nvvpfTv3z84aNCgUoCHH3648rrrrmt9+OGHq6+88sqSgoKC1Ozs7Mh7773XsX9OTs6wtrY2fTQaFXPmzEn66KOPtowePTp055135r777rvJoVBIl5GRMfzGG29s+POf/1y173kPRJ+WThZCXAT8H6AHntU07Q/7bL8R2FvdpQ24S9O0NQc77qGWTpYcHm2Lq2iZ3f13LKwGMn80qluVw4OhqApzds3hocUPEY6FsRlsPDr5UabkTTnovsENjTS9vqldAOjwXNMf2/C0/YqS8k1NzP7L6oT2K386iuwu+fh7Qg23+xS82+5ToIv7FDgn5aAzJ+pn77w1hDZqfOYsY+uObd22lZSUcNNNN+FfXkPz21u772jQkfnT0RiSLHzw2O/Ysrx7hsCcAYO56gf3Y7Jq8PQk8O1Tt+X8X8PZPyQWU9m2vJYvXt1MTFExmvVccPsQCofGfQpin/6aKtNUVq+yoKh6SgcKDK8/TmTFMhxTpqBPTaX17bcByLrveyQNNsKY74Dp2If8hdp8fD373yyf/Q5oGs6UNC6/7xdkFJUc87GcqhyN0sl9EX1wLPD5fDq73a7qdDqeeeYZzxtvvJH8+eefJ9ykTxeOS+lkIYQe+BtwPlABLBdCzNY0bUOXbjuBKZqmNQshLgaeAcb31Zgkh4auh4yFhlQLGA9t1cmgM3Bx0cWUppTSGGwkzZZGvrNnD/l9sQxOJv0Ho+JJjxwmDKkHTlFtsSWmVNbpBGbrwX/qOrMB66BkDN8dgtISQe80Ysp27NePQu80E2vzkpaTwla6i4KMjHiiqL2pnLtiSDZ3JFNKycmF5d23p6RnoLNYwCjAU5IoCpzxDI96vY7+4zJJL3AR8EWwJ5lJ6lLQqCblW8x+tgq0uD9W1Va46PsPkb5lPv5FizsEAYA+dyCMOz9eu+E4YHE4OevqmQw6awqRYAB3RhbOHnI7SI4vubk3Np0MImBfFi1aZPvhD3+Yr2kaLpcr9uKLL+463mM6UenL5YNxwDZN03YACCFeBy4HOkSBpmmLu/RfSk/1eyXHDVOes3tool7gvqgI/WEUrdEJHUXuIorciY6CB0IIES+61MtwyKQMK6MvLGBll5wF4y8vIimjdxkYhUGHKceJKbG0QgKGDAemwiADkgpZYy/D74+HCNtsNoYPj6cVNmY7MBW6iOxq92vQCZJmlHREcvQfcyZrv5yLvyWemMhidzBk0jQMe83m5/4CXrkiXqEQIGMo5I7rGIOurQZP89d4GraCNhTsY8AetxRs26Qm5A0oW9bMedNGUffHxzrarGPHYhkx+qgKAn9rmJodrTRV+0nJdpBZ7I4vazRshT1LIdgEuWMhezQY41Yng9lMeqGslCg5+lx00UVtmzdv3nDwnpK+FAU5QHmX9xUc2ApwK/Dx/jYKIb4HfA8gfz9x2JKji8FjIeWmwUSr/ajhGMZ0G8asEzubnNFs4IwL8skdnIy/JYwz2UJKrgP9YURNHAxzfi5CZ8RSH+KmS66lIexFGPVkZmWSkhJ/yjUkWUieOYhoVRtqMIYx3YoxqzOsLq3/QK79r4ep37kNTYmRmpdP6sAuyYHyJ8Bt86B+ExhtkDmsY92fYAt88jPY8H5n/7N+ANN+AQYzBkuiU7jBrMcypJTCN98ksmMHOpcTy+DBGNMT1/QPl0hQYen7O9i0uDOEfMikbM6+0I7x5cvB22UJ+obXYeDFR+3cEonkyOhLUdCTjbdHBwYhxDnERcHE/R1M07RniC8vMGbMGJnW7BhhSLIkOACe6JhtRnIHeo7JuUy5GZhywQ5k7qePwWXGcAAfjJSiYlKKilEaGghv3UrbV19hLi7BlJ8XLyaUOTT+ty/1m7sJgljBRYS9KcQ+n4uxZCAlZ6Sx9otyVKX9v4uAEdPy0RsMWIeUYh1SevgTPwAttYFuggBg/YIqhozIJs27j0/a3F9B3plgS/y+tGiU8PbtRCsq0aemYu7fD739xBalEsnJTl+KggqgS2k9coEEL0ghxHDgWeBiTdMSM9NIJKcBkcpKqu67n+A33wCgT0oi//nnsJQe4MYd7YySUIoupW6Nh9aPngZAmEzkPvccV/1kFDtW16NEVUrOSCejS3XCviIa7dGxGyXSQ5h1oBFiPUe0+ebNo/LHP4FY/Hgpd95JyvduR2/rfTEuiURyaPRlnoLlQH8hRJEQwgRcD8zu2kEIkQ+8A9ykadqWHo5xSlMXqGNDwwaq/dUH79xHqNEYkRo/kRo/aqTni7mk7/A1NlCzYyve8j2Et3U6K8ZaWmh49jnULiGEraFWNjZuZLd3dzw1ckp/cMbtE2HrOFo/+ryjrxaJUPPAA6S4Y0y4sh+TvjWA7P5JR7aMooShbhPUroeIH0I+qCmLWyyUznEmpdlwp3W3Lnmy7LjT7aDbJ6xx/J3gTKzeGamspPqXD3YIAoDGp54ivHVrQt/9EQ5GaSj30VzjR+2pjoJEIkmgzywFmqYpQoh7gDnEQxKf1zRtvRDizvbtTwEPAinA39sTnSj7hsycqqyoWcHPFvyMukAdHrOH3038HRNzJh5SwpcjRWkN4527m8CKeM4M2xlpuC4oPOmWC05W9qxbw0dPPIa/pRmbO4lp9/8Y/Z8fJ9YUd+4OrV2LGgigc7vZ2ryVny/4ORubN2LRW/jpmJ8yo2QGtplvwYI/oQQSV9Si5eXEvD4MyfvN1dJ7fLWw8C+w/GlQYzDwEiicDHN+BjpD3Jdhwt1gT8WeZObiO4fzzZzdVGxuJr80mZHn52PLsMC334F5vwNfNYy9DYZf1+PpYq2tqN7EVMuxxt4ZE5tr/Hz5z81UbW1BZxCMnV7I0Km58egUyQnFrbfemldQUBB+8MEH6wAmTpzYPycnJ7K3+uHtt9+em5OTE903uQ/A448/nnLZZZd5CwsLo/tuOxg//vGPsx0OR+zXv/51wnFPZ/o0o6GmaR9pmjZA07QSTdN+1972VLsgQNO02zRN82iaNrL977QQBFVtVfz4yx9TF4hncWsON/OT+T9ht/eoVADtNaEtzQSW18Y9PTQIfFNPaONJF210UtJSW8Ps/32kI+og0NrCpx/+G+MNnTdJ5/SL0btcBKIBHlvxGBubNwIQioX47bLfsqlpE2QNh6uewTg80R3HNn48htSjFNa3exEs+3tcEABs/g80b4uHR+oMxBprUOraLV5KlBRnG9NuKOa6n49l6o2DSMl2gN4AxVPhpnfhe1/CxP8HrqweT2fMyMC4r0OxwYAx9+ABSrGYyuq55VRtbQFAVTSWzd5J3c5EkXE0iUQieL1eFOWYZKM9ZTj77LPbli5d6gCIxWI0NzcbNm/e3BFutHz5csfkyZMT6hwAvPrqq6l79uyRSu8oItMcHwdq/bU0h5u7tQWV4DFfRghtSHzqCqyt32/+b8nRw9fYQNjfvcppJBgkbI1baRznTsNzzTUIIWgKNbG4anHCMcrb2oN7DGYsQ4aR9fvfo3PEIxvMQ0rJeOC/0TuOUgGhHV8mtu1ZijbgUoJT3qO+6rvU/SuAd95OlJXvwVNnoX/nZmxtGxKXLMyOjrDJ/WFISSHnT49hLCoE4j4WuX/5X8wlB09mFGqLsnNNfUJ7Q2WP95WjQlVVFW+++SZPPfUU//nPf2ho6HU+oJOKlyobkkcsKhuW9cXq0SMWlQ17qbLhiM1Q06ZNa1u5cqUDYOXKldaBAwcG7XZ7rL6+Xh8MBsX27dstAGPHjh04ZMiQwRMnTuy/e/du4wsvvOApKyuzzZo1q3jQoEGln3zyiWPQoEGlgwYNKh0wYECpEGI0wPr1682TJk3qP2TIkMGjR48euGrVqgRT6G9/+9v0kpKSIQMGDCi99NJLT+u4WJnm+DiQZEnCorcQinUWvNEJHcmWo2DmPQRMBa4Ey4C52H1MlzBOV2wuF+70DAxmC00V5Wiaik6vxzNxIu4p52DMyelwqHManQxIGsCWlu5uN2mWzgqwOrOZpCuvwDZmNKrfjyEzC0OS++gNOHskfPNS97b0wURSZ9A4WwEt/nTs/bQCMbEQp94IW+dA5Qq4/QvwFBzyKa3DhlH4z3+i1Naicydhyu7ZqrAvJquB9AIXu/epqOk+gtLfB6KpqYlXX32VQCDu+Llq1SoaGhqYOXMmVmvfnPN48FJlQ/KD2yoLwqqmA6iNKKYHt1UWANyck3rYJsbCwsKowWDQtm7dapo/f779zDPP9FdWVhrnzZvn8Hg8SklJSegnP/lJ3n/+859t2dnZyj/+8Q/PT3/605y33npr15NPPpn+2GOPlU+ePDkAsGnTpg0Ad9xxR+4555zjBbjtttsKnnnmmd3Dhg0Lz5s3z37XXXflL126tNt/pscffzxz9+7d66xWq9bQ0NB3+bxPAqQoOA4UuAr45Zm/5JeLf4mqqQgE9425j2L3sRWo1iEpBFbVodTGL2b6VAu2EUcvXl2yfxzCw5UX/Ixoc4DI+TG+XvYeJePPJLXfAPSG7v8t3RY3D5z5AHfNvYugEi+mdEW/KxicPDjhuKa8vIS2o0LxOZAzGipXxt+7ciBnDJF6C2jdowfa1mnYBl6Bfu1T8eiCxq2HJQoADMnJh+wTYTTpGTejiJqdrYT9cbFSMDSlzyIvGhsbOwTBXsrLy2lpaTmlRMGfd9Xk7BUEewmrmu7Pu2pyjkQUAIwePbrtiy++sC9ZssRx33331e7Zs8e0aNEiu9vtjmVmZkYWLlzonjZt2gCIFxFKS0vbrw/Bs88+61m7dq1twYIFW1pbW3WrVq1yXHvttR0mpr21E7oycODA4JVXXll02WWXtdx4440tRzKXkx0pCo4DOqHj4qKLGZA8gBp/DenWdIqTijHqj+3SmDHNRtqtQ4nWBtA0MKbbMBykaJDkyInW+Gl4Zl1HoSfdBpg243YqFQ3dvt757YzOGM2bl77Jbu9unCYn/ZL64TL3fXhhB8lFcP1r8SRKqgKpA0BT0cViQPfcAzqHQIS6VD00HvvcAukFLq79rzG01AQwmPUkZ9uxOo64wm+PmEyJx9Xr9RgMp9bltS6i9PgB7q/9UJgwYULb4sWLHZs2bbKOHTs2WFxcHPnLX/6S4XA4YlOnTvXV1NSYVq9evelgx1mxYoXlkUceyV64cOFmg8FALBbD6XQqey0I++OLL77Y+vHHHzvfe++9pEcffTR769atZcYeKqmeDkifguOEUW9kUPIgpuZNpTS1FIvh+Hj8611mLP09WAd4pCA4RoT3eLtVfgQIL66leXcA7wFKUxe6C5mSN4VRGaP6VBAEvBF2r29k05Jqqre3oOwNVXVmQPEU6HduPKuipwBT/0x07i73BAHuMyLotn8Qfz/oUkgb1GdjPRDuNBsFw1LJGeDpM0EAkJaWxsCBA7u1TZ48meSjEfVxApFuMvSYUGJ/7YfClClT2ubOnZuUlJQUMxgMZGRkxLxer37VqlWOW265pampqckwd+5cO0A4HBYrVqywADgcjlhra6seoLGxUT9z5sziF154YWd2drYCkJycrObm5kaef/55D8StDEuWLOlmvonFYmzfvt00Y8YM39///vcKn8+n33vM05FTS8pKJCcDPYXMxzR0Ao63j2ewLcqCN7ewbUXnk/45Nw2i9OzsHvsb02yk3TaMyB4faljBlGXFxCY49yFIKoTcMT1mKzyVsNlsXHLJJYwYMYKWlhbS09PJyclBrz+17is/Lsys7OpTAGDWCfXHhZlHVDoZYNy4ccGWlhbDVVdd1eEIMmjQoKDf79fn5OQor7/++vYf/OAH+T6fTx+LxcRdd91VO2bMmNCsWbMa7r333oL77rtPvfvuu2urqqrMd9xxR+HeY2zatGnDa6+9tuP2228v+J//+Z8sRVHElVde2TRhwoQO9a0oipg5c2aRz+fTa5om7rjjjtrU1NTTNmlLn5ZO7itk6WTJyUykso26J9eA0qkO9JNz2R2NMfbSYvT642fAq9zSzHt/XtWtzWTR861fjMNiM9BcE0BVNZIyrNich29ZCgeiNNcGUBWVpAwbtkMoxS05PI5G6eSXKhuS/7yrJqcuopjSTYbIjwszK4/Un0By7DkupZMlEknPGLPtpH1vGL7FVcRawjDQQ8htZmix+7gKAoBwIDHGPhKKEfZHWfjmVnatjd8/UnMdXHDbEDyZh+4v4GsOsfCNrexYHQ8bTM62c+HtQ0k+wYttSeJRBlIEnNpIUSCRHGOEEJjzXZjynKBqqHDcxcBekjJs6I06YtFOK0bOoCQaK9porGhj5Hl56PQ6dq1rYNOSas68ouSQQ1irtrR0CAKApio/6xdUMvGa/gidDIeVSI4nJ8aVSCI5DRFCIPS6E0YQAHgybVx6z3A8mTYQUDAshak3DCToj1AyKo31C6pY/fke0gtcCJ1AiRx6TYHaXYmZBcs3NhEJy0yAEsnxRloKJBJJB0IIcgcmc9V9o4gEY1hdJowmPVZnK0ve2dHRb9OSasbNKMJgOnRBk1nkYt0X3dsKhqRgMsvL0XFAVVVV6HS6k8+5THJYqKoq6NndGZCWAolE0gMWuwlXqhWjQUDVamq2JPqi7S3JfKhk909iwLjOyohpBU5KJ2bLpYPjQ1l9fb27/UYhOcVRVVXU19e7gbL99ZHSXCKR7J+aNfDPa/GUvJywKTXXcVilmB0eC1NuGMiIc/NQYxrudGuf5hGQ7B9FUW6rqal5tqamZijyIfF0QAXKFEW5bX8dpCiQSCQANNdUU7V5A77GerIHDCaz30BMFcvBX0+Bcwtrk0vwNcWzy5qsBoZNzUV3mE/3e+sTROsCRDY04vVHMRe6MeU6EcYT597UWLGHio3rCQf85A4aQkZJ/4Q01Cczo0ePrgMuO97jkJw4nDq/bolEcti01tfx3qO/pqmyvKPtwrv+H0M98aJKniU/4YqJD9GgG4qqt5JSOghPzpFlVYzWB6h/dh2qtzMhXsqsUqylR6nc8xHSUL6bNx/+b4K+uGOkEDqueuBhCoefcZxHJpH0HSeOJJdIJMeN+l07ugkCgK9efR5f8kiwJIGq4Pr6lxQvvZJ+2XV4co48S2Gk3NdNEAC0frITtYdcCceDio1lHYIAQNNUlrz1L6Kh0AH2kkhObqSlQCKREA2HE9rCgQCqOQlu+Q9s+QR8NTDoEsgbd1TOuW/9BwA1oKDGYuhOgEtT2O9PaAv6WokpCqdnqRzJ6cDx/58nkUiOO6l5+RiMJpRo55P7iPMvxpmSCvoMyBx61M9pzHHEbZVdAhgcE3MwHEH65AOh+MKorRF0VgOGlIOXNM4dPARE94IUo6ZfjsXh6JPxSSQnArL2gUQiQdM0qrZsZPFbr9FcXcHQqeczdOp5uNLS++6cMY3wzha8n+0m5ovimJCFdUQahj6ogxAu99H02kZiTWGEWU/Slf2wDU1FHCB6QolGKV+/liVv/Yugz8uoSy5n4JkTsbmTjvr4jgU91T6QSPZFigKJRNJBNBwiGg5jc7kP2E9TVNRAFGExoDMdWTVANRxDU1T09qNnlFfDClo4RsyoobSF8L60FaWhiy+AgPR7z8CUffCn/nAwgKrEsDqdR218xwMpCiS9QS4fSCSSDoxmC0az5YB9IrV+fPP2ENrSgrnAhev8Akw5h29S15n1YD46ZYY1TSOyy0vLxzuJNQTRD3Sg62/tLggANFCaQr0SBWar7aiMTSI5GZCiQCKR9JpYW5Sm1zah1AQACG1qIlLpI/3ukRiSDiwmjgVKbYD659aBEreAqqtaMEUF5gFJhLe0dOurd8qESRLJvsiQRIlE0muUpmCHINiL6ouiNASP04i6E60NdAiCvUTWN2Mamwz6zkRLznPzMGZKC4BEsi/SUiCRSHqNMOlAANq+7UfH/H+kiB6WIXRWA9s3L8c83k5+4VDMHjuGdBs6WYBJIklAWgokEkmvMabacEzJ7dZmG5WOIf3In7qbqirZvHgBGxfNp7Fiz2Edw5Rlx5i3j0PgWBvLPv03frUF+9B0THkuKQgkkv0gow8kEskhEQtEiZT7UOqDGJItmPKcR7w+37BnF2/+5ucEva0AmKxWrn3wETKL+x/ysZTmUHx8rSFUF9S07sSa5CSjZAAWu/2IxnkyI6MPJL1BymWJRHJI6G1GrAOTYeDRO+bW5Us7BAFAJBhk7dxPyLi9H0IcWtElg8eCwdPp9JhE3lEbp0RyqiOXDyQSyXGnuboyoa2xohxVVXvoLZFI+gopCiQSyXFnwPizEtqGTbsAvf7EcGCUSE4XpCiQSCTHndzBQ5n23Tux2B0YLVYmXj+LopFy+VsiOdZInwKJ5CRAU1Q0VTvilMInKhaHkzMuvJT+YyegaRqO5JRD9iWQSCRHTp9aCoQQFwkhNgshtgkh/quH7UII8Xj79rVCiFF9OR6J5GRD0zTCu1tpfG0T9U+toW1ZNTFf5OA7nqQ4klNwpqRKQSCRHCf6zFIghNADfwPOByqA5UKI2ZqmbejS7WKgf/vfeODJ9n8lEgkQrfZT/4/OtL0t725DC8Vw7pMrQCKRSI4GfWkpGAds0zRth6ZpEeB14PJ9+lwOvKzFWQokCSGy+nBMEslJRbSqLSFtr++rCmLe8HEakUQiOZXpS1GQA5R3eV/R3naofQAQQnxPCLFCCLGivr7+qA5UIjlREYbE/6LCqAOdNK9LJJKjT1+Kgp6uWvumT+xNn3ijpj2jadoYTdPGpKWlHfHgJJKTAWOOA53d2K3NdWEBeoes8CeRSI4+fRl9UAHdUonlAlWH0UciOW0xptlIu30YwS3NqC0hLINSMBU4D76jRCKRHAZ9KQqWA/2FEEVAJXA9MHOfPrOBe4QQrxN3MGzVNK26D8ckkZx0GDPtGDNP35z9Eonk2NFnokDTNEUIcQ8wB9ADz2uatl4IcWf79qeAj4DpwDYgAHynr8YjkUgkEonkwPRp8iJN0z4ifuPv2vZUl9cacHdfjkEikUgkEknvkGmOJRKJRCKRAFIUSCQSiUQiaUeKAolEIpFIJIAUBRKJRCKRSNoRcV+/kwshRD2w+zB3TwUajuJwTgbknE99Trf5gpzzoVKgaZrM/CY5ICelKDgShBArNE07rQq1yzmf+pxu8wU5Z4mkL5DLBxKJRCKRSAApCiQSiUQikbRzOoqCZ473AI4Dcs6nPqfbfEHOWSI56px2PgUSiUQikUh65nS0FEgkEolEIukBKQokEolEIpEAp6goEEJcJITYLITYJoT4rx62CyHE4+3b1wohRh2PcR5NejHnG9vnulYIsVgIMeJ4jPNocrA5d+k3VggRE0JccyzH1xf0Zs5CiKlCiNVCiPVCiPnHeoxHm178tt1CiA+EEGva53xSV1sVQjwvhKgTQpTtZ/spd/2SnEBomnZK/REv07wdKAZMwBqgdJ8+04GPAQGcCSw73uM+BnM+C/C0v774dJhzl37ziFfrvOZ4j/sYfM9JwAYgv/19+vEe9zGY8wPA/7S/TgOaANPxHvsRzHkyMAoo28/2U+r6Jf9OrL9T0VIwDtimadoOTdMiwOvA5fv0uRx4WYuzFEgSQmQd64EeRQ46Z03TFmua1tz+dimQe4zHeLTpzfcMcC/wNlB3LAfXR/RmzjOBdzRN2wOgadrJPu/ezFkDnEIIATiIiwLl2A7z6KFp2lfE57A/TrXrl+QE4lQUBTlAeZf3Fe1th9rnZOJQ53Mr8SeNk5mDzlkIkQNcCTx1DMfVl/Tmex4AeIQQXwohVgohZh2z0fUNvZnzE8BgoApYB/xQ0zT12AzvuHCqXb8kJxCG4z2APkD00LZv3GVv+pxM9Ho+QohziIuCiX06or6nN3P+C/AzTdNi8YfIk57ezNkAjAbOBazAEiHEUk3TtvT14PqI3sz5QmA1MA0oAT4TQizQNM3bx2M7Xpxq1y/JCcSpKAoqgLwu73OJP0Ecap+TiV7NRwgxHHgWuFjTtMZjNLa+ojdzHgO83i4IUoHpQghF07T3jskIjz69/W03aJrmB/xCiK+AEcDJKgp6M+fvAH/QNE0DtgkhdgKDgK+PzRCPOafa9UtyAnEqLh8sB/oLIYqEECbgemD2Pn1mA7PavXjPBFo1Tas+1gM9ihx0zkKIfOAd4KaT+KmxKweds6ZpRZqmFWqaVgj8G/j+SSwIoHe/7feBSUIIgxDCBowHNh7jcR5NejPnPcQtIwghMoCBwI5jOspjy6l2/ZKcQJxylgJN0xQhxD3AHOKey89rmrZeCHFn+/aniHuiTwe2AQHiTxonLb2c84NACvD39idnRTuJq631cs6nFL2Zs6ZpG4UQnwBrARV4VtO0HkPbTgZ6+T3/BnhRCLGOuGn9Z5qmnbQllYUQrwFTgVQhRAXwEGCEU/P6JTmxkGmOJRKJRCKRAKfm8oFEIpFIJJLDQIoCiUQikUgkgBQFEolEIpFI2pGiQCKRSCQSCSBFgUQikUgkknakKJBIuiCE0IQQr3R5bxBC1AshPjyK5/hICJF0tI4nkUgkRwspCiSS7viBoUIIa/v784HKQzmAEOKA+T80TZuuaVrL4Q1PIpFI+g4pCiSSRD4GLml/fQPw2t4NQohxQojFQohV7f8ObG+/RQjxlhDiA+BTIYRNCPFme737N4QQy4QQY9r77hJCpAohCoUQG4UQ/xBCrBdCfNpFjEgkEskxR4oCiSSR14HrhRAWYDiwrMu2TcBkTdPOIJ4l8pEu2yYAN2uaNg34PtCsadpw4hn3Ru/nXP2Bv2maNgRoAa4+mhORSCSSQ+GUS3MskRwpmqatFUIUErcSfLTPZjfwkhCiP/HKdMYu2z7TNK2p/fVE4P/aj1cmhFi7n9Pt1DRtdfvrlUDhEU9AIpFIDhNpKZBIemY28Bhdlg7a+Q3whaZpQ4EZgKXLNn+X172t1Rzu8jqGFOoSieQ4IkWBRNIzzwO/1jRt3T7tbjodD285wP4LgW8BCCFKgWFHe4ASiURytJGiQCLpAU3TKjRN+78eNj0K/F4IsYh41b798XcgrX3Z4GfEqxa2Hv2RSiQSydFDVkmUSPoAIYQeMGqaFhJClACfAwM0TYsc56FJJBLJfpHrlxJJ32ADvhBCGIn7F9wlBYFEIjnRkZYCiUQikUgkgPQpkEgkEolE0o4UBRKJRCKRSAApCiQSiUQikbQjRYFEIpFIJBJAigKJRCKRSCTt/H9pfwpdBT0PPAAAAABJRU5ErkJggg==\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "sns.scatterplot(data=clustering, x=\"Margin\", y=\"count_zeros\", hue=\"Study\")\n", "plt.legend(loc='center left', bbox_to_anchor=(1.0, 0.5))" ] } ], "metadata": { "celltoolbar": "Edit Metadata", "kernelspec": { "display_name": "Python 3", "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": "3.8.8" } }, "nbformat": 4, "nbformat_minor": 5 }