{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "

ipyrad-analysis toolkit: Dimensionality reduction

\n", "\n", "The `pca` tool can be used to implement a number of dimensionality reduction methods on SNP data (PCA, t-SNE, UMAP) and to filter and/or impute missing data in genotype matrices to reduce the effects of missing data. " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Load libraries" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "# conda install ipyrad -c conda-forge -c bioconda\n", "# conda install ipcoal -c conda-forge\n", "# conda install scikit-learn -c conda-forge" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "import ipyrad.analysis as ipa\n", "import toytree\n", "import toyplot\n", "import ipcoal" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "0.9.61\n", "2.0.4\n", "0.19.0\n", "0.1.5\n" ] } ], "source": [ "print(ipa.__version__)\n", "print(toytree.__version__)\n", "print(toyplot.__version__)\n", "print(ipcoal.__version__)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Simulate linked SNP data from this tree and write to a database (HDF5)\n", "Here we simulate data on a tree/network that includes an introgressive edge. " ] }, { "cell_type": "code", "execution_count": 29, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
01234567891011121314r0r1r2r3r4r5r6r70250000500000
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# make a random tree with root height in generations\n", "tree = toytree.rtree.unittree(ntips=8, treeheight=5e5, seed=123)\n", "tree.draw(ts='p', admixture_edges=(3, 4));" ] }, { "cell_type": "code", "execution_count": 30, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "wrote 11283 SNPs to /tmp/test-pca.snps.hdf5\n", "wrote 10033 SNPs to /tmp/test-pca-cov75.snps.hdf5\n" ] } ], "source": [ "# create a coalescent simulator w/ 8 samples per species and high Ne\n", "model = ipcoal.Model(tree, Ne=2e5, nsamples=6, admixture_edges=(3, 4, 0.5, 0.15))\n", "\n", "# simulate many short loci on coalescent genealogies (a few minutes)\n", "model.sim_loci(nloci=2000, nsites=50)\n", "\n", "# write simulated SNPs to a database file (.snps.hdf5)\n", "model.write_snps_to_hdf5(name='test-pca', outdir=\"/tmp\", diploid=True)\n", "\n", "# now randomly drop data from many loci (RAD-seq like)\n", "model.apply_missing_mask(coverage=0.75)\n", "\n", "# write a data file with missingness in it (.snps.hdf5)\n", "model.write_snps_to_hdf5(name='test-pca-cov75', outdir=\"/tmp\", diploid=True)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### The input data" ] }, { "cell_type": "code", "execution_count": 31, "metadata": {}, "outputs": [], "source": [ "# the simulated SNP database file\n", "SNPS = \"/tmp/test-pca.snps.hdf5\"" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Make an IMAP dictionary (map popnames to list of samplenames)" ] }, { "cell_type": "code", "execution_count": 34, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "{'r0': ['r0-0', 'r0-1', 'r0-2'],\n", " 'r1': ['r1-0', 'r1-1', 'r1-2'],\n", " 'r2': ['r2-0', 'r2-1', 'r2-2'],\n", " 'r3': ['r3-0', 'r3-1', 'r3-2'],\n", " 'r4': ['r4-0', 'r4-1', 'r4-2'],\n", " 'r5': ['r5-0', 'r5-1', 'r5-2'],\n", " 'r6': ['r6-0', 'r6-1', 'r6-2'],\n", " 'r7': ['r7-0', 'r7-1', 'r7-2']}" ] }, "execution_count": 34, "metadata": {}, "output_type": "execute_result" } ], "source": [ "from itertools import groupby\n", "\n", "# load sample names from SNPs file\n", "tool = ipa.snps_extracter(SNPS)\n", "\n", "# group names by prefix before '-'\n", "groups = groupby(tool.names, key=lambda x: x.split(\"-\")[0])\n", "\n", "# arrange into a dictionary\n", "IMAP = {i[0]: list(i[1]) for i in groups}\n", "\n", "# show the imap dict\n", "IMAP" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Run PCA\n", "Unlinked SNPs are automatically sampled from each locus. By setting `nreplicates=30` the subsampling procedure is repeated 30X to show variation over the subsampled SNPs. The imap dictionary is used in the `.draw()` function to color points. " ] }, { "cell_type": "code", "execution_count": 57, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Samples: 24\n", "Sites before filtering: 11283\n", "Filtered (indels): 0\n", "Filtered (bi-allel): 426\n", "Filtered (mincov): 0\n", "Filtered (minmap): 0\n", "Filtered (subsample invariant): 0\n", "Filtered (minor allele frequency): 4557\n", "Filtered (combined): 4983\n", "Sites after filtering: 6300\n", "Sites containing missing values: 0 (0.00%)\n", "Missing values in SNP matrix: 0 (0.00%)\n" ] } ], "source": [ "# run on dataset with no missing data and with minor allele freq filter\n", "tool = ipa.pca(data=SNPS, minmaf=0.03)" ] }, { "cell_type": "code", "execution_count": 58, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Subsampling SNPs: 1918/6300\n" ] }, { "data": { "text/html": [ "
r0-0r0-1r0-2r1-0r1-1r1-2r2-0r2-1r2-2r3-0r3-1r3-2r4-0r4-1r4-2r5-0r5-1r5-2r6-0r6-1r6-2r7-0r7-1r7-2-1001020PC0 (26.8%) explained-1001020PC1 (13.2%) explainedr0r1r2r3r4r5r6r7
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# run to subsample unlinked SNPs and run PCA\n", "tool.run(nreplicates=20)\n", "tool.draw(imap=IMAP);" ] }, { "cell_type": "code", "execution_count": 59, "metadata": { "scrolled": true }, "outputs": [ { "data": { "text/html": [ "
r0-0r0-1r0-2r1-0r1-1r1-2r2-0r2-1r2-2r3-0r3-1r3-2r4-0r4-1r4-2r5-0r5-1r5-2r6-0r6-1r6-2r7-0r7-1r7-2-1001020PC0 (26.8%) explained-1001020PC1 (13.2%) explainedr0-0r0-1r0-2r1-0r1-1r1-2r2-0r2-1r2-2r3-0r3-1r3-2r4-0r4-1r4-2r5-0r5-1r5-2r6-0r6-1r6-2r7-0r7-1r7-2-1001020PC0 (26.8%) explained-1001020PC2 (12.3%) explainedr0-0r0-1r0-2r1-0r1-1r1-2r2-0r2-1r2-2r3-0r3-1r3-2r4-0r4-1r4-2r5-0r5-1r5-2r6-0r6-1r6-2r7-0r7-1r7-2-1001020PC1 (13.2%) explained-1001020PC2 (12.3%) explainedr0r1r2r3r4r5r6r7
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# a convenience function for plotting across three axes\n", "tool.draw_panels(0, 1, 2, imap=IMAP);" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Run TSNE\n", "t-SNE is a manifold learning algorithm that can sometimes better project data into a 2-dimensional plane. The distances between points in this space are harder to interpret. You can see in this example that the gene flow between r3 and r4 is not readily apparent." ] }, { "cell_type": "code", "execution_count": 87, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Subsampling SNPs: 1918/6300\n" ] }, { "data": { "text/html": [ "
r0-0r0-1r0-2r1-0r1-1r1-2r2-0r2-1r2-2r3-0r3-1r3-2r4-0r4-1r4-2r5-0r5-1r5-2r6-0r6-1r6-2r7-0r7-1r7-2-200-1000100TSNE component 1-200-1000100TSNE component 2r0r1r2r3r4r5r6r7
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# run tsne on same dataset (try several perplexity values)\n", "tool.run_tsne(perplexity=4, seed=333)\n", "tool.draw(imap=IMAP, size=8);" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Run UMAP\n", "UMAP is similar to t-SNE but the distances between clusters are more representative of the differences betwen groups. This requires another package that if it is not yet installed it will ask you to install. " ] }, { "cell_type": "code", "execution_count": 86, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
r0-0r0-1r0-2r1-0r1-1r1-2r2-0r2-1r2-2r3-0r3-1r3-2r4-0r4-1r4-2r5-0r5-1r5-2r6-0r6-1r6-2r7-0r7-1r7-218202224UMAP component 123456UMAP component 2r0r1r2r3r4r5r6r7
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# try several n_neighbors values\n", "tool.run_umap(subsample=False, n_neighbors=13, seed=222)\n", "tool.draw(imap=IMAP, size=8);" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Dealing with missing data (filtering and imputation)\n", "Missing data has large effects on dimensionality reduction methods, and it is best to (1) minimize the amount of missing data in your input data set by using filtering, and (2) impute missing data values. In the example below we use the simulated datset with 25% missing data and use filtering and imputation methods in the ipyrad-analysis tool. This allows us to make inferences from large SNP datasets despite missing data. " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Below the `pca` tool uses `impute_method=sample` to sample alleles for missing values based on the frequency of alleles at that SNP in the population (defined by the IMAP dictionary). It also drops any sites that do not have data across at least 50% of samples (mincov=0.5). " ] }, { "cell_type": "code", "execution_count": 92, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Samples: 24\n", "Sites before filtering: 10033\n", "Filtered (indels): 0\n", "Filtered (bi-allel): 303\n", "Filtered (mincov): 0\n", "Filtered (minmap): 7\n", "Filtered (subsample invariant): 0\n", "Filtered (minor allele frequency): 3151\n", "Filtered (combined): 3458\n", "Sites after filtering: 6575\n", "Sites containing missing values: 5116 (77.81%)\n", "Missing values in SNP matrix: 9513 (6.03%)\n", "Imputation: 'sampled'; (0, 1, 2) = 81.9%, 5.9%, 12.3%\n" ] } ], "source": [ "# init a tool with the missing data dataset\n", "tool2 = ipa.pca(\n", " data=\"/tmp/test-pca-cov75.snps.hdf5\",\n", " imap=IMAP, \n", " minmap={i: 1 for i in IMAP}, \n", " mincov=0.5,\n", " minmaf=0.03,\n", " impute_method=\"sample\", \n", ")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The results below are nearly identical to the results above where there was 0% missing data, showing that filtering and imputation worked properly. " ] }, { "cell_type": "code", "execution_count": 93, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Subsampling SNPs: 1918/6300\n" ] }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAA+gAAAEsCAYAAABQRZlvAAAJMmlDQ1BkZWZhdWx0X3JnYi5pY2MAAEiJlZVnUJNZF8fv8zzphUASQodQQ5EqJYCUEFoo0quoQOidUEVsiLgCK4qINEWQRQEXXJUia0UUC4uCAhZ0gywCyrpxFVFBWXDfGZ33HT+8/5l7z2/+c+bec8/5cAEgiINlwct7YlK6wNvJjhkYFMwE3yiMn5bC8fR0A9/VuxEArcR7ut/P+a4IEZFp/OW4uLxy+SmCdACg7GXWzEpPWeGjy0wPj//CZ1dYsFzgMt9Y4eh/eexLzr8s+pLj681dfhUKABwp+hsO/4b/c++KVDiC9NioyGymT3JUelaYIJKZttIJHpfL9BQkR8UmRH5T8P+V/B2lR2anr0RucsomQWx0TDrzfw41MjA0BF9n8cbrS48hRv9/z2dFX73kegDYcwAg+7564ZUAdO4CQPrRV09tua+UfAA67vAzBJn/eqiVDQ0IgALoQAYoAlWgCXSBETADlsAWOAAX4AF8QRDYAPggBiQCAcgCuWAHKABFYB84CKpALWgATaAVnAad4Dy4Aq6D2+AuGAaPgRBMgpdABN6BBQiCsBAZokEykBKkDulARhAbsoYcIDfIGwqCQqFoKAnKgHKhnVARVApVQXVQE/QLdA66At2EBqGH0Dg0A/0NfYQRmATTYQVYA9aH2TAHdoV94fVwNJwK58D58F64Aq6HT8Id8BX4NjwMC+GX8BwCECLCQJQRXYSNcBEPJBiJQgTIVqQQKUfqkVakG+lD7iFCZBb5gMKgaCgmShdliXJG+aH4qFTUVlQxqgp1AtWB6kXdQ42jRKjPaDJaHq2DtkDz0IHoaHQWugBdjm5Et6OvoYfRk+h3GAyGgWFhzDDOmCBMHGYzphhzGNOGuYwZxExg5rBYrAxWB2uF9cCGYdOxBdhK7EnsJewQdhL7HkfEKeGMcI64YFwSLg9XjmvGXcQN4aZwC3hxvDreAu+Bj8BvwpfgG/Dd+Dv4SfwCQYLAIlgRfAlxhB2ECkIr4RphjPCGSCSqEM2JXsRY4nZiBfEU8QZxnPiBRCVpk7ikEFIGaS/pOOky6SHpDZlM1iDbkoPJ6eS95CbyVfJT8nsxmpieGE8sQmybWLVYh9iQ2CsKnqJO4VA2UHIo5ZQzlDuUWXG8uIY4VzxMfKt4tfg58VHxOQmahKGEh0SiRLFEs8RNiWkqlqpBdaBGUPOpx6hXqRM0hKZK49L4tJ20Bto12iQdQ2fRefQ4ehH9Z/oAXSRJlTSW9JfMlqyWvCApZCAMDQaPkcAoYZxmjDA+SilIcaQipfZItUoNSc1Ly0nbSkdKF0q3SQ9Lf5RhyjjIxMvsl+mUeSKLktWW9ZLNkj0ie012Vo4uZynHlyuUOy33SB6W15b3lt8sf0y+X35OQVHBSSFFoVLhqsKsIkPRVjFOsUzxouKMEk3JWilWqUzpktILpiSTw0xgVjB7mSJleWVn5QzlOuUB5QUVloqfSp5Km8oTVYIqWzVKtUy1R1WkpqTmrpar1qL2SB2vzlaPUT+k3qc+r8HSCNDYrdGpMc2SZvFYOawW1pgmWdNGM1WzXvO+FkaLrRWvdVjrrjasbaIdo12tfUcH1jHVidU5rDO4Cr3KfFXSqvpVo7okXY5upm6L7rgeQ89NL0+vU++Vvpp+sP5+/T79zwYmBgkGDQaPDamGLoZ5ht2GfxtpG/GNqo3uryavdly9bXXX6tfGOsaRxkeMH5jQTNxNdpv0mHwyNTMVmLaazpipmYWa1ZiNsulsT3Yx+4Y52tzOfJv5efMPFqYW6RanLf6y1LWMt2y2nF7DWhO5pmHNhJWKVZhVnZXQmmkdan3UWmijbBNmU2/zzFbVNsK20XaKo8WJ45zkvLIzsBPYtdvNcy24W7iX7RF7J/tC+wEHqoOfQ5XDU0cVx2jHFkeRk4nTZqfLzmhnV+f9zqM8BR6f18QTuZi5bHHpdSW5+rhWuT5z03YTuHW7w+4u7gfcx9aqr01a2+kBPHgeBzyeeLI8Uz1/9cJ4eXpVez33NvTO9e7zofls9Gn2eedr51vi+9hP0y/Dr8ef4h/i3+Q/H2AfUBogDNQP3BJ4O0g2KDaoKxgb7B/cGDy3zmHdwXWTISYhBSEj61nrs9ff3CC7IWHDhY2UjWEbz4SiQwNCm0MXwzzC6sPmwnnhNeEiPpd/iP8ywjaiLGIm0iqyNHIqyiqqNGo62ir6QPRMjE1MecxsLDe2KvZ1nHNcbdx8vEf88filhICEtkRcYmjiuSRqUnxSb7JicnbyYIpOSkGKMNUi9WCqSOAqaEyD0tandaXTlz/F/gzNjF0Z45nWmdWZ77P8s85kS2QnZfdv0t60Z9NUjmPOT5tRm/mbe3KVc3fkjm/hbKnbCm0N39qzTXVb/rbJ7U7bT+wg7Ijf8VueQV5p3tudATu78xXyt+dP7HLa1VIgViAoGN1tubv2B9QPsT8M7Fm9p3LP58KIwltFBkXlRYvF/OJbPxr+WPHj0t6ovQMlpiVH9mH2Je0b2W+z/0SpRGlO6cQB9wMdZcyywrK3BzcevFluXF57iHAo45Cwwq2iq1Ktcl/lYlVM1XC1XXVbjXzNnpr5wxGHh47YHmmtVagtqv14NPbogzqnuo56jfryY5hjmceeN/g39P3E/qmpUbaxqPHT8aTjwhPeJ3qbzJqamuWbS1rgloyWmZMhJ+/+bP9zV6tua10bo63oFDiVcerFL6G/jJx2Pd1zhn2m9az62Zp2WnthB9SxqUPUGdMp7ArqGjzncq6n27K7/Ve9X4+fVz5ffUHyQslFwsX8i0uXci7NXU65PHsl+spEz8aex1cDr97v9eoduOZ67cZ1x+tX+zh9l25Y3Th/0+LmuVvsW523TW939Jv0t/9m8lv7gOlAxx2zO113ze92D64ZvDhkM3Tlnv296/d5928Prx0eHPEbeTAaMip8EPFg+mHCw9ePMh8tPN4+hh4rfCL+pPyp/NP637V+bxOaCi+M24/3P/N59niCP/Hyj7Q/Fifzn5Ofl08pTTVNG02fn3Gcufti3YvJlykvF2YL/pT4s+aV5quzf9n+1S8KFE2+Frxe+rv4jcyb42+N3/bMec49fZf4bmG+8L3M+xMf2B/6PgZ8nFrIWsQuVnzS+tT92fXz2FLi0tI/QiyQvpTNDAsAAAAGYktHRAD/AP8A/6C9p5MAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAAddEVYdFNvZnR3YXJlAEdQTCBHaG9zdHNjcmlwdCA5LjUw/rJdRQAAIABJREFUeJzs3XmcHGWdP/BPVXd1ddf0MT3dPWdmhpCEZJQrQQGJAZRLWTEmrPqLHF6rqy6suCoo6IqusCseK8IqgqgQ0PUiApqVa0VIBMMmgQTNfcz03H1f1V1T3VW/P6o7GcIk6WTOnnzer1e/pruquvpbk8m361vPU88DEBERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERzUrCVH+gJEmyJNn9AHStoMVLhmFOdQxEREREREREM82UFehNjcF/9Pm8/2QTxTcCEMuLswVNe3poOHqTquZ3TlUsRERERERERDONfSo+JBRs+JDP571+aDh6g1bQXipoWsomiqLslNs97rp3d7a3Pbl7b/dpuq5npiIeIiI6ukXw25YLc+sWol6GdWHV2IxI4VFzX06BBADoQcZQUZzWOImIJloH3IICSexBpqSiiCCc4iphgbIYIScAGwCzB1ntv81due1IMAkS0YSZkhb0zva2XybTmZ+mUum1h1n/80gs/gNVzT83FfEQEdGRLUWL/aNClx8HezxVOAF4AZTKr40eZNOPmvsimxHRpzRIIqIJEIRTWC7MVRYj5FRgdwJwATABqABKO5AcaYdbUmA/NB8CgHm/uS25HgMjUxo0Ec1atqn4EI/H/WbZ4XhDLqf+wTTN19xz7pTlxkDA/8V0JnuvrhfjUxEPEREdngK78C/CmQ0SRBcABYAbVmHeAKC5vMwBYASA6IPDdY7Q5AsKTnUzomxJIqKa0QG3eIvwpoZ58DkliAoAH6wephKsvKcH4fRLEN2wLkxKKLeglx+uJUKo4RKhXThNCDgU2LEX6dJhPo6I6KimpAXdKcuBjo6239tEsVVV8xsA5ABAtIltTlk+K5XK3N4/OPTNqYiFiIgOWooW6RKhXemA2wkAPchqeRQdC1Ef0kWzYUge6UpIxUBwRIo2FxyR8peGCKAAIFn+WWk5Uu8wN/ewuycR1Yo7hPPqg3DKsIruBgCIOYqNUYd+it0UhCZN6ncXbTkAMqxcFym/1VZ+5GEV6tnyc/Qgq91hbkqpKHIgZCI6ZlM2SJxNFAVFcb3J6ZSXAAgB0AD0ZDK5PxY0bXiq4iAiIstHhS5lKVoaYbWIV7puOouC2fp8IP33G+uzK4qC6a5sr5TE5NK497Gzku4/wzpRHYB1QlqCdXI6shmR3rvMrekpPhQiosNSYMdihBxBwWlTzaKxGZGRKArmYoRs1wunBcub1e1XCm9+OpT6fNShL0H5HFkASvNzri0XRXy/rdftg7DOX1Ojdl8AMAgrBx4YS2kHkuo3zE0zbmylRfALPciYHDuEaOaakkHiAEC02QJOp3xKqWSEU6n0j0qGUQIASZKkULBhRSQaXzNVsRARTSRFcZ3c2d62BwC6w33zVDW/d7pjOpqlaJGWoqUNVquQHVaXdacumB2/aIt+rtelndGsORJvTCt7RkQDva4RX49Lq38qlLw24tAXvmPYvxZWUd9d3qUXQKo8gBILdKIZaqbmqyCcwlKhxbkIfgkAepDRnzLDhSgK42qFvgTtjuXCXK8Cu3VbpwCswgJzMyK5KApGZbu/edS3Pd4c/77NFEpLUu6+Rk3K5Gwlc4tX7dhVl1/c69QWXNUb+mlwRNoCIIiDFyYBoBFWK3r5E6AtRL3QAXeuB1kD02QpWhxvFVrqF6K+TkVRUWA3YV1E0HuQ1R4192U2I8JKnWiGmZICXVFc8zvb29bpenGHaBMDPp/nMz09fVeUDKMgSXZ3MNDwSCQan/I52YmITlSXCO0BWMW5D9Y95jYADc8FU1f1urQz3pLwxH263VzfkD4pYy+9ZrySl325c+p1e+nchOc5AB4Ae2EV5XVTfBhEVMOWosWxRAg6FEhyO9x1CuxFlAegXIh61yVCu/fn5q7MZkTyx1OoX4J2xyphgX+MVcJihNwoF9VxR7Hxd83x/7Qbwshc1ZndXZdv2uTLzhn9hrzNcP+qLXrV8oFAtrXg6IaVMxVYLeqe8vPEqLeoq4RTineZW6JT3VodhFNYJSxoXoxQJ6wB79wK7JULBTYA0Q6409cLp+mPYl//o+a+GdfSTzOboih1qqrmJmp/siy3NDY2/mysdYVC4TeRSOTuifqsWjAlBXqDv/4zmWzu+719A1+ziaLY0dH2cFNj6Jb+waEvT8XnExHRa3XA7YV1u5GvvEjSRGPOJl/uoo68nM6LhuuFxoQLAEQTmKs6zSZNQp9zBN2KJjwbTJ0XkXXvuwYbHhWsk9y9AJI9yHIkYyI6og64hY8Ib6jvgNsBq2Csh9Uj58DI6bAKXvsqYUFwFRYkdiCZ/bm5M1tti7QCO5YLc71H2UwCgA31mY8YgMcQTWOnO++uK9lwcs4JlyGWBmVdjDl0AQBS9pL/4TmRf1k+2HDfKVnX32Bd5HSVY8/COq+uVOPKQtQ33igs0W81N6TG+OxJ0QG3eKOwpEWBfS6s360b1u/SC6t1Pw0r9xcBqMsxtz0oOHvuN7dlD7tTIgBut3txIBD4gizLl8L6P5vUNO3JWCz2H9lsdvN49m0YRkFV1WdHLxNFUfT7/TcUCoWHx7PvWjQlBbpNFJviieQaACgZhjEwMPzPc09qfzmZTt+HUffrEBHNBB533ek+n/djTlleIkn2AICRgqaFVTX/VDQav6dkGIXpjnEC+HGwOAcApdc1srgkmJJoQnvZl3MBgE+348qBgNmoSQc2DLs0c21TQvirRz21XrfHl8W86wHEAWQ41RrR1Kq1fKXAjuuE0+uDcDrKiyoFLmAVkI3l5yqsgl0A4F6I+tKNwhLHHeameDVF+mKEpAPd2g/u217eZ6WILgKQ99QV3gYANlMQFqfqzIVZl9CoSSXZEAUTMP+vPmv+bygpyiURhmA6ftcU//DHC81fcxdtKqwiP17e76HTsDk64K5bipb8VE3Dtko4xaPAHoJVnLtgDXynlOMErIsKSVjjiKgAxKVoCWxHUluPAeZvGlNzc/MnvV7vXQBGmjRpt2yIL2mi0Twk44rW1tYr0+n09YODgz+odn+SJPlFURQNwxjxer1vi8Vij8Visa+O3iYUCv2DrusvxmKx+yf8gGa4seZznHC6Xnw5GGz4J0mSPABQ0LRIPJH8amtz02pJko52dZOIaMp43HXL5rS1/K9RMlLxRPKO7nDfJ3r7Bm7IZnO/VBTXsnnzTvpfmyhOyRSVk2UxQnZY0wdV2AH4U1KxAQDCrpEDLefv7X9tcQ4ALQUHLh2uR13Jhr/4M2/J24xWACcBkJZjrucO4Tz/YoRe+yYimnC1mK+WokUeVZwDVsEIWAVkA4AWWNM5NsBq9bWV10GB3XadcHpV541BwVlphBJh3TM+r/yYX/7pBwAVxVzOXgoAwLKYVzsl6xLqdXtJNqxTZAHAm5NuvDnhMQo2A6fkXKkR0VReqs9eAMCANUjcWBdBSihfAHir0OKqJubx6oBbXIh6BVZhLsH6/Yk4WJwD1iG5YB1/5XckLRGCUxIj1R6/33+p1+u926fbez/QG1r74Z6m2Ad6Q1L551qfbu/1er13+/3+S6vdp9fr/efW1tbvdHZ2bgoEAg8dul6W5Ta/3//1oaGhjxuGccLNhjAlBXokFv+uTbRJ80/ujIeCDX8HAEPD0Xt1XX+xtblxy1TEQERUjQZ//T9EY/Eb+geHvhRPJB9V1fyzmWzufyPR+AP79oev1HU9JTvlN013nOPRIbgdsFpOKpwAxHrdngWAkmDaAGCe6jKDI9KYX4zeoh2LMi4UBVPaqxTaYZ2oOgCYQTgd1wun+ZeixTHWe4loYtRSvlJgx1K0OC4R2n2wckXlQqETB7tgK+XNK63dEqyePlJ5GYJwyh1wH/X8VTWLBqzz3DZYBfroXqMSgCYALWFkdQOIAVAX5FxWc33RVsl7lbnOcXbSbQgAHIYIb9EW268UFsAqzLXyAzjYMl+Zdg0AxHa4p+QiyUJrgD0BB4twYOxzfRsO/v4r752ygaOptgQCgW8AiFw5ENjYkZcDo9d15OXAlQOBjQAi5e2qJknSqlgs9vmdO3e+7qJbY2Pjf6TT6XtVVe0e672z3ZT8Z9R1Pbt77/53OWU5MLqrVXe47yanLH9XUVxLpiIOIqIq6LAG/HkdmygKNtHmAmp7fppF8NtgdckMwjqJswPQ5+Tlfrsp6EXBlADg0JbzipJgCg5DgF+3vkJSUtEF64TUgHXCVwAgrBIWeHeYidh4R2EmosOqiXy1XJjrXI65HliFYQuAAKzcUyj/dMGastEGq7g0YOWlUnkXdTg4Sjra4bEfbbyLzYiMrMKCehx58EpfFHknAE+TJimaaGiiKRjC66chNl0lUXCWRGRsJaFhxG4OOHUfrK7iRjk2rfy8CCCHg793Y6pyoCLYBVi/MxPW77J0mE0r21XqAJ3TrtFYFEU5WRTFM09L121o1KSGsbZp1KSG09J13Vu9ubMVRTlZVdWqZobQdf2FRCLx2zE+8w0ul+vvBgcHTx5v/LVqSq+WFTQtNsaygYKm/X4q4yAiOpx4InnnnLaWZ9zuurdms7kXUG5pFkWx1eN2Ly8ZpX6toG2a5jDHZTsSpYWozwEYhnWy7AQgOQxBeFPSve1Ff+Z0ACiIB27zrNwHCl0wK01KxRHRsAOAwxArJ9RFHDzphgK7bTFCjqcQ1kBEE64W8tUqYYHzErT7YOWQelgXFNzl1XWwWtN9sLpcizhYHKdgtUKXys8lWBckcAzF5NF68ehL0eL+k9GTytoNFAVTFmDlt9FFuiEAmmjaiqIJyRSKg46i4S3aBgDshtXyb8Iqyg28vigeCSMzJfefbzYj+nJhLmD93prKsWiweiaMvuhQgnVBpHJhVduBBAf4pLF0AICvaAvDurA2pvL6s8vbV1WgF4vF3rGWBwKBG7PZ7IO6riePPdzZYUq6uBMR1YpMNre1O9x3hqrmX1ZcrqWKy/UBxeW60inLbfFE8raenr4VpRq/H2qzGSnAOilLwiqmM+WfkWUx77Oeoi0FADvdeWiiAUMAioIJTTSEkmAKdlMo5m2G0Ouyzufm5OW9sE4Ai7BabQ5QBDu7TRJNkpmer4JwCpegvdLCr8AqmA/tmuOBdR+6A1bh7oN1fmovvxZh5ZVKN1gzjMxRBzMLwmWDVdDnD7OJDquQdXSo8vqcrYRhWYcJICUVURJMsySYpi6YGBEM2566PHTBRN5m9GXspUY5bzwRRWEYwCCsXKrj9cV5QUVRe9Tcp2IK9CBbiqKgweohVTluFa+9R77S/T4FK940AGO9OXi43xOd2Cp/O8GjbFdZfyyDUr7uSpskST6Xy/XedDq9+hj2M+uwQCciOoSuF3XDMPoz2dzq3r6By7rDfX/XHe77ZCab+31DQ/3y6Y5vvHqQNdZjQIV1UhyHdQKnAuixmcLAW+Ke5wAgYy8JT4WSYkE0xKJgCgBMyRRGsvaSuc2j2vYqBczJyz3NmhSG9aU8VmtVTV/MIJrpZnK+WoyQjIPnmk5YOaIIawRxwCrWpfJrGw5e4BtdgOuwLii6ADg2I6IeY5dxFVYBnS9/TuWiZLq8X7MvGfk+gOKL/kxatRmJtL1kz9hLdl0w7SXBFHN2o/BcIF0CUOy2xtxIrk/u/dat5oaBzYgMwGo9P1RBRTF5v7ktOZW3+dxtbkmrKBYAhHGwCI+WY6w87y+/HgIwcr+5LbUdicN1h6cTmK7rWwDktnrUI90qgvL6XHn74+Z2u98JIKGqak33VBwvtmwQEY2iKK75ne1t63S9uEO0iQGfz/OZnp6+K0qGUZAkuzsYaHgkEo0fen9izYmiUOniKMAq0iutVJklKffesGtE2eZRL3rVqyItlUpnpurSsiHKCano2FtXsO9VCnAaYu7yIf8aWCe6OVjfKa/pJhk1C5y2h2iSzPR8pQgHpjkTcbBVHLDyhQmrhbwEK/dU7t2uFJV5vP4+afF+c9tYxfDr9CBThNW921Xez5it2CqKhR3qcHdzWvw8vN7/XNsUzy7MKmjSJDgNcSTq0JNbvWogZyvZAMAA5MHBwZWapkUA4C5za3YR/PnFQrCuA5668nEWepBRnzLDhakeg6MHWeNWc0N8lbDAXIwQYA2AJwPYh4Ot6jYAI5sRiT1l9uZYnNPh6Lqu5vP5n8Dluu6ZUPLRiyL1vkO3eSaUTKWk4oX5fP5uXdfH1VvE7XZfls/n15+II7ePxgKdiGiUBn/9ZzLZ3Pd7+wa+ZhNFsaOj7eGmxtAt/YNDX57u2CbSImvEXhXWCZsMq7BWUB5gaPlgw781jkivvOjPXN3j0hp7XJp/9PubC47d7x5q+FWDbo+X31uZbuhAN8koCiOcV5do8tRAvqoMZGEe8hqw8k8l91Ra13VYBXrl4mFlVPIcgKQCe15FsaoTdxVFPIVw9BK0tx9hM/MphKMAMDg4+N1SqRTz+/3f3urNubda6x04OC+7qev6HyORyOez2exrWve2I1HabibSsC5WTrsoCuZd5tb4JWjPrRIWVO7/P6AHWe3H5t/S1cwnTzQ4OHhLZ2fnBS/VZ68YkvU/vznpTsolMabZjMBL9dn6Hpd2hWEYWwcHB2+pdp/pdPqBbDbrPHR5LBa7q1QqnbD3nlewQCciGsUmik3xRHINAJQMwxgYGP7nuSe1v5xMp++D1S1yVlAP9iCtjD48Auu+RAnlbqZviXtWvynh/q9B58gbh2T9koJoNNhNYWROQd4/J+/ogXXSV1f+mYHVXbIyiFPxbnNLakoPiugEM9Pz1Q4zOVIuDU1YLeQ6Dk7/BVgF+wisorwJB+9JrVz0y5ff04eDo6VX7VFzX2qh4Hd2wB0aY7XZg+zgo+a+Ay3ykUhkdTKZ/I3b7b5UFMUuWBcsd+m6ntM0baumaQPH8vnT7SmEtc1mJLIYIbncm8HsMbP6ZkR44ZSqput6OhwOX9jc3Hxfjwsre1yvHfdV07RHBgcHP6bretUXqHRd3z/W8hO9a3sFC3QiolF0vfhyMNjwTwVt5C+6rmcKmhaJJ5JfbW1uWh2Jxa+d7vgmynYzqS8WDpyzFmEV5iYO6aIumYLWnpc3tuflGKzWJCesE2wnrK6cKqwT6CEAQ1EUijuQ0B419+U5vRrR5Jrp+Wo7EqUdSOYXot4FK1fYYeWLykBxJVj5pwDrPvFKEV/EwRHIK6OjowfZYyosVRRxq7lhaClacpcI7Q0dcLsAmFEU8usxEBtr8DZd19Xy1E+vm/6pFkVRMJ9CuMDRQGg8NE2Ld3d3X+l2u7tkWb4M1mCOKU3Tnshms9umO77ZhgU6EdEokVj8u53tbf89/+TOeDQWf08kGv/90HD0Xqcsz2ttbhzX4CczyXoMFJZjbp0Ce6Wr++vuK4NVrBcBYDMiry5GaC6saZJeM9qvimL8DnPT9mM9eSai8amFfHWXuSVzo7BE6LBmVsvB6nHjhXUOmobVY8eENXhZZYTx0eXkgSL6KTN8XPe3rsdAdr05kD2uAyCiA7LZ7DYW5JOv5gc6IiKaDE5ZDpQMo6Drem7UshZFcS2JJ5K/H72torhO7mxv2wMA3eG+eaqar2oO0Om2GCH7R4UuvwK7COv+89GjtJZgtWgZ6zGQvd/clgvCKS4X5gbb4Ql0wO2IojCyA4nIo+a+GFvLiaZPLeSrpWhxLBGCsgJJBiAvRL1dRVGIoiB1wC3g4HgYo+9LVVEeIX09BtL3m9s4FRgRzXos0ImIxqlWC3QA6IBbXC7MrStPh+SAdYJcApDvQXbkKTOscqA3otljJuarxQjZlwtz6zrglgDIKoqSArsOYIR5iIhONOziTkR0AutB1rjL3JrBDBhQiohOTJsRKW42IxxUkogI1gA/RERERERERDTNWKATERERERERzQDs4k50iEVnd9lhXbwqbd+wrTTd8RARERER0YmBBTpR2aUffIdjxXUrPYpXOfD/Qk2rxTV3P5LZ9PTGkWhfZDrDIyIiIiKiWY4FOhGAFdevlN9z3cp6WC3ndeWHongV21U3X1246uarh6N90ey6Nc+pa+56pDC90RIRERER0WzEe9DphBdsCwnvuW6lF9YUUy0A2sqPJgBBAO0ATgm2Bevec91K37/99jbP9EVLRERERESzFQt0OuEtW7msMv9zEIAH1jzQSvnhB+ADMAdAFwB/R1end8X1K53TFC4REREREc1SM7ZAt4miU1FcFyqK60KbKLIYoknTsajTDqsYr4NVqIcABMo/PeWHNOp5/VtXnO+dnmiJiIiIiGi2OuI96JIkzan3eVZWs6NINP69iQnJIjvl1s72tj8CQHe4b56q5vdO5P6JKqJ9EROAE4ALQEP54QEgADAB6OXnCqxCXQi2BRsUrzKsptXpCZpeZzrzFRHRsWC+IiKiwzlKgW6fHww03FnNjvgFQrWqe1vPCAAbrC7ucvkhlFcLsFrVbQCysAp0ALAtuehNjnVrnhuZ4nDpMJiviKhWMF8REb2eKIqi2+1+uyRJ83Rd35ROp1+a7pimwxELdKNk9Kpq/k4AEG2i3SnL1wIQVDX/OICYJElnSJJ9WckwTshfHs0O69Y8p19189WC4lXk8iJzjM1MWEV6pXAvhuYE7R1dnUXFo9i6zulyLDq7S1TTKrZt2KavW/OcpqbVsfZDk4T5iohqBfMVEdUaURRFn893udvtfiesQZUHstns/6RSqbWGYRgTsH9ne3v7Y5IkLdR1/VVZlu9UFOW2wcHBfxt/9LXliAV6QdN2d4f7bgCA1uamzzplua63b+DMTDa3tbLNnLaWuzzuug87ZbmloGkDkx0w0WRQM2pO8SoFWF3djfLPShd3rX9Xf8PG37/0d3s27u7MZwset98dnv+m+X+48f4vrPYEPEJ5+xKA7JKLz8quuG7lyH1f/GFy09Mbi9N2UCcY5isiqhXMV0RUSxRFmd/a2vorURTPBFAEMAyg0eVyfSoQCLzc39//XlVVd4/nMxobG2+UJKl57969XYZhqH6//x2hUOjxRCJxj6ZpkQk5kBpR9SBxkmRfCqCgqvkdo5frur4ZQJ3HU3f2RAdHNFWCbUENQOUEqFR+mKViyXjyvicu+clnf/T5Lc+8skwQhI5QR8hfyORP3/Q/G2/84ae+/+yejbuXwbo/vR5AB4C5ildp/vR/faZl0dldR7wIRpOD+YqIagXzFRHNZLIsz2ltbf2TKIqnrAzUr3t44UlPPX3q/J0PLzzpqZWB+nWiKJ7S2tr6J1mW51S7T6/XuyIUCl0VCoU+ecopp2wsL7s2lUp91zAMFQBSqdQTvb29l+i6rk/Wsc1UVRcPJcPoBqB0dLT9dzQav6NkGNE6xfUmn8/7VQAoFLR9kxYl0eTTAIwAiMG6cCUBcP7hB2vf9fKTm89taG3ARR+5BB2ndsI0TBRyBbz0+F/w0mMbAo/8x6//4yP/+Q83B+YEs7D+T+kAIgBiX1x9i7RuzfNDD9++OscB5aYO8xUR1QrmKyKayRobG+8QRbHxxjlNz1xa75FhDaqMJsnu+lRLsDjfJT9/R+/QRY2NjXeEw+EPVLNPSZJO9/v979d1PRWLxX4my7IXwLx8Pv+3QCDwIQCdhmFsTaVSv52I7vO1puoCPRqN36EorpVOWV4xp61lxeh1BU17OJPNbZn48IimRs+2br2jqxOwurQnAOT3b9l35stPbj7HE/Bg1deuNrwhrymKomkapiDJkrDs/10gOpwOrP/lOs+fHn726pU3/f3PyrszYBX7IoDcW1cscwPAfV/4YW46ju1ExHxFRLWC+YqIZipJknwul+vvz6hzbS8X569zab1HfiKR3v4K8PeSJH1S1/VUlbv3dXd3LzYMQ1MU5WQAaG1tfUDX9f3FYnG/y+X6lNfrvTocDl9pGMYJNa5T1V3cC5o2sG9/eHEqlfmarhf/DOCVgqY9Ho3Fr+np6bt2EmMkwtLzu2y3fetq5bEnb/au/tUN7lXXLHNM5P43PbNRhzUIXGUwONvLT24+D4Bw7oq3lOqb6iGKogBAFERBsDvskBXZOPOSxfC3NGDnX3acqRdGKhe8RABeWK3wPgB464pldR1dnVX/f6PxYb4iolrBfEVEM5UkSW8EIJ1R5zriOBjl9VJ5+6rk8/kXDcPQRi/LZrOP79u377JwOPyPvb2958my/C63233x8UVfu47p/lhd1+PxRPL7BU172WYTzWQq8yQApXQCdj2gqXP3fR+vW3XNsjocHEEdl19xVt2NX1o5cs37vpt69ZXucf/9KR7FAaAAq9uOyzRMefdLu84AgJPOmCuM9R7RJkLx1Zntb2gXtjwTl1ORdCDYHhwqr64MHOesvD7r4rPknm3d+fHGStVhviKiWsF8RUQzlB0A3DbxiI1Mo9YfS22ZHfU8BQDJZPK/KwtUVd0DYKckSV0AnjqG/da8Y2rRa21u+te5J7X3NjUGHwkGGm6VJHvz/JM79/h83rdPVoB0YrvtW1e7Vl2zzI1RxXlFR2fQsfqXN9T76pVxf86is7tssO4dLwEwo+FIq6ZqHgDQC4cdm0KwO+wYKVhTodslWwhW6zvK+5FgdXVvO/ayAAAgAElEQVSvxM4W9CnEfEVEtYL5iohmolKptAuA+Uou7znSduX1Znn7Y6aqagxAjyzLJ1WWiaJoAxCCdevpCaXqgqHBX/8en8/zVQBJXS/uAgCjZOQAGK3NjQ9LkjTmfQlEx6ujMyh84rrL3EfZRlp1zbJx/+11dHUWYRXUAGDXNf1Aq8Wul3ZWngqHPExd083ev4VRV1+X9zXVG7Ba4A1Yxb4JID3qY9gSMkWYr4ioVjBfEdFMpWnagK7rz65P596wp6CVxtpmT0ErrU/n3qDr+rPaOKaETKfT9wQCgW+43e7FiqLUNzc3fwuAkE6nHz/uA6hRVRfoHnfdewHE9u0PL0ql0w8DQEHThqKx+D8AaFYU19LJCpJOTEvP75JQxd/o5VecNe6Tl+0bthVhdXGXAEi+xvq4IAolANjw6F/E4e7hsd4mPH3/k8gmsjjz0sX7BUEQYRXlRRycI7JQXmZufHqjNtZOaOIxXxFRrWC+IqKZLBKJfB6A9Nl9fac/mcwkR697MplJfnZf3+kApEgk8rlq96nrenehUNg+etnw8PAdmqY90dra+sKcOXMSiqJcMDg4+E5d15OH289sdSz3CcgA4gVNi3k8dQcWFgrWxPEOyX7Elk6iY9XRGRrz3u8x2I6+yZFtfHqjtujsrhEAOQDBuvq6fPsbOnb0vNr9Bl3TsfoLPxXf/qGLzTdecCocTgfS0bT5p4eexZZnXhYVn6Kd9/dv3QTr/pk4ABXWNGuJ8nOsW/N8rmfb+O+Vp6oxXxFRrWC+IqIZK5vNbhwcHHxPc3Pzf9/RO/Se7w9Euuc5ZWNPQROzJaMTQGpwcPA92Wx2U7X7TKfTPz10mWEYpXA4/ClJkj4tiqJT07TMRB5HLam6QNf14kYAV3a2t/0EgAeAJxRs+IDfX/95AKVMJlf1PwpRNbZuOWpBKwIQUil1zC43x2LdmucKy1Ysy3V0dY7AKrRdb7v27b9Z/YUHTjEMw17IFrD27t8Jf/j+WkguB7Rc4cDFg7d/6OJNDpcjCmAfDhblOQBRANq6Nc/nHr59NadYm0LMV0RUK5iviGimS6fTf9A0bZHf77/OcLsve6VkNBqGMZzNZh9KJBJ3a5o2OFGfpeu6DutW0RNWtS2UkCTJPfek9qdtonjOoetSqcxX+geHvjaRgSmK6+TO9rY9ANAd7punqvm9E7l/mvl89Qo27/jPoM+nHNpCLgOow8GW88T657Zlbv78w9nxjOi+6Owu+xdX3zIfQAes6dHqXn1261t/d+dj15aKJel1bxCApe9965YLr3n7BgA9AHYAGFDTaqZne3d4+4Zt+Y1Pbyyw5XzqMV8RUa1gviIiotGqLtABwCaKDp/Pu8rjrrsQgK9kGP2ZbO6XqVT6uYkOjF8gBACfuP4yx23fvNo/alEdrDnGnbB6gGgAkgDUVErVrvvYvYmOzqC09Pwue0931Hz1lW597eMbR1JJtarPu/SD75CvuvnquQDaATQAcA/vH1r0wm/+/M7dL+2aW8gVFEmW9I5TO/vOXXnelpNOP2kPrG7tgwD2ANj179fcNrB9w7Zxt+rT+DBfEVGtYL4iIqKKYy3QRdkpz4c1UvVrqGr+lQmLCvwCoYM+cf1l8m3fvNoLq+W8BYAC629XhDUYmwBrhPRs+TF65HT0dEePab70JRefJV118zVtwbbgQgCNAJphXRSww+pyo+Pg/50cDl4k+OuTD/xh+8O3P8S5zmcA5isiqhXMV0REVFH1PehOWT6po6PtCZsonjLW+m07dh9TsU9UrXvuekJb+9jG6Opf3dB66umdJoA8rJZ0GYAbVqFuA9AGa+T0XbDmHwdgzZf+2JM3+y88+5Z4T3fUPNrnbXp6o759w7b9V918TeytK5adBKALVpFeLH+OrfzZAqxiPQ8gEe2LDrA4nxmYr4ioVjBfERHRaFUX6MFgw03lL48kgO7JC4no9Xq6o2Z7Z8gJa9qyBlgD6Ug42NXdDatwVmD9jfbDmt4MAODzKfYbv7TSdd3H7q2qr7uaVnHfF36YUdO53Zd+8B2Ve90rLfd2ABlYhbkKa1C5yKan/y899t5oqjFfEVGtYL4iIqLRqi7QbaLYBKDUHe47VVXzfZMYE9HrdHQGBZ9PscMqyn0AHMWi4X76D39d8vyzOxcnE2qTw2ErnL64fc+7Vy5GIOjWYRXOIygX6kvP75JRnvasWuUW8V2XfvAdeQALYBXpRVjd2jVYLeiqmlaH1tz9CFvPZwjmKyKqFcxXREQ02rFOs/YuXS+ecJPF0/Rr7wwBVjHsAiD39yXn3vTpX/zj7p3D8+x220iw0Z3IpAstL67fc9qDP1p/6Y1fvvyb77zi9N/BKqbjALSOzuBxzZf+8O0P5Tq6OvsXnd2Vg9V674TVrb4EQIv2RQfv/Kf/jKlp9ajd52lqMF8RUa1gviIiotGqLtALmva8D57knLbmn2WzuUcPXR+Jxn88saERHbT+uW0mrAHZfNms1vHPH3voc4MDqcBHP3n+jvdfdY7p8TpdhmGmXli32/at2/+n+d++9NiXPV7XyFsvWPAXWAO8DfZ0R6PH+/n/fs1tmbeuOF9btnJZZtHZXUq0L2rbvmHbyMan/y+16emN2kQdJ00M5isiqhXMV0RENFrVBbrHXbcSQMgpy+92yvK7D13PLxCabD3dUXtHZ9D+03ufv7SvN9H4xa+8a8+7Vpwpp1N5pb8v6TINE4ve0FL8/o+vjX70Az9u+dZta28457zr/lWSbEOwWtLHdY/4ujXPjaxb89wIrK7zNIMxXxFRrWC+IiKi0Y6lBf0lAA9OYixEh7XqmmVSR2fQBSCz9rEtZ80/pTF5wUWL6v66tS+oj5Re83csikL9hRctyq351cbQxg37zzp36bxXAQgdncG9q65Z5vj56udHxv4Umi2Yr4ioVjBfERHRaFUX6EPD0YcBPDyJsRAd1tLzu5wA5O59sVIingsuPX9BbO/u4Wa73WY0t/pKTpckGIYpagUd8WhO6JwbdAPAvj2R+ecunTcMa3C44Kprlqks0Gc/5isiqhXMV0RENNoRC3RJkubU+zwrc2r+L5Ik+RySfdHhto1E49+b+PCILB2dQQcAwTAMAEA0kgnYbKI5p8MPu2Sz2WyiabOJptstw+N1Yc+uYQEASiWjBGtAtwYAcwAkMM6u7jQzMV8RUa1gviIiosM5SoFunx8MNNwJxG9VXK6TFcV17eG25RcITaZUSgUA99x5IZvH64zt2jEU8AfqBMlhE5xOybDbbUJlW6dTMpMJFQCEjRv2d1794fM2wvpbr1t6fleDr17pSyWPabY1qgHMV0RUK5iviIjocI5YoBslo1dV83eO6MW/iKK2C1brI9GUWnXNMvnyK86SATgANJx6+pw9L6zbHfjLn/di+ZWLDbv9tbOnpVN5PPqbTQCAF9fvmb/1ld6O086Ysw3W9GiOU0/vlNY/t02f6uOgycV8RUS1gvmKiIgO54gFekHTdneH+24ADgxb/bOxtpMkSZnowIgAYOn5Xba77/u4D1Y3dRsAryBAAIA7v/kkRAF495WLYbfbTADC317tx+1feVwYHkoLDQG3GY9lhT89s739tDPmbAFgApA6OoPSemtOdZpFmK+IqFYwXxFRLRJFUVIU5SxZlls0TRtQVXWjYRgTck7t9Xovbm5ufuqQxYmdO3c2TMT+a0nVg8QBQCjY8FHF5brs0PfJTvmMnbv2zjva+52yHJKd8pkOyR4AoBcKWo+q5l8uTdA/LM0+n7j+MgWABMAHIAtA75wbTP35+d0wDRPfvO1/xB987484eV7IjEaz6O9NCLJTwpe/vrx493eeFgEI3fuiPlj3nccBODo6Q8f0d0+1abz5iohoqjBfEdFMJoqirbGx8V+8Xu9NAAKjVsXS6fQ3hoeHv2MYRmk8nyFJ0nxd1zcPDQ39y6jFJ2SNWHWhEgo2vC8YaPjR8XyITRSVjo62+52yfEXJMHZqBS0l2kQxGGhoB+COxuKfikTjvz6efdPsdurpnRKAOgACAAWAdOFFizb+/MEXLz719Db9Pe89S3jujzvE/t6E0NLqM995xWnm8iuXGC+s223GolkRACSHPQagH4AGYGTp+YtYoM9y48lXRERTifmKiGYyURRt7e3tv5RleWVbvdxz0SL/pkVNirl9SBWe2Z5YCHjvkGX53HA4/L5qi3RFURbb7XaHYRhFv9//gXA4/FlZlucVi8VXVFV9dpIPacarulBRXK53AyhGY/Gr/P76W7WCtiWTzf06GGz4mq7rm4703qbG0OcAOHbv7W7RdT0zep3HXXd6S0vTY5lM7oWCpvUd32HQbOWrVwRY9547YLWi4/TF7eFzl87b9uL6PV1nntU58tV/X1GSHDYBVhd28Ynfv2p+5z+ekAQBpmlCeNPZJ22DNc2aDiCx9PwucbqOh6bGePIVEdFUYr4iopmssbHx07Isrzz7JO8f73r/KarHaaurrPunC+Zsvf4XO/ds2J9e2djY+OnBwcHvVLNPl8v1br/fv1QUxTfm8/mNAGCz2eYB6A2FQl+QJCmYz+f/nEqlfmtUpnA6gRxLS6ICoD8Sjf9SluU3S5L9jHgi+WsAelNj8NdOWb65oGk9Y71RkuynxROpew8tzgEgk81taShoz4s2cQEAFuj0Guuf22ZcfsVZwMFW9CKA0kf+8fw//uXPexc98KN1jv95fEvpTeecZEgOO7ZsDov79kRsLpdDE0XR7nbL6iXvPPUlADmUp1hLpdQT7j/6Cei48xUR0RRjviKiGUkURZvX673R47Ttuuv9p+RHF+cA4HHa6u56/ynqxXdu3gV4bxweHr6z2lZ0URTf0t3dfaqmad0AIMvyPFEULwbwCwClUCh0v9frXdXd3f3eiT+yme1YWhLjANpCwYb3aZq2xSnLSz3uunMVxXUuALvHU7fkcG/U9eKfG/y+mzzuugWjl9tE0d7gr3+XorguMErGq8d5DDSLrX1sYw7W4HCVv9UcgPxpZ87pufoj5/0vAAwPpW1rH9siPfabzfbhoXQ+GPKkIUA0TRM3fvnyuzxeZ6T8vjgAY+1jG0em5WBoKh13viIimmLMV0Q0IzmdztMANK04M9TvcdrGHLTS47QpK84M9QNoKm9fFU3T1lWKcwBIJBJ39/f3nx8Ohz8WDoc/0d/f/zZZlld6vd4Lx30gNabqFvRkOvOgorg+FAw03LNvf/i0YKDBPqet5YXyaiOn5g9bYA8NR743p62leU5by0uwWkBz5VVNJcPY1j84fG1B06LHfxg0W/189fMjN35pZaqjM1gPoASrFT0NQPzUp9/+tM/nKv34nucuUNUR2TRNMZfV3LmshkDQPfSV29/zvQvevvBFAAaAQvmncc/dT3AS9FluPPmKiGgqMV8R0QxWDwAepz0CIHi4jcrrD2xfDcMwhke/jsVi941+nc1mXwawU5Kk0wA8W+1+Z4OqC/RUKv0cgEvqvZ4rC5rWF43FV/n99TcDEFOp9HdUNb/7cO8tGUapO9x3k00Ub5Gd8jwAjbAKph5VzQ+N+yhoVrvj648M3H3fx/2wWtF9sO41jwNIX/Wht2T+bvkZm9Y/t2vu4EDKLwiCdvL80LZzz5u32emS0uX3VAp747qP3Zt89ZVudnGf5caTr4iIphLzFRHNYDEAyBSKLTjCiOrl9Qe2P1aSJDV5vd73JxKJHxwybZsMaxypE4ow3QEcjqK4Tu5sb9sDAN3hvnmqmt873THR9Nk79MNmn09pgFWgO2G1hgPWvXs+WBebdABRWL007LAGlrMBGFn7+MZXbvncQ5Ge7iiLc5pwzFdEVCuYr4ioWqIoivPnz+/2OG2lpz+9eJvHaXMeuk2mUCpcfOfmrkyhZNu9e3eHYRjm0fYbCAS+oijKyeFw+IMAIEmSe+7cuX3pdPr2wcHBb5S3+X+BQODH3d3d8zRNG5j4o5u5jtiC7pTlU5sag1+rZkfd4b6VExMS0etd97F7Y6t/eYMJay50NwAXrCI9B6swVwCkRr2lWH4AgNbRGSywOJ/dmK+IqFYwXxFRLTAMw0gkEl+H33/PBx/4W/iBD74h53Ha5Mr6TKGkffCBv7kzhVJnIpH4RDXF+Vh0Xc8ODg5e09zc/KDX6/2wYRi6KIqdsVjs2hOtOAeOUqCLNjGoKK4V4/2QpsbgTU5ZPudI2/ALiI5k7WMb9es+dm/itm9d7fX5FK28uDJwXBJAA6yi/VBFAAOnnt7pvPzdZ+XWPraxOMY2NAtMVL4iIppszFdEVCtisdi9iqK8efsgPnrxnZt7V5wZGvI47YOZQrF5zcuR+ZlCaY6maffHYrF7j2GfX43FXtsbPp1OP5bP5ztcLte5NpvNls1mX9B1PTnhB1QDjlig63rx1WgsPu7COZXK/FRpcXU5Zflt0Vj8q+PdH52Yfr76+ZG1j2+MXn7FWY6OzpAdgPn7xzeOrLpmmfKJ6y6zAdBg3atig3WfugbrXvUCAJx2eqeDBfrsNVH5iohosjFfEVGtMAzDDIfDH/P7/S8iELjlwRcHzxq1en8sFvtYIpG4/3hbz0fTdT2t6/qT491PrTtKga5HI9H4mtHLJElyK4rrzQ7J7sup+b1aQdtaOso/SEHThoaGozd1trc9EYnGfzwRgdOJKZVU8fPVz48AODBVWseXVgBWK7oE677zynzpIzh4rzowg8dcoPGbqHxV4ZTlkOyUz3RI9gAAvVDQelQ1/3LptYOXEBEdM+YrIqolhmGYsVjsR7FY7EeKoiwA0AagT1XVXdMd22xU9SjuADCnreUrHnfdjbDu90UwAJQM49Wh4egHU6n0piO9V1XzQ93hvo+OI1aiMb36Sk/x8ivOAqx70Q97MpJK5UpTFhRNu+PNVzZRVDo62u53yvIVJcPYqRW0lGgTxWCgoR2AOxqLfyoSjf96qo6DiGY/5isiqhXlopyF+SSqukAPBRs+4HHX3QqgqOvFZ3VdH5Cd8lk2UTy1tblxrarmF+i6njnSPlQ1v3G8ARMd6uern9Nu/NIKE0duITfWPrZRO8J6mkXGk6+aGkOfA+DYvbe75dBtPO6601tamh7LZHIvFDStbwoOhYhmOeYrIiIaTTz6JhbF5VoOwOwfHL5o9979b+sO931gz579i1Q1/yMATR533bLJC5Po8Hq6o8Y9dz9xxItDd3x9TaanOzrue2OoNownX0mS/bR4InXvWCfEmWxui1bQnhdt4oLJjJ+IThzMV0RENFrVBTqsSeJz2Ux2XWVByTBMNZ//Y/l5bqKDI6rWLZ97KH/H19ekUin1Nd3YUym1dMvnH0p+4+uPFKYrNpoWx52vdL345wa/7yaPu+41J7U2UbQ3+OvfpSiuC4yS8eqkRU5EJxrmKyIiOqDqLu6ZbO5ORXG9q6Wl6b/iieQdAJKSJJ3m99d/uWQYTxil0quK4vIDgFbQktUObEI0Ub7x9UcK99z9h8LS87uk007vFNc9t6306pbuYiqpTndoNMXGk6+GhiPfm9PW0jynreUlWIMNVk6Om0qGsa1/cPjagqZFp/6oiGg2Yr4iIqLRqh7VurO97buK4vp0Ndt2h/saVDWfOP6wAEVxndzZ3ranvL95qprfO579EdGJYyLylU0U7bJTngegEdZUfT2qmh8aax/MV0R0vJiviIhotGMZxT0PoNqim63nRDSdxp2vSoZRVNX8DgA7JiwqIqLXY74iIqIDqi7Qu8N9XwTwxUmMhYhoQjBfEVGtYL4iIqLRqi7Q57S1/Kuq5p+PJ5J/HL3cKcvNLS2NP9i3P7xi4sMjIjp248lXTY3Bm5yyfM6R9t8d7ls5UbES0YmN+YqIiEarukC3ieK8psbgrR533Q/7B4dv0nU9HQo2fDgYaPg2AP8kxkhEdEzGk69SqcxPlRZXl1OW3xaNxb86RSET0QmK+YqIiEarukBX8/nfKYrrckVxfWLuSe2XGyVjhyTZLwFQLM/VSUewFC3SEiHoWgi/XUURO5AYWW8OqtuRMKY7NqLZZjz5qqBpQ0PD0Zs629ueiETjP56ikInoBMV8RUREo1VdoEei8V8lU5knO9vbfiNJ9otsotgBIN7bN3BpJpvbOIkx1ryPCl3upWipK7+0K7DLQbSIS4WW0s/NXcNPIcw55Ikm0Hjzlarmh7rDfR+dglCJ6ATHfEVERKOJ1W7ocded1tne9mtJsl8Ea67NCICGlpam34SCDVdNWoQ1brkw1zmqOPfA6q6mAHACqFslLDjpVuFsvwJ71VPeEdGRTUS+UtU8LzwS0aRjviIiotGqLtAb/PWfkyT7xSXD2NrbN/CW3Xu7T1HV/AM2UewMBhoemswga9Ui+O2XoL0RVlE+B9b8pApe+3sXOuBuuFFY4pmOGIlmI+YrIqoVzFdERK8lSVJ9e3v7L2RZdk93LNPhWOZB1zPZ3K0DA0O3lwxDB4DucN+HGvz1Pw8GG+6dpPhq0lK0OFYJCzwK7H4AbQBsAAIASrCujpdgzXmagTWnqdQBt2sxQupmRIrTFTfRLMJ8RUS1gvmKiGY8RVHm1dfXf9btdl8Gq/Exkc1mn0gmk99WVXXPRH5Wa2vrD2RZfp/NZvvERO63VlRdoPcPDt+g63oWAJyy7BdtYp1W0PriieQTmWzutMkLsbYsRYvjo0JXPQAXgDoAAqw/Yh8ACUABgAari7sdQApWwY5FQr282WSBTjRezFdEVCuYr4hopguFQu/3+/0PALC3Nft2t7V4w30DablvEB93u90fSSQSH4xEIr84nn0riuJVVTVdeR0IBK6WZfncCQu+BlVdoOu6nvX5vJc0NQa/bRPF0wC80h3uW9na3PRQb9/A+3UgfdSdzHJBOIVVwgIvrKJcAWDAKtJ95deAVaQLABzl1waAYQDogKfqWw6I6PCYr4ioVjBfEdFM5na7l/r9/tVet9x/9+3v2XH24vZKDYMNm8M9193821MA/0P5fL43m82ur2afgUDgM263uwHAAlmW379z504BABRFOSkQCHwrFot9OBAIrJ2kQ5rxjmWQuMWtzY2/K3956OXFhiTZ39LR0bbGJoon/CBnC+GXFNhtsLq022D9fl3l56NVLowoALywinRsR6I0RaESzWrMV0RUK5iviGgma25u/jaA7IPfe/9fRxfnAHD24nbpwe+9/68AMuXtquWVZfmfdF0f6O3tfRMAiKJoa2pqWp1IJL6az+d3TOAh1JyqC3Sfz/uPAMyh4eiF0Vj8NgBQ1fz+TDb3BZsonqUorrMmLcoaERSch/ZIcMLqvj4C617zikrBXvnSFQGYm81IYXIjJDoxMF8RUa1gviKimUpRlA5RFM9Z8c5T9y1a0KiMtc2iBY3Kineeuk8UxXMURWmvdt+GYfT19/d/RlXVjQDQ3Nz8BcMwsrFY7J6Jir9WVV2g20SxEcDOeCL5p9HL44nk0wDgdMpV/4PMYpUi3Cg/l8o/07AGhBu9nQkgCyAHwLEeA7keZI0pjJVo1mK+IqJawXxFRDPYyQDQtSAUPdJGo9bPq3bHmqb9tfLc7XYvcbvdn0mn0593Op31six7AUCW5XpZlse8MDCbVV2glwxjP4CuULDhQ5Jd8gAQFMXV1tQY/CwAFAra7kmKsWZsNiMj5acmrIHgzFGvkwCGYBXlCViDw6UBmJsRSd5vbstNcbhEsxbzFRHVCuYrIprBNABIZzXpSBuNWq8d674BQJblMwEEQqHQ1jlz5sRDodBmAAiFQnsbGxtvP8aYa17Vg8TFE8nvedx11wQDDT+pLOtsb+sFgIKmPZ7J5rZORoC1pAfZ0g4k8wtR74LVMq7C6uJe6dKulZflYd1nlgEw0oNsfDriJZqtmK+IqFYwXxHRTKXr+lYA+Wee3+2/7sPnJQ+33TPP7/YDyJe3P2bpdPqhfD6/pvJaluXOUCi0ORKJnKyq6tDx7LOWVd2Crqr5/d3hvnMz2dxPAOwHkCgZxqupVOZfe3r63jdpEdaYu8wtmR5kNRwcnT2Ogy3pJVit5ikAMVgFe3K9OZCfjliJZivmKyKqFcxXRDRT6bqezefzP9u2a/iMB3+1cWSsbR781caRbbuGz8jn8z+rTBl5HJ8zoqpqovLQNC0NAJqmJTVNU8dzDLWo6hZ0AFDV/B5VzX9ksoKZDVQUzVvNDclL0C4vEUKyC/Z9HXDHYQ0YZwKozHOuA8g+hXAsioJ52B0S0XFhviKiWsF8RUQz1fDw8Bfa29vffvv3/njhhs3hdde+96zE2YvbAxs2h2MP/mqj/+nnd19oGMb+4eHhL1S7z3w+/ydd17ccbr2u64lYLHarrusnZCPmMRXoVL2nENaeMsMagPRyYW58OeZ6cHDqNQNA6VHsyzxq7uPI7URERERENONomhYNh8PLWltbH3r6+d1vf/r53YB1u64LAHRdf7a/v/9qTdOOOJDcaKqqPnuk9eUC/avjibuWsUCfAo+a+wpPIVxYjJAUFJy2HjNb2oGErh5oTCciIiIiIpp5NE3r27dv39u8Xu8yRVEuliRpjq7rvaqqPp1Op5+f7vhmGxboU0RFEesxoMOEPt2xEBERERERHYt0Ov08C/LJd8wFuqK4mgDMhXVPdVzXizt0XT+WIfWJiKYE8xUR1QrmKyIiAqos0CVJqg8FGj7p9tRdYxPFrkNW5wua9mQ8kfpBKpV+YhJiJCKqGvMVEdUK5isiIjrUEQt0myjagsGG6xv89V8BUF8yjJ2qmv9JQdN6DcMoybJcJ0n2U52y/PbW5sbloUDDuqHhyHWZbO6VKYqfiAgA8xUR1Q7mKyIiOpwjFuiyU17W4K+/XVXzP40nkvdksrkxh8OXJMlZ7/Os8Hm9n5nT1vLAth27z5yccImIxsZ8RUS1gvmKiIiOiyRJzYriajmW9/h83iUT8dmK4jq5a+F8s2vhfFNRXCdPxD6JaPZiviKiWsF8RUREh3PEFnRd18AmksIAACAASURBVAd1/fWDjjtl2dfUGPymorgu1vXizt6+gU8VNG0vAKRS6U2TFCsR0WExXxFRrWC+IiKiwxGP501NjcH/UhTX+zLZ3COiTZzb0dG21iaKtokOjohovJiviKhWMF8REdFRC3RFcTWPseyyeCL57d6+gc8NDAx93CaKC2WnvGByQjw+J4Xc/5+9M49zqj73/+dsOcnJnskye1hlggjCIKIIeF3Aa93A64KCtrXY26qtVq9tqb+2tmpt1aqt7W2l9mqptdZWKlYrKi4sKggoi84ga2ZhtsxkP8nZf398E2YGBhiWAQbO+/XKaybJyck3M9885/l8n+f7PNT0SIibHgmxQwIO6niPx8TEZOAZrPbKxMTk1MO0VyYmJiYmfXHQNmvhqop30pns623tsQcVRYkDgKbrDU6H46J0Jvs3t9v1XwBkXdNbB3y0/WB6JMT+8OpxJdMjIQGADooyQIF5/7PWfELSsks2tanRlCJ/+lmjmkiKx3u4JiYmR5HBZq9MTI4F/ooANeGiWt7uEhgAet3qOrl+TZ12vMd1qmPaKxOT/SO4BGrChRO5QKWfA2BE66LK+rfX7bsvxMTkJOSgAn3bjui5oaD/uyOGhT9PJtOPdXR2/batPXZXeWnw1XBVRT0ALZlMfycvSYljMN794hEseObr5zqvnFhVAYCFYLNSVt4HmhbAcdr5VRU0KCp31Wy6DYItS4X8yScXvpO56/+9mDue4zYxMTl6DBZ7ZWIyUPgrAvTU2VP56pqwo2ZShBdcggUAD0AFkAOgXHU7UL+mLvfkbY+nxZRoHN8Rn7qY9srEpJsJF9Vys26fba+OhDkANgBuADqALAAZgBZrjslP3vZ4sqEuqh/PsZqYDDT9Tv0WBFtloMT3Y97KXxiPJ+5Pp7OvCoKtVhRzW/KSFD3aAxME27BwVcV2AIg2Ng8XxdyOAx2/7L4ZzumRUBAMY6c8LjcsnBe6bgXLCrBwTrCsBIaRwbIJqGorOK6LCvkz73/whTjry79NmNF0E5OThxPdXpmYDAQ3Lphrm3HzJU4AHgBFYV4Ccq1PA8gASALIA0BDXTT//676QfI4DdekgGmvTE51blww1zrj5ktcILbKDaC08DsDsriYKvzMiylR/NlND7aaIt3kZKZfReI4juMBlLe1x+5ram65wu1yza6urngXADcQF49DZfroUnb62cNLqNLgMCpcMQp+3wjwfDUEoQI8XwKKYqBpNhgGBUN3wsaXgqZGQlUD0889rbLriyfK/+9XX7Z73MLx/igmJiZHyIlur0xMBoJZd8wuOrh2EHEuADgDwKjCbTyAMwGcBoADgOpI2HrerGmW4zNiE8C0VyYmNZMiTA9x7gBZVLQAcIGI9WEAJoHYsBrBJYz46T8frJl1x2zr8RqziclAc9AUd6fDPqqsLPQmQ9PVANRkMv29bTt2XeF2u6YFSnyP+Lyeezo6u+5NJlMfH4Px7oPHLeDx+68JUuX+amSyXuiGB5IkAPCAZRmwrAWABkPPgWEYMCwFlrNAU1NQFCdYNgFAuPm6c7Vxp1exF8x+LGlG001MBicnur0yMRkIBJeAGTdd4ijctYGI9HEAfCBOrwZAAYmoVxSO2wgAtRfV8isXL5eP7YhNANNemZgAwMybLxHQHS0XQIS5D0Sj+EDslly4lQCQAHivun12GEB08a9fzh+PcZsMDIIgDLHZbBcAoHK53HJRFLce7zEdDw4aQff7fY8AyDc1t0xOJtNPut3On1t5vjKZTC3fuathcjKV+k0o6P9zwO+bcQzGuw+P//Ra4cwxVQHkJS8UxQFF9iMnhSArlchLQyDLVdB1PwyEQNFDwDGl4JgAGMYBigqBODIsAJw5psr67VsvNFfkTEwGKSe6vTIxGQiqa8KcQArAMSACPQLACRKN8oGki1YA8BeOKQXgBYDC60yOA6a9MjEBaiZFuMKvPIi9Gg5iq0KF+xYQX90FIuDLQWyb7arbZ4f8FQGzU9MxxOVyTS0tLb2/qqrqmdLS0vtdLtfUo3XukpKS2ZWVlVtcLtdNDofj+srKys2BQOCWo3X+wcRBBbqV50dKeemddCa7Op3J/BEAQzP0UADQdN3oiHX9ffv2Xaen09l1Az7avfC4Bdx87Tl2I5cvgapaISl+yIoHuu6FpgehqKXIy2WQlRAYxgua8kCHB6ruBUXzoGkKgBXkiw4AuPm6c23H+nOYmJgcHU5ke2ViMlBEzo4Ur+XFiJMLRIC7QJxeFsTJDQIIFH4PAICYEs1q7scJ016ZmOyBBrFJpSD2ikL3dh174eYGWYC0ggh1NwD/hItqzW06xwCe5yuGDh36bmlp6XKXy/VDm802x+Vy/bC0tHT50KFD3+V5vuLgZzkwXq/30Uwmc9/OnTvPj0ajF6dSqe96vd5f0DTdry3ZJxMHTXFPZ7JLnQ77LeWloZgg2C4E0Kko6ic9j9F0XdUkqXPARrkfzjy9ikFGFKBpLPISiQyoOgeOsQCUFQYAw6ChKEEwDA+W7gQMGoriBUWJYGzF1iVFoZ4fUlVy0L+JiYnJicmJbK9MTAaKutV12lW3AyBinAMR5wwAA72LwRogkfUsiMOLujV1EgA4BD/lEAIMAKM1ZrZgOxYMdnvVc850JaOarJjbA00Onfo1deqEi2oDIOKbBikGx6JboxQDZyKIjcuAVHcHAEc4EnaApL2bDBA8z/urqqpW0DRdMe2CUe9cN/fs+ISJ4ZL1a6OdL/55tXf5O1vOq6qqWtHY2DhJkqRYf87pcrlm8TwvAHB5vd6vNTU1nU/TNJNIJF4pHpNKpVa6XC6f1Wp1iKKYGrAPeAJyUDHa0tK2gKkos7vdzq9rur5rd2v7fyqKkjkWg+sPhihykGSyf8UwaMDIgmKDoCgVhsFB1ynQDAtNZQGbAYoyoOsqaNpm5PI8JdgAms6DfOnNfSwmJoOYE91emZgMBA31UVVMiZrgEliQ/eZ84Sm5x+8AcXxpEPHONtRFpbX//FydOWWBO+Sv4VEQ87IiqnU7lmY31C82r4kDyGC1Vw7BT00Zf6sj5K+xoXsBSKvbvlTcsGWxaAp1k0OhoT6qTriotpjmboCIbRZEhLMgNktB96KjjG5/3Thv1lRh8a9f7oo1d5gtIweIYDD4ME3TQ+68d8Zb18092wJSCwATJoZLJkwM48U/r37viV+8eXEwGHy4sbHxa/05J8dxY71e73WKoiQ7Ozv/Iopi+osvvggXn+d53l1SUnKrrusfn2riHOiHQNd0XWxqbvkmb+V/BwBSXtrQ83krzw+nGdohirkNfZ9h4Pj0s0YNmk5BkhgAOmhKgW4ABiTQEEFRDuiGDiAHispD12kYtAKGEcHQGnRdNbKiRjkdBsgXH59ubjSL5ZiYDFI0XRejjc3zAczv+XjA77sAADpiXe8cl4GZmPSB2yNQc+ZNtbjddhaAsWljVH59yTr1UM8jpkQsfurl9I0L5gZBHFcJJCtMBRFQpFgquRWvcclFC15Oz5iywGfhBKZwDAcAFk5Qxo2axTpsfmbVJwuzR/xBTfpkMNorh+CnLjv/Aa+FEzgQv8lS+KlFhs+0hPwR9l/v3XfKOdMmh8eYcWH6ov+I2HxW6DkVnKQhqxuwgkTL7SAZPzqIQKcB5EAi6ErhFDIAtWZShFu5uMP03wcAjuMcNpvthpGjQhsK4nwfrpt7tuW1VzZs2Lql7QaO4+48hIVGdzQaHa/req8MiKqqqudsNttNALIdHR0XHOlnGIwcVKC73a4JoaD/ZYamwwCg6fq2lpa2WelMdjMAhIL+PwiC7fy6LduOeZGGRFLEcy99lL152vA8ABo0nQcFGQZSMAwVFKXAwvKgKBq6kYSuJcHyGYBSoKhpAEkYBg2HQYGiDADG/Y++ajojJiaDlIDfdxV6p/QCAPwlvl8Wfv0OAHTEuhYfy3GZmADEGf3S5bUWALTLIzBz5k21ut1Cr711DdGYNO/aJ1KbNxxaj983n3tDmnnzJV3+Cr8PQCOAMIjgVkBEu1z4XQYQAxA9LTjLaeEEHiQlngeJTqkgDrE2vHoq29pZL21vWHHIiwYmB2cw2qsp4291FMS5HWQfMI1uX1L1uatt42pmyWb2xalLddhPXXpFrcXttjMA9NdeXSfvbc/cHgEPPjLXPmfeVHtcgr0zb9BZlVJkDdaUAjVHLE4byNziQGxTAkAXiEgvbsPpBKAHKv2n3B7lYwXHcWcAsE2/oCaOPuxVkekX1MS3bmmzcRx3hqIoH/bn3Llc7qO9xTkAtLe3f5NhmB94PJ47AoHAvzOZTERRlPbD/xSDj4MK9FDQ/yxD02FFUT8FIHMcO7GsLPSa0tB8Vl6Sjvsf67l/ruu4eXJ1FhSVA8cKoKADegpgaGgaBUVNgKF0cJwIRe1EXsqApmkYRhY0TaILuk6BYaT7H3019cq/PzUdEROTQYq/xHdAR9Zf4nsZADpiXWbVV5Njhtsj4KmFt9ovvbzWDuLgsAA8hacz6LG9qjrs55e8ucAzftRd8WRCPKSUzZWLlyeuun12O7qj5wEUxHaPw9IAtrZtyWZ97moHgCqQKGhPpMLrHZFhM1kAsXgymutKNhzSooHJgRls9srCCQj5a4oFuhzoLuLVE2dVaS3aYvXR1lidCgDDq6dyPle1xecO062ddWpjyzrJnEsnJw8+Otf237fPdKKHkLv3vlnG66+uy94+/+lsMkG2PxTEuQMALAxUALDQyBgG4OORTVKwZBQYIPUySgunEgu3LIhg7wCQAoBoXdSsmzFw8ADgcPIK9v2+76Hw/J7j+8meSDvHcW6O46pEUdwsSVIWQDafz39/xIgRX3e5XP/R2dn54uEMfrByQIEuCLYhDE2fIYq5PxTSsODzeq4IBf3/KCsL/qmhofnSYzPM/fP+563Krp1tW4eUe7zQOBUsYwXDJKDLDAxDgabJoCw6WJYGx9GgaQq6LoGiOqDrKgADFJW5YNZjTe99sEU56BuamJicyBgAKEVR31EUJVl8UBBsFwCAKOZOuJRRk5ObMePC9EOP3OibMi1STNVUQcQNDeLIOEGclCyIUNfdboH779svsf78gZdzh/JeS597Iz/hwtpEdSTsBRAF0A5S1d1aeO9E4fHsp2/UySwmDMO+DpcFZPGABYmIeqaMn+8EIGfEWOy9NU8ku5INpjN8dBhU9srnDjMg86LY8qpniz4WRLgzPne1Z8aU7+dlRZRlRaQcgn+PWAv5azBu1Cxje8OKzKpPFpqb1U8iCuLc1cdT1KWX1zqeWngrs2p5nTFlWsRx6eW1xX7mksBC4WhogCFpBsWpOni3Bfm8ipxqIAdiFy0gYryzcD+JwnYdMSWq9WvqTP994NgOAFu3tPlRWBDpi8Lze44/VDiOG1lZWbl6586dvh72kAXA9RVlP9k5WEqIDQAURd1dfKArnliSTKa/b+X5mWVloV/QzPEvff/V361qNWTlcyMrxo1Upt2IJ3dCzO2EonbBZuMg2FwQbDQsnAUUxYFlZTBMBrKSS3Sm275y558aTHFuYjL4aWuPXaHp+naOY0eLudw/oo3Ns6ONzbMB7AKwq8d9E5MBxe0RsOilOx3vr36gdMq0SAVIFGgIgBEAhoL0JXeiu/dvsWc5DwDXz5tqPdT3FFMifnbTg+0NddEmEOc1B6AZwDYAdSCOk7xy8Yq21Od+HkS496Rn6nKx53AxGmJxCP7AZec/4PO5q83e6UeBQWivDBChJGBfce7s8RgDwGPhBK9D8AdB5hnT43lqePVU51ljbjTb2p4kVIf9VCFyvj9sl15eG3nwkbnDLr28tgLE/kUAjKaAIWUCWAsDi5UxMhYGIg3oVhY6gDiA3QA2AdgAsn2nA921NIznH/pzWkyZaz0DhSiKjbqur37tlQ1Dt25p6/MPvXVLm/jaKxuG6rq+WhTFxsN5n3w+v0HX9Yby8vLHBUEI8DxfUlpa+iiAbCaTOaEWK48FBxTXiqJuBxB3u533lJeGHhEE29kA0Nbe8VhekpY6Hfa7rTw/7ZiM9AC8X9emXf34e9sTorwJhtEJCyfDYc+jxMu0ytSQ97d0jX53ffO43W2pEAxYwDIyaNoCmqJ+8r/Ldj/34gdmWruJyUlAVzzxr+3bd41JJtO/9Zf4Fo4YNmSZINhqjve4TE49Fv3tTmchpd0B0q+3GKnmAIRABHAxCkkVHqcKj3HVYf9hieCCSN9dv6buMxDHNl64ZcWUKP7zqZd3vfCTf6QKInvv1OmiIPeBCC5vj/EX94IK50+680COuEk/GWz2qjVWp8uKaGDfFFb7XvcVkHnMF56rBuAHmVclIPMMkeEzHT2j6yaDl0uvqLVg//uTrSALlFYQO+hCd0s1KwCPhYG93A7NzYPjaUO0c0ZcYJFqqIu2Lvze05+uXLyiEd2t1QCQyPnC7z2dWLl4uVkcboBpbW29G4Dzm1/90+nr10Z7BTTXr40q3/zqn04H4Cgc1y8URYnm8/n64n1d15X29varOI4bW1lZ2R4OhzsEQTi7tbX1PxVFOeUKTx4wxV1RFLkrnvi2z+v5o9vtvEdRlYwo5lZrum60tLTfUF1dsZSh6YnHarAH4pW1jer7ny/e+c4fviqPqwlPrN/WMfJn/7vypo8/ax1S3MRHAcb4mtC2791U+/wZp5dnKZ83d9X109gnX1xzXMduYmJy9NB0Pb+7te2niVTqz6Gg/4lwVcUGEIdx2/Eem8mpwZRpEXbKtAhp/0nSxXs6rkVH1gayN9yH7qrECkiU0t4QjR12wVIxJRo/m/dgqjoSztRMinB2l8BE66LK+rfXKQBQ6o8Ur/1ZEAEOkIWCYsG4PZFOkHRSFsSpToNE0nmfu5o29xEfOYPNXm1vWJGJDJ/ZM7jDYd9gT7GLQM8CchaQuUSDiHYWQCrkj3CZhhWmwBrkFArC9UVx0dEKMlds6J4zRfvBAxBYCrmAFYrfioyqQ/77r19OPvLAyyIArFy8XFn865fFCRfVWuwugYrWRbX6NXWKGTk/NmQymVXxeHwu4H3utq/+qbKs3LOtrMIttTQn+ZbdiREA1Hg8Pi+Tyazq7zlTqdSzfTy2IZVKTeR53g4Ahb3opyQHLRLX1h5blM5kV9oF28XZHq3U8pLUtX37rnPcbtdVgmA7a2CH2T8ef/gGx5nTx4b/9Jc1Fz767OpbZUXjBRuHkUP9qN/eAUlSqfX1bSOvXfD6j0eGfV/ceE3tH6+/ctwzx3vcNfAyF1OVQhWcHAA0Iq28ZTSJ9Yib+/xMTA4TUczt3Lmr8Uqf13OJ02H/GoCG4z0mk1ODOfOm7kkL7+PpYq9yR+H5YjRCKzyXBYBVy+sOSbT4KwJUzaQIF6j0U9G6qF6/pk5pqIvqDXXRvvbu6aAA3UpnDQoOSjM4WjYoGL1Slw2QgkzF11OFMecAwMLZWXSnmZocIYPFXm1vXJGIDJ8pAhAMlmIMlrJCNzhaNVTo0EHmTHEu9/Qx9xZwPACrQzCrb58k7G+xjgMR6G50t3xUC49JKNiTwuMWADkKsHI08h8sr5P2smvaysXLJVOUHx86OjpezGazaz0ez90tuzGzZXciCCCeyWSeTiQSj4mieFh7z/viVBbmRQ4q0AFy4RDF3NN7P67putoVT/y9K574+9Ef2qFx83XnWm6+ZnLJU79//8tP/WXdnp6iYk7Bhs9bAACBEgeu/tIYfPxpI+q2tp/240ffenjpe1/U0DT9VV3XD6la7tHiFioiTEFZAOR/YQBQ/LDK46mA7S00pl4wtpqtSkxMjoCueOKNrnjijeM9DpNThx7p6UUR23OvLYXulPbiIiwDItSLIlh9YdGKfhfFuXHBXNuMmy8pFp4DAIgpUXv+oT+n+kr/ZGrK2ezpNhd0WJisplMaKEo2XGxKs9B5vWe/dBE9KswXzt9TwJscZU50e9WVbNBjUmOLq2rYcIOFjdJgpTTDplFQKMlIsWlNBinstXfae1/zxZoRY2Yg4iTgtVfXyffeN8tA39tmHOi2e3rhlgeJqlMgdobu8Vp21fI6cch5tfytCy9x9DynmBLV5x/6c9pMaz8+iKK4XRTFbx7vcZwK9Eug7w9BsHnDVRVdANDfPugcx/Ecx3oBKFJe6tKOkjD+xlfOtz/0q3e//ad/fPI1ABAsDM4b4YZgYRDLyFjfkEFHZwadcREPff8SfLalDb/70+rMh2ujXy4pKfmko6PjV0djHP1BAIvxCFhqKI97CsoqQZyhnquPKoDUxahy1SOhfYIOs4CdickRcjj2ysTkcGiIxvQp3XdVkEhRMapOF249w0B5dAsYFYCaTIr9Ei7zH/66/bxZU4t73JnCuWXBJajzH77VK7iExJvPvbFH7Fecd6HljPl3evMZLW9tVnhNYFJ0TtdpA4zmpAWDBsOIenEcMewbGVMB6F3JqHldGkBOVHvlqzmDwawxmrpVytKSbhgMcrRC2mQZAsUpForlYmoOZN70FOl9zReqsWWdOY9OAjZviOqvv7oue+nltY4eD1MgAr1oQ2SQhUmA2DwKxC5K6BbuAJBb20EZM26+pNiuzVY4ThdcQm7+w7eyAOKmSDc5mTkigX4ohIL+r7vdrtsYmj4d3av8mbwkvd3WHvuuKOa+OJLzv/7Otnv++s9P7wKASUNceOCq4XDxDERZQ2dWRUdawsNLG/DSqxsxangApw3z48HvzVTuvv9fjQB+nEgkfq8oyoCX8Z+CMsscaqRLAGtF7/6zeRRSG0H+L24A8SupofZPjI7EQI/LxMTExOTosGp5nTRn3lQ7ukVJ0bbz6BbiCsie7qIQVtG9D12qDvvZzRsOLIJrJkXY82ZNLQNxYO3oXgTQC+8pzrp9NrX+7XVtseYOgxPsiNw43wUAmoOR5SDSlg7FoTmZjKHoFiOHjMFQAi0bEqUa8cKYeqIAkOq2L83Kiplmeipyxvw7nbqV1nND+TZrs+yhFMMCWjcoFZzBQtMdNKVbKHCdmkpLuozubRx9zeWMOY9OHm6f/3T2qYW3UoXimAAR4yqIMGdB0tl7VvvPoVucpwB0AcjHJcjDJ0aK3SRK0XurkAEgdeOCuVi/bG2bme5ucrJyRHt/pLyUjDY2D482Ng8/0HEBv+/Lbrfrjrb22Hd27mr01W3ZRn2xdQcTbWweI4q5d8JVFW9yHHfYVWEddmHEi698eg8AjAoJeHruKFR6eAgWGkGnBaPL7agsEfC9mdUQLAyefXEtOI4BRcH65WsnJgB4bTbb1MN9//4yBWXcLVTEI4BlQFJ+uB5PW0EMVxEGgK0ajr72MJqYmBwi/bVXJiZHyguLViibN0bzII6pCuJUZkB696ZBoueJwi0G4pyKICJGB5A9Y2z4gFXcBZeAGxfMDaH72tGzsjZdeMwmuISSmkkRKwAEJ0zmWKG7mJPiZWRxGK/LpaxFDnCq4mMlqYxLSBVc0mCovcW5BAoZqcKSy0ZsxpAZV/A2f/CEieyebJyI9spVPYyy+YM8AGgOWhOH82mpjMvKIa5NDrJ52c8xip9TpQqLQyrj7FIZx+g8JYHM/Z7oAFJdyYb0Mf8QJgNGMiFi3jVPZK6Y8VDn755ammmIxmSQhcLdANpB7ODe2zZ1kPmxu3Cs1pChdJBFx0rsW8eDAuAWXELlhAsnHnIrShOTwUK/BDrHcYIg2MZwHCcU7pcE/L5v+Xye73Ic5xfF3I4DvV6w2S5ta4/dm0ym3s5LUhIANF3XRTEXbWuP/VoUcx9yHDv+cD/E7CmjvmIY5Es8e3wANEXB6FG6nQIw1MfDaWUxodqBppYkGpsTsFhY+bTh/nzhM1Uf7vv3lznUSBcZEWh0t9TpiQW9RfuB2laYmJj0gSDYIgG/7+6A33e3INhOKz6u6brOcdx4u2C77HiOz+TU4IoZD6UKIj2F7tRNFcQZjQJoAxHwPcWLAiLa9YZoxwFT3M+bNc1aHQkXUz+5/RxmB0DNumO2BwBsgVDPrDkagMdgKUZ1Mjk5wCZVF5PWrXRGDnCZfNiSNBiqszCemOpi0pmxgpI8247qy6501dw43zP9sWcCkRvn2zlh7+3GJv1lMNmrwuIODVLgy2ewlFMpYS35SosnN5TX5CAb1220pPM0VC+T0JxMPjfMmtZtdAJknmfRHSmVNmxZbIY/T0JWLa9Tf3DPn7O/f+qNBHr/z9MgAr2nSJcAtKBQLO6FRSsyb71TZwAI4MAaxT519lTPQIzfxORE4KAp7m6365Ly0uALIK1iOmOdXdd7vZ6FDE0PKR5jt9m+v7u17eH9nUPT9V1Oh/3STDrzhqbrvfazWXk+yFv5M0FWzw6ZKydWsaCYM4r3h/pJLR5jr3okDE3BbWPhLwSkE6kc3E5r7osdHQYAsCw9oCu5NfAyAtji37vYT7Z4v2ekgkd3KhgtQjV7tJuY9JNQ0H+Dz+v5EwopdP4SPBTr7JrbEet6CQA8LuftgmA7vyPWdcxqTpicmiQTojF90n3JS6+ozU6ZFkmfMbbaPmVahANxVgUQx1RGd7qniu6icfrrrx54b27tRbXFdPm9e1L3hAHA+Cv8XHUkTKnZTM/rb7Fiu6X4u+plaSal2RnoWV2gM0qAFS2tSlZ1M3xuOC/kKy1JdLdjowFo4RlXsDZ/kEk17JQB0Go2o7Wt/0jOxdrNInIHYZDaKzeIQC+mKltAFoJk3UandRs6AOQBNmGJqQLXqdqlMs5m2yH13KpnbG9YkTb3n5/cvL5knfzgI3OLheOKHSH4wq0YgFIBxDdvjEq/+/VS8YVFK+TvL/qBAJIZdCComkkRJ4DWgfsEJibHjwMKdIammfLS4DMgBvkzACF/iW8pAEoUc88qitridjtvc7ud93fFE3/MS1J7X+eJxbp+Xl1d8dppI4ftEsXcGhT249EMXWHl+dpkMv2QKOYOq+fnuLCP3ZXpjjI3xfPAUBeKtecMwwAo8rSFoSDKxP/xbJs73AAAIABJREFU++xqoMSe+tdbdT4A+ozpp3206G8dhzOEflEFBwtijOwgF7cgiFPEg6wmJkGcs54rhton6MjBxMTkoDA0zfq8nicASMlk+mkAstvtvNlf4luUz0s70pnsuuM9RpNTj9eXrFNfX7IuDRI9gtsjYMmbCzxjxoaLwrq4CFtcuNV+8cDiZDJx0OAiDSLwD5YJRwNQBKfA7K7fXOyzXizOJKBHhXmDhq66mYzBUbLBwAYDstGWUnJjg7zsZxOgYEfvivQAIAQnTDaCEyY3o7DAUHPjfH3bP19IbVv8lwGv6zJYGYz2qmLqhTaQPcEe9PZnKGBP5XYOwC4AkP2sqAm0wiY0K9+sGLSka40t65S6HW+KrbE6M/hwktMQjRmFwnFeELujgQj1PcZt88Zofvqk+5I9X1e/pk6rmRQ5mF0zABg1kyJU/Zo6czHQ5KTjgAKdt/IjAJSnM9lfNjW33G3led/QIVU78pL0TrSx+SsAoOla3Of1/EIQbBPzkvR6X+fJS1Ln9u27zhEE20SrlZ8AkroiAXi9paX9hv0J+/5S7bNu/WhnCgDwzw0xXD0+WNTkMAxA1w1QFIWOjIKV25Lw++zGtLOHdiz+16b0ex/smHBupGyJkMkOnDoH0IiMDiLKQyAOjgBiYIoF4awge3T2VKWMIZ95wdhqCnQTk37AW/nhAALJZPonu1vbfgQAXfHE00OHVH1YVhb6q9LQfNZxHqKJSXGfZnLRS3e6xowNW0Fsvx1ESBsAMtfPm2pbubxOWbW87kBp7jrI9WLvNm5IdaRcO9Zvq80mRR9FU60jzxr5BoD2VMMOvX39R2JwwuRiD/a9xTZAQdHsdEqz05CDnFq/9rVMTWA+B1I3Ze/j7eiOdHlB9tMDAD3iqjkeNZtJ7HpziSnS+2Cw2SubP0iFJkyuQrff6ACJoFOFn0Wfxgki0i0AZE2gFU2glRXP/TzeVb/JjJifIowZF6Z/s3C+s2DjWHRX9N8j0DdvjOaumPFQX9mrORC7tmf/uZgUbdvXbRuf6kyV0gwtV9ZUrqsaXd3QUB81xbnJScnBUtx1ANA1PQ8AeUnqSmeyv1cUZU/FdYZmelaN3S+arhvpTPbjdCb78ZEMeG9eWdsg/+Eb0xf/bV37nQCojU0ZPPlOI759QdUeka4bBlqTEh789y6IsobpY8q7fvLoW9Sr79RPCNosLf8zqur3zu1dFcNrx3bev7EukVKO7Brih5WaQpVZa+DlAKAece1iVLkBlKPbmVFALnBG4XcegA9kv58hQk3/2FjTIkI1jY+JSf9QAYDj2PLiA3lJ2t7WHpsbCvrfKCsLvszQjO/4De/QGO120W4LxwAwPkuk1CO1SyYnDg3RmDF90n3J99c8YB0zNhxCt/OaBqBUh/22JW8uYK+Y8VB81fK+I43r3l4nFYq/dYJENVk5L3NvLXzzqxve/uRqQzf2OLfvPrfsgVwu99LV37rmntHT7BpXJnp3dlgq8zJFyxol52Q6r5N9YWphDEUszvAwDoWipXsNgUXvNFQHyD7TPb7AiFk3OJtXLpMUMQuTfRhU9mrErBvsrGB3gvgrGrq3ZhR9lGJtHRVEtBfnQ5ED+ognAucEShiQz6V92NFp9mc/TMaMC9NL3lzgc7uF4hzJg9iLAIitygFo2rShIeN2C9g7W2jd2+vkq26fHQMQ1HWdW/XiytkfvLTyK6qi9mzhhuCQ0AeXz7/yjg9f/2BjQ13UzMgwOak4YJVYXdMTHq/7yzabdabTaR+fz0kfdcUT/8hmxU8AYOiQqoUOu/AtAFJrW8fduq7vXZ3xsOE4zutxu74NAMlU+leKosb7Oq4tmTeun1ydMCimdGt7bjwAfNKYwQc7kjAMICNpeLs+jof+vQs7YmR426Jdwhc7Y45pQwPbHp502h+duzNJqLpwusflOL80oK3vSmQ7pMNb9J+CMu671ARfDbxWP6ycH1auBt4AB7oozouLIsV0n2JxnzTIBWxnAzIdTxkbO2LIm+LcxKSfFOzVHN5iudjjdl2m6XpKkqTPcvn8dquVt9kF21yGoUMAEOvsuv9ovnd/7VV/mFkeYhdNmeT+5qjhrmvClbZrwpXC3GHVNp5h9I9iXaYTchLg9gh4c/mPSsaMDY9Ed+s1A937eVkAypRpEer3Ty3t87rasnO3OvlL5/CCS9ABUIqkUC/88PlffLF6y3/CAE1RlEZRlIxCYVKO404XO7NXnTf7nHdCfk7OybQboKwURWlWi8HIGpXSDSoD9CrgIqtiNmvzh1zYt5qyFb2L0ykgjvieOUpzFjrb0iSnG3ae8OLsWDPY7NWEb//ABSKwGBS2YqC3OFdAMiPzIIEGFO4bqpjVPnvutyfsKs3M8hD3yvnneuYOrXYWbe4tI4ZYCzbXFOqHyB8W3eYaOaq8aC9okHkTBLFrOsgcYs8YF7bOuWkateytTfn2tuQeu5OMJY3I2RHNXxHgXn38lbs+fnX1LQ6fI+Xyu9KqrOq6prMA6GwiW9XV2HnN1x75+iehcLBt04qNZl90k5OGA+7x0HRdb2uPfRlA0srzs2mG9vZ83srzVwPQYp1dX1UU5bAd0iPllt9/kP7J5UN/evYQ10fFxzY2ZXD/azvxtUX1eOytBjTFJVxdW/3Fty+oWfv980cvf+nayf98aFrkQ0dnTjck1QWy2usZ7XYN/eXEsYe1al0DL3MLFfGiO53HDcAPkvrnAGDJsJrlXX/ymv+rbn/w8eG7f/uHcNt9bwcSV8c51QqgeRVakj821sQbkDEdGhOTQ0DTdaOtPXaDpuvbOI6daOHYmuJzLS1tPxDF3BPoLsJ1QjKzPMQunFzrqxRsvcSQi+OYuyIj3Y/VjhWO19hMjh4PPjLXPmZsuAJ9V18vVsl2V4f91inTIn0upIspEU/e9ngi1hzLA0h8+I8Prmqqazyn8DRlGAZrGIYVPRbiM/HM0H//7+v3AoCd1zOCRVfcNg08a+R8dn1vAW4AENvXfSSh724ie49LQR8+xV6V400KDEJ7xaB7Cx4NIrSKYjxbeI5C7zHTALDrzSUnbDu1a8KVloWTa70ujuv1XXRxHHtXZKTnR2NH77sNxGS/VIf91JRpkZ6FKz3orlnQEx6Axe0WPIv+dqdr7/M8edvjqTf/8EZk07sb57gC7p0ALB0NHSE5J7sMw9hje+Sc7Fv8878/fv61FwybcfMlZltik5OGg144k8nUe6KYq+I4dqSUl3oVcmtrj80Rxdz6vCQN6P7tg/HD2WMFwcJoz9xU8+P/+6Bl1rMftl7VmVVCAEABxmnlrt3fmDbyg3ND3k5d0ngto/KGqlNKc0rUs4oAoAzkYqMDkEa7XaFbRgxJf9jRpXyeTPU7in0lNdQBYnSKRVMAkhZoBeDabs+PWVLadZdE6y6HynSW5S163KI413oyV613Z2bM7PAuiCXz9Ufzb2NiciqRTKbWZdKZ03grP1JR1D2FZzRd16KNzXdZef5hQbBNOJ5jPBCP1Y51Y99WjBoKnR2uCVc6/97QLJnpl4OX6rCfmjNvqgPdae19QYNcN6xnjKtm97cXvaEuqt99wZ2J82ZNy8a3dd7Y6wQsg6rRVZrT50RHQwfVtqOVBoBta7delmiN/87l8hu6QSGdp3lJNfIAYOUMLq9QxV7sKQBK88plii9yRjo4YfLei0M9F5EVdPdw74WazZhzdT8MMntVnBM27JvinkF3HYUMukW61rxyWfpELRbo4jj8aGzEhe6uOsUth3vm8S0jhjhfijZJnydTZtCkH1SFA8W6BEB3h4i+goHF+WOtDvttl15Rm3l9ybo92TdiSjQ++Puqr3Ecl5ZEyStl8y4A4AXeCJ8xxGAtLJrqm5DqSNJdu7sirz7xyoJZt8++683n3mgZyM9nYnKs6NfKtqIokqIom/d+vCueWHr0h3RoTI+EmJumDXcDcNIUlb1lSvkLt0wpf6G+VayQOUso6LXl/JzVIcVhV9NyhSHrlJaUDT0tSelUntuZzY3QDIMptXDBAMduB1ACwPujsaM5AHEA+kvRpuzjdVsTTWLugGJ9FDw29BbnRdh2XqlaXNb5PdoAP6XLFfPJLNdkkzXOoCwG5OYUq1W8EYw/8h+qd8ctWev7q4zWbD3ipmNjYnKIaLpuiGLui76ey0tSW16S/n2sx9QfZpaHWBfHsei7GJeGwv7k/6qusH3Y0ZnZ5wQmg4JCq7X+RHo4AFxDNHbQReL1S9eWV1ZWhov3q06v1q/8zizNHXQXr0VGw+ao/s9HXmbSXWkmujl64ZgLfB+qGuUWLDojWGg2mWM68gpl5BUmAxIZNZpXLhMVMYtdS5d0BSdMdqH3vJRBFhEUkLlpoEeR0+L7tq3/yCyecAAGi71qX/+RFJwwOQMyByzoPYeL4j0LkDZrqpiV1j/5YHtX/aYT1o+5JlwhuDjOj3194TzIQoMBgLomXGm9f+PnZs/2/tHTXlmwf53R67gzxoa5ngIdADiOOw8UeCmb5wBg/CW1+kVfvVi32CwUABi6YXzy5np96e/+zdZ/UHfNOVefu7BmUqStfk2duZhiMugZ9KlnN08bXqyAS4HsdxIAoKZUaAZNt1CCxZFrFqvU3ZJXFfWcLurszkTO8Zemjsi7HYlhimGwKLw4Ilh3fb3U/+/TBetOkKh6HYDENeFKz8zykP/+jXUNL0WbDrQS7MC+4pwB4Hw7kLhepQz+7IQTO4S8f5VPBgXAptFQaMMNAAbArQvkH/5Gtuy6KVSZtAotLc8YdaYjbmJyCjDa7eJA9uq5QOxGsUp3HkSslQJInxMooWeWh/IphfgyjVlRO9jiocmJg9sjFK8RCg4s1CkAxuYN0YMKXH+F31383VvqNa7/0Q1a0YktUj0mTF31P7PVRd9/jvti9Zap4y46c5vLptmzMm3xCBpvs+iy3aKJNM1RHSnWSDfsyNc/vzALAF31m7T29R+1BSdMdoNkiTEghZ56jk3EXhH06JtLMmY/9JODXUuXZIMTJvMggQsN5H9dXLBRQfyvtsLzRt3zCxMnsjg/J1DC/Gjs6ADIZ3CCLDYZIPNaAvGP4wBwusc16H3lY8Wq5XVaMilqhQJxFPreHgP0th37O8ZRlPE1UyL6pbd9yehxLEXRFDXhklpdTGTx/vPv0Rvf2fif1ZHwqsjZESNaF9Xr19QpYspcVxmMCIIQAuBVFGWHoiinZG2BgxqdyKgRy0EcxgNSt2XbmUdlRIdIOOBg0O3k5FCIOgAAdF030tm01qXYDUX3iRlVeKyuefJbbfGRxl4GwQDwuZgfcvfO5vnfrwz9dbrb8TGISDcAxF0cxz9WO7YKQMNL0ab9TZa99+RRACxtvFLeaJNq/TKnb7PnqC6LSk2OO7WJCYdm1WhGonVmmz2PlSUpLcmpoz93ipNHp4WPpqCsKkbld71i7DRbrZmY9IOjaa+sPB/grfyZFo4tAaDk81KDKOY+1XT9qEcEKwUbdU24MgCSwVOEBrHRHhAHWAdgrRRs/MLJtSEQ+5IHIH8U6+r8ztoNSVOon7iMGRemUoksNm1oUNEtaGzYfy0Y7YVFK5L7i6D7KwLUrDtm+yZcWDtcldWqX3/lCQMANXn2OfuIcxSudxU1lbS3zIfta7eNz6flf3AWi+5gNFWiac4A3ByLjiF+hROSu7V1iz+WzxvzNaesiEZrZ71c/+zTGQBGcMLk4oI4QCKmzsLn6eUJN69clql7fqHpHR+AwWSvuuo3aZsWPpGI3DjfYAW7BjJ/vSALNgqlGp1sStMYUfegIZnwx1xMUvBTGfHgGSDHg8dqx/oBVAAYDuIzFheXVJB2gTYUbGylYDPOCZRkza1F/eP3v16aufe+WW6Qv6UKMkd6YoBcu4qoANTqsJ/ay941g/yPMPX66Tr6EPy6plMTLp2I959/D831jRff8sStfwfxxQ0AUqw51rpy8fLOpc+9kTfF+tFDEISRIP+bZlEUtx6t8/I8X1FeXv4XjuPOBZkjSjwe/1ZHR8efj9Z7DBYOKtAVRW3hOHbqsRjM4ZDIyj1X1ACSaicA4GGAVkS6Ehznk3VZuOeTnRdvSmbLRlh5XBvwGGc57EZC1fBOMk39tSNOKYYB1TDYnze1XXuajd9cZuF4kOh8HsQRsf5obMSzdHdb+35aHsnoXfTHAoD6dyj+VQNgVMoQE5wqzGz3quOT9mLvUFA6rY7K2Fivwsp/qeywvetPfv+0jO161qByF6Mq9Ap27jrqfzgTk5OQo2GvGJoWqqsrnrHy/OWarn8h5aUkzdC0v8RXBcAR6+z6Zkes6+9HacgAgD+cU+uqFGx7OzEAieo4QZzHBEjhSSeI7S4WbcpO9vtK37jwvNZrl6/eae6VPHFwewTce99sYc68qTa3W2ABIJkUixePPEhasAvdIr3YC90AkHhh0YrOvs5bHQkz3//TD6oFlzAahYraLr9rdyqWqqgYVQkUou97v06RFLpqdBU2LuuytNYlqjzBQNrQQVOUlnF4KMlp0ZTELoNSW1xDxgyb0aErxIkeXj3VnhFj8nvPPpHctviFXLB2Ml8Yq96+7qN2Z/VQNlQ7mWcFO5WLtWvNK5blT+To6YnCYLNXzSuXKV31mzqH/scV1hGTvyQC6GzduFoNl4x3sHGtnM7rVianczDYinGjZknjRs2Kb29Y0fLx5ucTsiKeEEL9nEAJ81jtWG+lYKsBUA2StQSQBQcRxLZWg3w3BQBNlYJNfXHq2f6PYl3i1z5clzZbXh6Ynz/wcv7SKyZwY8aGOXRntu7ZboPurkUA8ZNt9943y3fvfbPQEI0pf120IvvzB16WMtnMqw674785K2cEhwT7jMZrikZRFIXQsFJ0NnVWg2SaASRCn/FX+Idfdfts+4QLa9t+dtODSVOkHz40TVNer/eWkpKSHwAY0uOpXZ2dnQ/G4/FndF0/ou95eXn5rwHQTU1NFfl8PhYMBr/h9Xr/L5vNrj6aCwGDgYMK9G07dl0XrqrQBcF2fVt77Py8JDUei4H1lyXrGqUrJ1btXbAkC0BUKVulZhhWHXL2hS/ah29KZsvOcgi4syJohDiS2e5kaNwc9OFMu8347q7dlGYYkA3D8lIsMeNb5YG/onu/VRYAXBzHnxPwcUt3t+1joWPIp/yw9qyay3RYlNJWXq4FgASnCgGZ08cne9cGYgyK0gDdpTK2URkbtjhyw3YK+ckjs7ZPBLB0DbyMuR/dxOTgHA17FQoG7gFg2bYjWqYoSq/qw06HfWxZWWhJOp39MC9JzUdjzDPLQ8xot8sKInjy6O4tbQNpTVMUb3YQcd4J4vQAxDZJAODiuPI/nFOrXLJsZYPpQB5/3B4BS95c4B4zNmzt9bhbKBb/00H+d0mQSGSxyJYOoHPzxmh7X8XhBJeAb//mroDgEkagx4Jw9ZjwZ5vf21QRb43TwaEhg6KoniLdUCSFzqVzlJwnCWCGxrBqDk5NBnQFfDJquNUcmEwrtoCCbLHDyCcggggV1SH4LTOmLPD+6737urYt/kuvrK5Uww65eeWyUzIN8UgYLPbKIfipqrJam4UTGIfNbxkuT7VgeVoHAA9G+5GQykH8yZ4FwQwAwvDqqVxVWS3/5qqH2rqSDcdVpI92u+iFkyf4XBznBYn+FbvucCD21gay97y4dzoH8nlkAJjs9wl/OKcW1y7/6IStSn+8+e59s/mv3zHTXrBzxXbCMohfXGzFWBTnThDfek8Lvuqwn7v3vlmeqrA/s+jvG59INiTmqpJqF1OiIbiEXgJdlVVomkbLeVmTczLD8ZwOsuCSL7ynDrLoEqqOhMUbF8zTFn7v9+a20cOApmmqqqpqIc/zt1jt1qaxF45bZ3VYW/OZfOnGZRtCJShZ6HA4Jjc2Ns7vr0gXBGE8y7IWXddVr9d7Q2Nj490cx83s6Oi4URTFdgBobW39jcvl+h+bzXbBqSbQD9hmrUgilX4OAPKS1CiKuR193QZ2mPvnlbWNcrQj09tYUhQFt9trCPYySuCDtIN3L26InWGjaVxZ4kYJu2/XmnF2Gy73uQwDAEtR+DgjRtAtznv9nUa7XX22vfkEHTmQKNce77jJJhdbp+gAMDRr3eMw9XwtAxh5WkdVjjcAIMarp5GH4Ub/CgqZmJjgyO0Vx7FndMWTT+/t7AJAOpPdKOWlFTRDjzxa4x3tdhW/3zq62xUJ6P7uF+2NG8QW2dF7O01RpNGVgi10y4ghZlug48yceVMtS95cEBozNlwK8n/zoNA6DaRWSVG0i+hOr82CiPUmAM1jxoYtTy281bH3uSdcONHir/CXoHcVeHbcRWeuBYDN722i85k8o0gKpSoqrcgKJYkSI4kSne5Mo+nzRtgc9qzN6rVIaTikFKzZdrikNDxiJ06TMygzVFhojqJB5pmn8BMWTmAjw2aa8+sociLbK4fgp6aMn++bffEvR5015sZh40bNGj28emoEZPufG2QeF4VuKXrbJQpkjnotnOA5d/ytB03lH2h+PG60w8VxDIAQur8/ErqrzltBFkUB8r20gKRf7xGGk/0+YWa52TqwL55aeKtw732zPD0WIQ2QffztKATO0C3OWQBdIEUFi/UMijbSPmfeVPfdP72u5YIbpj1lGAa17rWPWUmUoCoqVFmFlJMoVVFpAGrb9hYq3tKFqtHVDSD/VxdIi+NiK0sagPu8WVMFf0Vgf/vdTQ5ASUnJrTzP3xIaVrrytme+9dnF82emp86Zbr94/sz0bc9867PQsNKVPM/fUlJScmt/z2mz2a4IBoM/LS8vXwJgJMdxls7Ozjsymczy4jEcxwkAPLqudw3E5zqR6ZdAl/LSJlHMPalrevLgRx9bEqKM2Y+/15IQZVJIiaYFyusOGVbrMEqWhxg5aWhLWhzWnpVcYwQrHAwDlur7+3mey2EAgGoY6FLVYtGdnnuTAEBtEnN9ppC+YuwUG5Ap9gWNA4grlKEBgFdhOwDS9q2v10q0Yci0AYtOdQCAAaN4AaCmUKV99co1MTHpgyO1V4qifuDzur/rdNh7ObUMTbM+r+cyQbBN1zV9n64WR4FiBFJBb2fFAyLqikV3DPSupt3TjvPXhCvNXunHCbdHwPtrHnA9tfBW75ixYR+6Hf7Swq0awMTWluSMT9c1XL3xk8bLkgnRDmA3gO0AtoIU2lIAYM68qfY586b2sv/hSDUHMh+KUAA8Q8YNbSs/rWLblg/rsWbJamTiGVrJK7Sck5lMPI347i588LcVyMQzGHHG+E1SksopWWQVEbqhA7oCXZNg0SSjWpXh1mSDAZlndhBnlwGAqrLavrZimBwmJ6q9cgh+6rLzHwgOr55ahu4oczFD0AMyn0cCKM7zYhR6b1wALD53tb3UH+kzuHEscHEcJvt9VhS3QHb7dQaIcMwXfudBFhyKC2klhdueBbHJ/hLzO7AXl15Ry86ZN9UJcj1yAijJZKTqjZ80XvrJ2uglO7Z1hED+5ulVy+uaQMS5BnKNKy+8hkd3azbf9DH+8ktnT3iXFyzpj17+EF98VE+nO9O0Iim0oRm6nJOllq27qbf/+DYNAGfOnLAe3YvaTOE8ZYXz8gAof4XfXFw5RGiapr1e731WuzU696GbMla7tdf8t9qtfOHxqNfrvY+m6X5py8K5z4lGo+c2NjZeoSiK3NnZ+UdFUboAgOM4V3l5+bO6rrdmMpnXjvbnOtHp10TNS1JztLH5zoEezJGQzMoJj2DhKZfDD8PwGVmp2tA0p6HqNlnWrADg41hoZJ95nyLd3SOy7mXZDIjjI6N7dTUPQPuwo7PP/FERKn5hrE/cQY11FFqupcokbiMAVOQsiRyjl3zmFNnahEN1aIxBG4V8fFajEpzKUIAhMnoMQNAnc8VV89x4BFhSUN7ExORgHKm9amvv+FVlRVlpZUXZxyDRk2L6XUjT9brdre035SUpdlQGC6BJzBVby+RBnJQgiDNY7MmL1Wkx8kY8df5mMVcm6QYz1Gppnu52vH+Zz/22haJ6LhjSlYLtuDnBpzoPPjLXPmZsuFhcikX3/nIaQPizjc2RXz361mWbNjSebhgkKscwtHTe+ae98+17Ln6orMKzA93XGwDAnHlThRcWrdgj3gSXfe+Llw2F7iWXfevyF5+954/fWfnX5bZURxJVo6vhLfOibWcbtXHZBrTtaIWnJLg7MuG8jZoMa6ylKbB14/rTY60NAUWWGZvNmyxxj9xcVXlWXs5YMiApvgARKjqAnEPwn3AL9YOZE9VeTRl/q8PCCR50R4/5wu8CurP6vCBznFe1PNfatXpiPPPFqLyc8DO0Je+2D/k85DvrdYe1fDcAOuSv4Vpj+27bOBac7nEVFzh5kO9YtjD2nq3AhDZFrXgpFh+1NiMOiSma4GWZjrMcwqrZfs9zlRauEUDqdI+r3wLkVGHOvKnFgpfeZCJnf+qXb9++9LVNVyuKtmdhoyrsq7vj7ot/OfX8yFYQAV3c519ceM6ge6HaztLgRlbYm2Z/Y+ZvX3js1Xvf+N9/U2MvHIeq06sNzsoxbTvamI3LNiDe0oWxF525esTEEfvYT5D/t4B9izib9BOr1Xo6gMqxF45bZbVbrX0eY7dax144rmnNktVTrFbr6aIoburPuSVJWilJUrTnY4W97nNLSkp+oSjKlsbGxosURTnligccUKALgm1KoMT3ta544rF0JnvAFViGpjmfz3Ot1+v50hdbd9xwdIe5f4YEHNSy+2Z4PYKFAW+Jw2F3QMxbDFVXoMOArttK7bzOMZS+KZujZ3ic6FI1BLl9P3q92F1UcqJD2ArikCggKVAqgMwz23ZlDlQpWYRq/NxYn/bDmqmCk/OL1reg8Z9vceYClTl+504hP3K1N83WZGygDUrXKYMunowCWtd5MkFep7uGivynhffvEMCaFwMTk4NwtOyVputatLH5uwxN/4C38sNBBHMeQIMo5tqO9riX7m5TfjR4sc/cAAAgAElEQVRWUQs90C0g6ZdOAFANQ3i0uf3ytxPpc1mKUsYI1ixP0+rnYr7yty2x//5XV+rin4bLFlRYuKKQUlKKou7vvUwGjuqwn5ozb2rPtPNiXQEACLy3rP6cH9778i2KonEAQFGU7vEKGYeTF5e/U3/JhvUN43/7x5tuGjo88Bl6OJlTpkV6XayidVH1vFlTcyDXBxoFcQ4AgXCwfe5DNz25+JGXb964bEPZxmUbeo0xVDnkw7MvmL0cCutZt3rp2V9sWHMGAFgsDoWGhUrEmxyJeFNFY/Pa2lGV191nt5bWF15aLNBky4gxc6/5UeBEtlcOwU+F/DV29C56W9xjXhTnDhCBy2VyzeVbml76sqJm3BQog7d4FVXLuTqSm6o6kptmlPkm/yYcuuhn2H87rQGnMSsWFzIZkIh5MVpejPrS7yczY37e1DZDNgwmzFvax9ltmSZJFpZ0JW96PZ66+t7K4P9c4Ha+9lkiZe5j3otCQThnR1va/fWbn/1jy+5ETc3osh1jxlZq69bs5KK7Oisbo12R7377xd9XVPm22mwcRpwW2nrJZWe8MemcYZtB5oYTZJuPDmI/Nb8V+qyrx/+FpintH79Zevva1z52rX3t4z3ziKIpddKVk9+94MsXvgPyf9z7+qejO/sMYlo06zkdOiUAYHVYW0Cyqfqk8Pye4/uDruvtPe/TNC1UVVX9jeO40zs7O++Ix+P/ONLCc4OVAwp0XdObeCs/ubKibKOiqO+lM5l/pDPZVSB75DQADo7jIk6HfYbTYb8RQFAUc48ei4EX+eHssYJHsBCHx2bloKoWXdEZ6Lqmq4aTsrKwseCmDPPH3tvaEdyel0ABlEDThoPp1r2yYRgvxRIUAHAUpVzj96wEWc0rFvHJLN3dlnm8bmu/VnFiyBsxkGo83i7vtwKBwNJGm6QDUNd5MiwNKGOTdk2hQauUYWnnFWOVL8XnGN1zSbv3J7xOZ0BSHXMx5E2DYmLy/9k78zApqqv/f2+t3dV7T8/OzLDDKJugiCJoJAKviRo0xrgnMZiNuPxMzBv0TWJiTGLia1ySN5GYxKBxwbgmKi5RWUQxoCwygwwwPfvS3dNrdVdX1b2/P6qbGZBNHGSA+jzPPNNT1V1zu7vq1Dn3nPs9B2Cw7ZVJqaGq2a0Ath7OcSd1HXc3bEv+eNIJI2Hd/PKwMpf8H7ui570aT50+TBbN71SWal6eV0QC6uS45JvJdPQv3bHht4Y7b/+/UTXXOTiSBRBf3tGt7fcf2hwWZs6uL667BCznv/hY2LkjMuani5+5uhicl4TciEbSXF8s400msp6x9ZWRpq3doe9f9/jdjz377YsEgYsXj5tI7H7LWf/qOu3yxVf0wRJDKir676JyTFXnN373zV9tXbO1bPv6pupMPBMIVpe0alFlWXXgjJweY6dv2fbarKataye6lSp9ZNX8nMT7HYyayJtprb13NYmlGkONLX//6cSRC78uCZ4k+pXl4VZCXEWonj9SmdBjhaFsr9xKabFEeCADW9rKsM47U9MT/sbWx75mmFlPwD2G1pbN0RxyCSMgSKrh6M6uF6XO2NuLOE5siyVa7vqkYztU2tQsa1Oz+WGKk8EK4rzo73cubsxkx97R1j1PIoT/VkVJYoLi9HMAdfFcpkXLv3d3e+/oX7Z23x0ShL4T/d4XhylObl/LHY9HkokMAULSbbc88+POjvi4hd85881UUqtf9ve1paZJdwXUjIGLRtLjJEnINH3YM+bF5zd+7pz/OvH5xT85716HUzRgTQIV72EGAGGYm/V+6YuT/zJ91th/v/16w7xt7zeP0/OGUDaqqnfkjPp/h2pCxWqlPbtXmLACdANArqUhrLU0hO3v7OMTB4BcOleKfSzTHbB/1/MPhZqamgcAuMLh8KS96WocT+w3QM9pWnj79ubJ5WWl3/b5PDcFA/77gwH/3p7KdN14qTca+1EikfzP4Rnq7vgVCf975cnKVbNHVaNw02C6IRNZ8kIQ3NCNIKdwMs0aBITlvzFj+I41O6Ohv3THuK+WBwGA+AWeeQUeCcNkv++MkLCWJwRgN1WXPVEtiathXdSptyOxnmXhNnVZuO2QMlN9fX2vCYJwRSAQ+CsKn/m7/rT4rj9tlmuilBYoMrxJALhP6/P8dErC9Qis8isTAFajM7fPg9vY2AAY2vbqQAxTFBEDxCUBCC1avvSZaPwzVZKIb1aEeA/POQUCkqUMKdMom6Q4jctLA11/64mN+Vdf4uKLSvx/Tup69GAnEW0Gl9q6j4gPaQCc8T7Vf/21D1+XzerO+Z+fyL7+7TNZRaWPcRzBWyub2H13vcpt3dJZKsuC2t7ad+KrL24+e/55k55F0f6vaNgtYx1p72VL/vuB7oW/vNYNa037noEy40UhdsLsE9ecMPvEXefCPd+5u1GJzazT+R5x+4dvz1AcZcbIis/B0DWPofcWnyZUBKcjm4+wrBYpa+tdcdnIys/9Af1lpwaAXHlovNwVabDPs0/AUWCv9hfIyABgmCq2dzx7iWFmPR6lBpUlMzjdzDgNVYUk+nSvUocJdV8x39/xf6mO6OpbAelvADr3c9zDym0bt6SXzJimw1rzHMUATY/fd0ZmmYzx364qRa0seQQCZjDQXt0IKRxX8t/DyrsWhzuCv+uM3PjH0TVvvDX/M/JN6zbGl4Xb7IoSAJs2tDAwfvi6tc1zP3NOfcOOpt5Jry3fEjh91hgt0pviP2zsEs5bMIX+1/mTKZi1PlxxSRuW/P4N9ysvfnA+IYTd9ssF9xYOV/xMcwBAAFLmRMpf62wYfdnUnekvTnNnTQh9GpyRHHhYwXlxiVhxfTSDNdFtwNL40Jf88IHjOuA7VHK53CYA3Rtf21A167Iztzlcjo/o3OQyOXXjaxvGAOguPP9jI8tyUJblL0ej0etEUZwmiv0FPLquN+q63nXIb+Io5IBr0E1K8x1d3b/t7um9z+1xn+1yOs8URWE0rKA4qmaz72XU7Iuqmt15+Idr4VckvHbrXN/kuoADxffAcRxxyG7wfCmR4UPWdBIOBucQQTWTjC139/zigokbFz+7edLvOyNcUOAREASSpRQded1aC0iIeWNV6b8+6/esRX+AnH6waae2vKP7E5WN9vb2PpbJZN7y+/3fVRRlDsdxJZTSSLesB9wGnzoxpbw9Ne7+W3VOahr4ughy+VdYa3Zfx7WxselnKNqrg+HiumoFlr0hsJwK8Y1EeiID+PODPnOSy6kpA4RX4obJ2vK6PNnlLOEJMV6Np065qMT/52Xh9uT+luDYHD5WrWgwb8aCgZtUAI4//d+bF/X2pMouvWoG/fYNc5ggcLv6+c6cPYZMmDSMfeWSJayrM6EAwNurt8+cf96kf8I6H9gf7lv+kUB41dMr8moq03T54iszBdGjalj3Qg1W5VUH+p1cqEk1u/7VdcaEC5Bpb39vJGOMC3lPhGHmOIF3UklwgyMiKDOQN9IYFprFbWt/GtFUw9kjKz/3f4X3kofVvxgVJfX8Bjw96J/h8cZQtVexRNiAFdgY6PcTdVhBDweAM0zVubn5rz/I5WPDBF5ho6sWMJF3wmQ68nqSaHpcpDRPnHKpVlt6tr6983n/uFGnXwmId7d2rjsifSCXd3TrDzY1R68ZPXwUrAmIHgDOFi0/qimnVZ/qUdh0j5IPCgIlhWs0R6nRqumCSmnlFJcz825aPbktr9cMk8SWu6ZN8rep2eia3uhxX02yekWDmtfIDADE53OyZ55cH1jwpWnq3HMnkG995SH5vAVT6OLbzmMASLxPZTuaejnDME/49b2X7Pz+dx9vfvmFzedfcNHUFVNPqdsKy95kYJ1/DIXJIomDGZSRCcosAwDPvdyY48bV+2DZvzSsAN0PK0jXCsfoAtCnJtVES0P4uP+eDgVKqZlMJu8EcNeTP3+i7Yu3fCnjcDl2LefKZXKZJ3/+hJLL5MYkk8mbKKWH9DnzPD8eAF9SUvK7PfdFo9FrotHonw/9XRx9HLSaoUmpmUgkX0kkkq8czgEdDNf9V72zEJwDhXV4xOuxSv0oZUQQJRBwMFkejKWIyAU4QWBn1ZeHH/U41D+v3DFsZWtf+facJgOAxJH8dLcrfHlpYNUYp7wDVnmGCUvAQv9i7TDHYJSNqqraoqrq94t/KxDIOaRGOQcVpQqEEKwbQjkKYnQAjBAc5s1kauA99OZeYa05FYbtfNvYHIChZK8OxGmlJVyh9Y8Mq4yUUkB7JZ46gycwh8mi8dfumNyj68TN82yKy0ln+9z6CFlSm3Kae4QsseZcvhpAYl5Vue+2jVvsipsjwOoVDUYioeqFFkOAZcczLz6/cU5JyI1LrjiVCHuRE/H5neSqa2ayO29/gQBAX59ajkJJ+Z23P53YWy90AFj/6jp9/avrWhb+8ht9ZyyYVQrLMTXw0cynuf61dcV1flTN9w4HALdzGCeLPiqL/VlbDiIE3gmBk01RcPO6kS5Le9K66A0yMEb4NBWFlKl1Re3y9sFkqNmrvK5ie8tKdVTtLBnWeQVYAXoGgJvxhNvZufzSXD42DAC8rjomi14GWOeQyCssl4+RvJESeCNt+N2jdQBIZztPmXnSwtGrge3dkYZ8/gjoPp3o9+qwlhFUwPKB8/+MJU8CgNk+d7YnbwhPRRJiZ14nMkcwQXHSWT6X0ZHXSVAUXAAQ040pwySxDQD92qjhrjW90eSn/kaGGKtXNGgnTR0vA8Dbq7e7fH7FuO6mc8x7fvOyUxB4fOv6s3f5rk5FYv6ggnhMlVLJnPO7N322c/WKbXWvvLh59tRT6hrR3/quKNS8N7+X/ek3T6VPOHd2/owFszhYavBxAEXBMQLrnI0BYOtfW2dX/HwCenp67pFleWZ4U/OFf7rugZZxM8ZtLR9Zwbp3dJGtb2+dmOiJ12qa9lRPT889B3vMaDR6WzQa3fW3qqpvffjhh3YbvAJHZbuB6+aPH1hekYck+sFYCXSDA2UmOE4kAqGgjIERCkI0ZjAC3czXOMXOW2aM6jInGmq2L5eNJ/PEr9MeWI5NHNbF3FX4zQDkD0fPy5NQKlxD6kMKhBJYhiQH6/uQYanyqrBatfG1cPO1cEvnkBrnnWx9vAVpew2Njc2xRXF9ZwJA2RY1N7ozr1cAwK3hTp4jQIAXWNI0uZf6kvyfu6PCrTUVmpfncyqlDo5Y2dJhitNxWmkJb2d0jgyLFj6QXPrEDUEUHMvmHRFFzeTLJk2pgeKS9vm6qafU7Xrs9ysJAPT8uXfEVq9oOGDl1iN3LE1PnTNNVLxKDv0iXhysQF2LtEfij9yxNA0AaTWiExAJAHhOZgOD84FIoheS4KG6keZydVLQ9EjFwTM+TVOpiG63FTnGeXfzI5mAr04I+mqL4l15EBiGjzc0I+WLJrbM9HlGdWfU9nIC8pEAyiEFmGHmWN5ISrLg1QGAEOKURKXyM9Ov19NqJPbWe0sSR0jLoKjxIwNQXkukzgCAJ3rjUljLCwRAUBCYSil5JZ7i/9zNC18rL9mV9Sdklzhjel5V+b4v7OOIlnCEbdnc2gsA3V3J6rnnTjAdTpH/sKGLGzWmlAaC/fqZlDLi9TppPKZymbTmqBsRcnh9zvbWllgVLN+boj9A36sg36NLV6ZXr2gwV69oyDa809Bx+eIriOJViorwJiyfWgWASHtEf+SOpXaA/gmglJqtra1fKisr+38AfrD2uXfOGbA7mkwmb+7p6fnfQ82e23yUo1Id3K9IAwNmRgTBB8as90KQh2H2EYGLEZlPE5kn4Lgs4ZAFpSnGkR7IQgdkoU0UxaYSn3MnkYQorFKndljGIYL+spoMBll5NAQHuYbUBxQIgT2ObaC/16iIftELAIACQVxEJnkHcyw2NjZHljW90aLKLGAF6Nr2rFZT3P+5oJc+OX4EfXz8cPbsCSPpfw8rp3nGcHNzu2NrNsd15HWMdzqKKtD0BJ/nqJx4PRZ44bl1xpVf+m2sJRwpOvMDqxmIabK9VkBR2r99+ukjXwBgJBLqQVVLqUmV/c8XbulrXNuQglX1FQXQCyDSuLah53++sDimJq1jbW9dqTmkklYAyGq9u4TfBsAAMMPMsazWSwDAiMbGgkBgInFRmfi1YWLp2IXfGDt6wWVu2Byz5HUVL6++I7Fh69O9aTXSDYKYVilqhk/IxLUdDjDGlfon9zikYD6ptnCM0T0SBwSioDBKdS6aahABwO2s7oXVN73crYQcZ02/PhD01X6qfugH8aSO/irFbEQ3WMIwywAgrOWF84I++ui44Za9rR9Jb6+rpArP4Z6OHml1Mg2BkHyNJDXD8tOsN2oDn1/BpVed/hYAMMZEpyX4xjiOQ7GlZGEfMwyTcyqiAQCmZftMRhmB5XMXA2njztuf3plIqLut8U8kVPPO259OLFr4wK6Ae9XTK/I3zbmhuXFtQwssXz5WOA5rXNug/s8XFvcVbaDNoUMpNbu6un7d1NRU2dHRMSMajV7Y0dExo6mpqbKrq+vXdnA+uBytjlxRIdeaxRSEDAiRwBhvZdBJlvCskxlEIE6xlwjgaSrvZ0RIEYeQIoyKLE01zit3Imt4YVLezBsZWEa7B1ZJjQHL2TG3JJKDul7qAjJCUSA48FGVVBH9bU2Uwjgk9CtaIgSHfBJKxffQe0TWcNnY2Aw+y8JtuYvrhvkKf0Z2avlSAJgX8OK/Al7i4qxujBIh+Kzfg3FOmX1nexu5q71HAIDPB33/RKFv+pZEyq6wOYK88Nw644Xn1sVmzq7nv/ndeXmnIrVt29o9LJPWwHGEczolynFkoFPP3n17JwOAsnJv82fnnfAKAPK586ZJmzeED0p/JNLey35x5c9ToerSdGFNOiLtESPS3runU0r97lHvdvf956rexEZS4j2BEcIP6PRp0dr7BigzrAA9E68mBuumEsmYTi5HnZwGwD/6C5fWlE89tX3tLxYndTUDm2OPvK5iQ+PTuQ2NT+eGzz0/M/7UhXmpR3fp7dkzAEASPTTknxRr7nypoiP2Nq0uOX23YJsQASbNoyOyWiYcnw/66jcAcMEKboOSqEQmj1vgen3tPZ+aeNeDTTtz14weXvTvvALpz/6f6nHhmvISFDv8cASY4XFhnNPBvrW9FRHdILN97o1+gS+Hldnta1Ozti8GYMKkOv7kU0dsG1tf8dKHDV3zP9jUzgEwx9ZXmM88uV6M9KZYSciDbFbnAEDXTQOA5HCI2cYtnclUKnfSsJrgs7AqR00A5r+eX6f96vancjNn1wsTJ9fymza0mJs3ho1E/KPJcDWpHqwNtPmEUEr1dDr9Tjptdxs8nByVGfRwb7poEGUABJTqEIUkBCENQjIgXIa4nL1cRaCZOZRuBPxNXFXpeqa4tjGn0saVeT8Uq33reL8zxrnEPk4RewF8UPjphlXqXjQSeLCpeVBLY8YhIGHvkyMDtxXbnHzkebXELe65zcbGxkJRnPWloeBNpaHgTYriHDtwn8/nvag0FLzuSI1tX2xJJFMoiHoZjLHX4qnZQYHHeUGvoVFGmnIaSZnmrnRnqShgptdFGcDJHEmd7nGthpUxMAsZIpsjzOoVDebmDS25eedOeDwaSePxh9/RKWVEVfOcphkwDMoMw2StLTHy0JJVAgBccsWpT8oO0QfA7/UrH7t0NtLeyxrXNuiNaxv0fTimoiR4wgCQyOwkH7Y9yeXyseI+Zpgq29n1Iuvu+4+AQnZdKCnrZgSGqXAqdXLFigACwOmpHekaveCyjyj62hw8R4O9Co6fKFTPmuMGAZcvFzNsZOlWAFBzPQj4xkcdUjDf1vO63Nr7umlSbdd5l1JbEO5+GXkjTaqqZr8gOn0qrHXBCizfxllTOc2x9/96eGhTs+ymdRsTsBIffX6BzyscF+MJ0S8s8Zk7cxrXqxswmVXVYjIGnTGc7FYIAJzkcjbDWpc/CkD1mt6obW8tGADcfudF/y8QUHZs29rt/NPv3zRPmzlKZ5Tif3+xnGQyGm+alJNlIdvbneIIIdSpiK2/+fmL1QDYZ+efsAzWfdAEwDZvCDPA0vf4w33LtdUrGvYanA/kIGygjc1RwVGZQX9o5Q71RxdOklDIQLN8Pk8csgMc0QGiwzTzcLgJEQUmKFySQjAZhByRxQzN6YIZy2agizk+4OwTKz3N2fc7qN6e3GumYlm4LbMs3DaoBjgEB4e9i17sWSq1r9Ipu6TKxmYvlJeFLgsG/H9DwTaESnBHJBq7ojcSWwYAfq9nkaI4z+qNxO7d74E+ZZZ3dGs/nnRCHwDXzly+LEtpcLrHTQ0GrkIS8r26Ke3I5QkHAASgDBgmW/HbZ3yeVbLVAz33diSW9YoCSeq67ZgMAf71/DrtmRd/+IdVb25b8NjSd0bG49nObyw6K+hyy7JpUrz3bhj3/OYVxGIZVA0LbF9w8dR/Fl4qfnPRPO8Lz61L70sk7lAIemt5l6MiLAmecN5I1fWlt5F4uol3yCWMIwJTtV6esV3/jgCEUqaD8STP5ZlAHRhYbsoB4Ovmnq80PLLEXt95CAx1ezV6wWXy8LnnewTFxcMKSkUAhmPE6A/AcWq465XxzR0vFttaoT2yWuqMvgOnXEpNqlmTP4TQiqozXqysnv06NMYAlkV/ckgCwFWE6klXpOFTs1nLwm35pK7Hbqwf4z7B56U5Sj0jHLJYKYmZtEmdHXmd68jr4AlhJmMEAEY6LHvboxslsCYYEgD886rKS+5ucKaP9+4ZmzeGKQCzpi4YW/Lw1y7++uV/fuHBP6wo/+KXT8EZZ47F66828Jqms29df3ZUEHg5lcy5o5F0w+/ufrX8g03tY+Z/fuLvTz51RMOA431iYWYbm6OZAwbodTXVv4AlWrZfwq3tNwzKiA6Cn/5jgzY85EpfNXuUpTqR13VoeQ2yZN0oBD4F3ciD41zgeTfvceTgkHtR6soByIAxjRmUENEq7bvjiZUJf7hDuLhumBOFG0dS1/UHm5rVuxu2DboicgvSZi3cexP/2dMRo9i9NzIAQGXGJ2r5ZmNzLMJznBAM+H8LQEskUg8AyPt8nqtDJcGluZy2I5XOrDvSY9wXbWqW3bZxS/LHk07gdMZSAFAtiXHGWCCim3yVJOQoA5+ljDfBWMakcHKEB4A6WQrDchjVGaEgfWv+Z5S3IzHtJxu2pLYkkna5+xGkti7EBYIu8/4Hr1z4ve88eu9Lz2888eV/bdKrqv1mIpHlU0nr9lJe6Wu+63dfXuxUpGKpLwNg3L/kWs9J426MD9Z4JNFFQJCrqj3rkeYdzy/mBUfa6xqe60s0hgAQWSkx3CMn6LqQF/PRLiHX1my2vPSnO8pPu+D/fCed/uoehyvqtPDB8RO5WOMm+1z7GAx1ezVx4Q3O6jPmDNS8obAC6rK+la98FZQqDBTOutG6p3KcbvT1kdT2DaJuZIRMrrMQgBM68sQv/z3oGbsBFCbA0rDOGw39VYIslgh/6sHt8o5uY3lHd7zlwnNTrOD3RXRDKRWFbKkoEJVS0WQMWcqIRinvtJYZEZ0xAdZEBQFgeEXRfc3oEZ7bNm45rpXcE3EVjy5dqV565SxPdU0gded9l/z0B9c/cduTj70bkh0i5XnOfGtlk/jWyqaSAS8bD4DN//zE399867k/H7CdLf7eI/a6GZvjmgMG6IrivAjAmIM41qcWoAPA1/74VqY5kun+0YWTqgEILJ1JEjAJfh/A8wYoBbK5JLR8FgFvCoXyTwAAISgG582t0fx9D6/IxRMqblq3MX1aaQmXyOtsSyJ52G4Y76E3Vwu3DCv4HliunkehbVxhXx7YLWMBFQZ9D727bbOxsQFkhzwKQGkikfppR1f3jwEg1hd/YMTwmjWVleWP6S3tpxzhIe6XB5uataRu9H3vxHEugRB1a1bLXxIKdIe1fGmrpjs4AiYQQnXKOAaQLWrOBMCPckhhWGKWxSwmmREKOp6Yfar0pRXvxLYkkrZwyxHA51dw/5JrfQCydcNLuh/+xze+/syT7y144bkN5+5o6hnBGKS64SUtZ322/pXLv3La0x6vI45+e68CoLV1IXnC5Dpu84bwYAW/VC8RjIB88qoMjT7T2/zWF+LJD50A4Bk5SQ+eMc/UkZMkavIuAEY8ykdffobrWfPct8Sa6h7JP3IDrLW3JoAs+ieV7aquj8lQtlfB8RO56jPmeGAF5CKsQNoPoDK5fs3ZyXWrr3DWjWr3z5qrACRAeIHInMjEE8fysdf/Cb0vAgAoG3bam/7g+O3IU4AxjuiMg6Xvk4Pl69BYokU7Eu3WBhIShZa4abpdPOfo0Q0PAZjIEdNk4EzGOJ4QSi3xxdIRstQNy94CQCWA1Il+rxfWmvbjmlu+/7A6c3a9q7Yu5Jg4edj2h//xjRse+tOqC1a8vvXUnq5klSQJ6XH1FTtkWTAZg1pTF/xgzrwTHh2YOQdAFy18IHEwHSxsbI5lDhigN+0In1RXU/26KAqnJBKpH+iGHvk0BnYw/PQfGzIXTKvpmVwXkMEYWCoDZDUODlkkHMcxwzDfXB9ObpT82vXXztlNER0A3t/cqn3t+r8m4on+m0NBUfmw8gprzZ1ESh21cCdh3fSKYnEMluF3wTL2exp89ijbloggd1yXUtnY7AMDAERRqCpuyGna9u6eyBXlZaGXKivLnuI5PnjkhndgloXb8jfWj4nP8LheXZ1Mn/dhTmuc7HKyPsNwZkwqGYzxboE3GFh8QyYbColCZJLL+T4GCEkW8Yoi96fTpnlPf+n1viPwVo57zj1vmuTzKS4AbgCi7BBxyRXT/3XJFdOXwwpsnbB0VBKwbH2x36+K/skW+HwKj4/2Nj8kWrrWaZ7TpsjUySUqXef+QwqUJtvfe/Yq3u2BcvIpYp5mRAAwU0ma6wgTM5XgpLJK5Nqahciqf15aNeq7bSDEAavXcFEhiKZadtiTQB+fIWuvyqfNcAMoQWGdOEzclywAACAASURBVCw/xQ2gom/Vy5cJHl+y/EvXJDhJTqY2rXPkWrY7aS4L3uWBb8ZZ6HvjRWZmMyTatf5UX3B0k885/ANOZxqhMNBfeWEAyP1n85HLlJ5WWsID4M/0uV9/MhL/2gdq7t3P+DylCdP05SmVeY4wF8flnTyXuqejJ+DguNypHtcW9J/7PADPMMVpL/EokExkAIQoAD1Y4uq78QfzHr7xB/OWwjqXip0jVACdAIyWcKRj9QorPt+0sUX/430v5VrCEdvHtTnuOWCArut6JtYXv6u8LPRYPJl8UlWzOz6NgR0sX/vj6uRTN57lqyt1WwuEDIMibWgMwIZwn3bRXa8n42qe3fPAq/mrv3y6POXEGqG5NUrf39yaf+jxt46IuIcKA3ey9YnvkknmOPgZrPJUS/AOSG5FvLUGbqZA4Ae8Rn+QNaRs9XYbm72j5bQdJqUfKorz66NHDp/SG439JpFIPh7ri7+sKM47PW7XD470GA+GYYqTX1gh/H59Wj3ztpbO4TdUlfWc5XPTEsHK2mzLauzO9m533DClH9dW/kMgpKifIQK7HODisSS7L/qRYcLkOgX9gc3ASikT1veUgDUha8DKRlP0V1ENZNAmjbe3rMyN813LUwfxw88TFpMMAAjN/nwnLzvLmaGT5H9WmZmmLQL26AiXaw+f0Ldy+YLA7Pn/gtXGKAbAaF/1WsZWcf/4DFV75a0dydXNPT8EK1GgwGr36gYQ0rraR5qpZGlg1txWms0oXY8+UJ5ra959CSQh4GQnA0ComZd2fPD4ZSeOX/jfCvX3FJ4hwroG8qvfW9LZFTlymdIfT6p3A0heWhp4+I1E+pzfdUYm5ynb/oUSX59AiBcA35nXjTvCnaGdubxnYUXJP/0C31+NaV2v/DDFKV1cN0xYFm47brO+554/Tbh/ybU+n08phTVxYcKybxSWaj+H/omZYqUoautCjkULH4gOptaGjc2xwEGJxOU0bQOADdij3HoosCHcR6ct/lffVbNHOS44uUaGddGbz/6nVbv3pYZdWaXm1ii97dfPH1TLmk8DFQb7FVufqoU7cxIpFQGQFpY23kPvLgM/HgEOBaPWiD57fZ+NzX4wKWXdPZHLystCj4micLIkCuOL+zo7u2/hqytlRXF+Fx9tbzjUYNWSuO1ndZU/+llr109ub+2qe6BLUIc7JDOiG9zOXN7FE2JcX1X67BleVyOAKliTe1lYmQkN/WXImBEKinaA/ulz6ZWzPLCCm72dbwKsSVkD1ne11wxcIqGag+m4BiZOErRqEXyaciDg82baAQBEknhq6lzs3/9k+e52wTPpFOabPouJpZUwUwmWeOdNknh3JRd/69/zXeMnvyyVVbpgZVZJKryjd7DGdzwxVO3V6AWXumF9ty5Y/ocIwAdAMRKxEgDgFZfW8dD9w001LfhP+wz1nHQaE30B5KM9LPnuKpJ8/20eAIKnzPkwuvaV8R2tr88fXb3gHljnfQ5A3z/fuHV7LNFyxALaYYoTJ/i8MoC8j+fDvxxe9d3FzR2//UNXpH5pT6yvXnFwKdMk27KahwG4vDSw+pJQ4D8oBOWw7Gyx6iU1t7LcuSzc9qm1ixtKzJxdzy994oYArPOlWCFhwvqcXLDOoRQ+Ko6cB0Du+M3l7jOn35r4FIdsYzPkOagAXVWzjQ1bm6Yc7sEcKnE1j3tfasjd+1LDoAu6HW5akKYtLL1XtcpCUG4H5jY2B0kikVyXTqXHyg55jK4bu274JqVmuLX9Rocs/1JRnFOP5BgPxPKO7vy8qvLsZJdz21/G1P306Wj8M++m1Ylbs1owwPPq+UHfhi+U+FfUyGI7+vUqGKwshQCrXFqA1S7StBXdP31mzq7nfT7Fif0HV3kAHlitPffKH+9bPqiNZsdfvtDNeJI3fHya6MwJWUwDQD4RCRitfch3t5PgWedS/8w5u14j+ALEe/IZIKKI+Fv/lnpfWLaw+ivX/QjW5EJy9ILLfO2rXovqasY+zz4mQ9FelU2d4YC19hyw7Euxwo/ybm8GAGJvvFhN9bxQfsEVUMZNAOF5AoBJpRUkdO4Xwbs9Rt/KlwUqsICjsq4t1rV1FoAfwgpqswByRzI4B4Aal8Kjf9ljZrgsvffnMbWXPRtLXPpWMnP21qxWpnAkf07Au+m8oHf7eKcjDqvs3wmgHVZ2mMGysxmfJB6VXZEGgx/cusCNfmV+A5bdU2HZt6KGgQxrcgboryLKAsCESXVybV0ILeEhs4LWxuaIc1AGRVGcIzxu16WJROqJnKY1AYDP573AIUtjY32Jv+q6bs+g29jYDAlkhzzc43Z9MZFIPaEXgp+B9irWF3/xSI9xfzzZ0qbOqyovA6B5eK7vqrLgM1eVBV+A5eA4YDk7XbAcIoJ+pwewbLoDluPjARBf0xuzl8V8+hQVqvcUAh1IUc16YNupXTy6dGXqV7c/NWiTzt7akcQZKpNgOc46E4nuGDNuPV4HS295XzQzKUillcw/8+w9RsmYqaY5ZdzEfK51p5Rr3TlZ62z1yJU1KgAIiouvPmOO3Pzyc0fdBPlQYAjaKwes89aAdV4W+5Qbcnl1JxGlHNVyTveJJzHnyLGEFNTNBx5AGTuBS763Bsn314Q8E05py3WGvZlcV87lqCjq6hzxyZzWjGoWxlEcu+bguM5LQoHnLwkFXi9s88DKAPvRXw3D0B+E5gD0oH9t9XFHbV2IzJxdLw/YpMK6V+VhTWL4C9sl9N+rMrCW+BQTUKSmrpRvCUfsSi8bmwIfcQr2xOfznl1XU70pGPD/nOO5YcXtkihMCQb8d44eWdfg83mnHd5h2tjY2ByYY8FeneDzCugvU++DFcBlCo+TANoKv2OFfXsuPSo6S+LbkVjebrV2RDBhOf67lhrsBQogumjhA7E/3L88uXpFQ271igbt0aUrM+fPvSOyaOEDgyo8JSiu4oSODoooMZkuhSp6nCPGvpPvboeZTsI1bgLbU5A9H+0hzDSJ4PKYzpFWFXauZcc4DKjuCtZP3NckhM1+GOL2KoXd27xSIgia96QZGwHAUTOisG/3uNRI9IFqOU4ZfYJhqhkxF95OAUDgHbsy5q2d6454j+s2NYstieSetpPBCiSLAXdRbT4NKxCPwrLDOQAtsDLpDgD4IP6RYx0X1NSVFu1KkeKyHaCwnKHw24B1X4uiIBA34DWsNdxrB+c2NgPYbwZdFEW5qqJsKQCXrhvP6Lqxvbgvo2Zf8XmNWaIozCkvC/01nUpPMik9LmcQbWxsjjzHir26uG6YE1bQ3QErSxOE5QCZ6G8XqRQedxeeO9BBKpZV62t6o0NGd+N4orBuXIc1CZ6A5cQXNVKKYnA5AOrqFQ3mo0tXZlEo9xwsQtWl3ILvXugbP73eUVIVQnhDlmxrzpXn200ZJmTTxxPTxQlVp1/0eLj7/nGGmgow0yA0l2WE50HzGoxkgtC8RgSP1wAhjJOsuR9mGAS7B29D8loaygxVe9Wz/u1s2dQZXljnaQKWjamFZWOoGCrrAABmUikf6YbgC4B3OEF1HWY6CTOr8pzDacoV1bkU4NZ6O2pk0bdVFv1uWMFvumHHy0NC9fzuhm3pJTOmSei3nxys61AobOMA9MISyXMX9qVgfS5RWOe9nNR19mDTzuPS1iYS6t4mgLOwAnAnrPsUg2XzOrGXjiObN4a1Ynl7qLqULPjuhcr46fVSqDrER9oj5qqnV+Qa3mnI159aLwLAulfX6S0NYTugtzmm2W+ArijOswBUqWr27+HW9ssH7lPV7Fs7m1vOGTG89lVRFM6WHfIkVc1uOJyDtbGxsdkXx4q9GqY4i3ZZh5Utz8EqtSSwHFyC/uxsURnZCauUmsByjFIAcm1q1s6eHyFeeH5d9NzzplXBck73FoDrLeFIcrDXXSpeBQsWXeiZe/X8agACGNC3g3kE1SGVGHmf6iU03QudRJmXjxOHoLh6y+rPfqpj3bPXZBo3Eeeo+l2TPYTjmBgI6bzTqWld7UqudScDQMRg6AMMCMpjDZuOW/XqQ2Wo2qvm5c/1lU2dEUJ/OXcM/ROFvBQq7wCAfE9HRqqoVvRYhN81U0MIBK9PF/0lWuo/qwUAYNSUystOeQFAAFZwlqupmCp1RRqOeEC7vKPbuGndxr4fT6r3eUVxYPl6UcejWMoew+5tb4u6HwBAb9vYkGpTs8flJNXmDWHWEo7ka+tC0h679MJPCv3f/d4qJ9ji7z2SBoDa+jr+h3+7JaB4lV3aHaHqkPiFRReGvrBo12Qn+8KiCxFpj2j3fOfuZEtD2L7H2RyT7LfEXRKFkQCgZrOv722/SSlLJJMvAYBLcY4Y/OHZ2NjYHBzHkL0a6OjlYa03b4dVKtgHS5QoDCuzVRSSzBS298HK7OQAsDW9UXv9+RHiV7c/Xfw+9uZA5gEkbvn+w4MqAgcAC3/xjf7gHECqE4qWgJRPw6tIzMHnWbWD0nohbfp4lXqEhBkqqT15oxKq2673RZDv7aKCy6NJZZVZuXJYGgSG1tWuUE1j2Z1bDSKKEUfdqLeK/89QM2b7qteOeMny0cZQtVexxk1G+6rX2tFfgpyDZVuiAFRHZd0Hgssby25vpJI32EQ4jgIALzs1R8WwtOgv0fKd7WZq8zoJANye2vcrAqe8if5OE676UfP8FaH6IdFNY1m4TZ//2qrobRu3JJZ3dMdhBeJxWNneor1NFR6nYF27xd8ZALE2NXtc29lbvv/w3hTai7BEQo0lEmpyzx2JhGouWvhA3+oVDabiVfDDv93iHxicwzpnfLCqj0RYE9UAgFB1SP7h324JhKpLCWyOSaqqqn4ly3L5kR7HkeJAInF9ACAKYv2+nqA4nWMBIK8bx2V7CRsbmyHDMWGvtiSS+gk+70AnpbgGMg0rk+NH/xrnvU2yZgBgeUe3erxmdYYCmzeE6ZVf+m33/Uuuze+h6J4HkF+08IHEC8+tG9TM8xkLZgtTPzvNj+K9nUHMJ1mIE+AgHPxmhvmFPPPqIDzHwEiO8tAgE4OZw04877ltK//w7b7X/yXT086WHTX9MaGZSmTib79OqZbzBGbNvYNX3MWAnDY8siRhK7gfEkPWXjU+siQpKi5aNnVGANa5VNTESPMacZdOn//3ztef+E73Px4KlM27+ENTNEeaWlY2O1vlfHcH+t58EczQWUnJxHdGVJ57P2fyeuE4LljBrXdUzRnJrkhD5tN8X/uiTc2yB5uacw82NedemnNG2wk+b0VhV3HcRQxYJdut6F/iQT+IJ4/rCpIXnltnLFr4QPz+Jdf6sMc9KZFQjUULH0isXtFgnHveNKm2rlS0tmeMR5euzCfiKgOAMxbMlhWvsmdM4kR/S0oZu6vBq4pXweWLr3Df8527h+z9/FhDluWKQCCwyO12z+M4roxS2pNOp5f39fXdr2la12D9H6/Xe7Lb7b4pHo//cbCOebSx3wA9lcq8GSoJGj6f5zu6oa+LxeKPmZRas6UcxwWD/ssUxXkVgLSqZtd+KiO2sbGx2QvHir26u2FbZsmMaTL2VOuyKJYMFjPnAx0iVthnbEkkczet2zjo2Vmbj8cLz60zTlpxY+TSK2fJM2fXiz6fgtUrGo1Hl67ItYQjgx7UTvvsNCf6W2Q5DQ0eABI1IGpJBKgOH8+jVATLmRSqqUMGQATNhM8ZTA2f8qUHwhufXNj35osO3u0B7/IyUDOjR3skRqnoPXnmksDseU8BoLHGTVrT049mYo2b7LWgh8BQtle6msH6e36eDo6fmC2fNkP21I7gs5EesXzqjFJRk2v9w6d+YEyKvti76d/zW5f+1uesHhknosjnetoCZjpJOE4wRo24cFnIdeJ/iMk8sAIqCVaAlQcAt1LqQGEycSjx9TXrok/MniEOU5wlsMbtQL8tNmBVNO3KmC8Lt2WS+nGdQAcAPLp0ZX71iobIpVfOdsycPV5IJFSsXtGgP7p0ZS4RV3c9Bx8VNQUA1E+v37NEHrA+ew/6bRrQL4IqA0hO/ew0E9Z9z+Yw4/V651dUVDwGwMc5nM1yeZWpdXeEvBx3i9frXdTV1fXlZKHq5+OiKIpXVdWk1+s9paSk5F5RFE/G/tuUHvPsN0DPaVpnKp250+N2LQ6VBB8JlQTvzmlaIwA4ZHk8gDIAiPXFF+u6bl8gh4gCAbXwCOOIn9/K4mYj+o7r2Vgbm0PhWLFXyzu6jbsbtiVvrB/jxV6C9GXhtuiTLe3Zr40a7jzR783ByjKQYYoz93Ykpi/v6MouC7drttM4NEjEVfaH+5bn/nDf8sPehqxQHlrUKlAKj0F1CGYODlOHl1GIggzGp6FTg+3KAnI5wF86rlM497of9Wx9c266ZcvkfHd7iAiC6agd9bbv1DMfVkbXr8lGevQ3b7rGbq36CTka7FWscZMZa9y0S9DN+7N7iCLXhPi0qZbXn/2Mu2zk5mjDqvmZrp2jTD2nCJIrFSqZsqOqfOZGh1zSRHLUgFWa7IdVIq/AKnPXMESd7zY1y+a/tqrz4rrq5LyqisCMUDAPK/OfhbUGepd/tiWRzN62sWFICN4NBVrCEfar2586JG0BxavsbUK6KCw4EDLgtxeAUVtfR1oawnYVz2HE7XZPq6ioeIZzOJMl51zwjGfSKcX2eUhtfPf96CvPzqyoqHiGUnp6Op1efzDHLCkpudHtdgcBjJFl+ZIPP/yQZLPZnd3d3T+UZbmytLT074ftDR0FHLAPemdn9616KBgLBvy3AChzyHLZgN29sb74/3T3RI7bEoRPygVkhOMc1LgVCNbNigAqDONRti21Gp3HZdsOG5tD5VixV3c3bMu9HYnpXxs1XDmtNCh5RRFvR2LGsnCbuizcpgPAmt5osezdZggzYXId5/MpPAC2eWPYKGaTDhMU/erJ4CWYhIBREzwYeGpaLaEYBSE8nLAymIzxhDCJiEzm3JIn2F3xmS8+wngsAZA3/EIfrKxXGgCcoTKfqLg6dHXIJT+POo4me+WtHcl5akfypFNXwcATA9TtG77Nc0pdBzGZi5gAlzY9XI5xnM6cIJZ/yUQiUYHITAAHEI0wiERnmVgyPGRnEJO6jgebmjMPNjVnAOC00hJugC3mtiSS+oNNzbll4TZbe2EAM2fX87AmXujqFQ0fK9HU0hA2xk/fbbUHQaGF3R6YezzHGWnvtYPzw0xpaemvAehVV3x7o1Re5R+4zzPpFL9cXr2x7U93nVpaWvqbdDp99kEe1ivL8nfS6fRDvb29JwOArusRXdffADBycN/B0ccBA3STUtbdE7kr1pe4X1Gcp0qiMAIAzeW0JlXN/sekdMga2SNFCA5yARnhOgmlsgKBU2GY76E39yzbmY0gt8uQXErGOM9BjXfP1ysQhGtIfUBhQvwVtNo3ABubg+RYsldreqPmmt7okM302+yfCZPr+KVP3OCprQvJAzbTP9y/PHPL9x4elChd8SqYOudkqf7UeilUHSKwnONiezcQDkzykXw2xhg4EMbAUx0OMw/GGETGgVAHJ1IHZ4CDQSVCwOAETygIBEKZCYokuN1aqnGe2pG8Xdr+yTma7FVw/ASR05ggxA2OT1OAAw9CJGIyGRQUDBoYPIwHhYEcCECdnMI48CBgTOZ8TOJ0RuAyXVwuNPeS3vrXFa3p6b+rQ32yZ01vlBYmRG32wrnnTxN//usrPLV1IbG4LZFQzT/etzz1q9ufOigfdvlDy3Nzr57vRn+GvGjLivaTK2yTYMUuOgBDTapUTdpFDIcTWZYrRVE8yzV2wrtSedVeK1+k8ireNXbClsyHm8+SZblS07TOgzk2pbS9o6PjxsEd8bHBAQN0nuOEUCj4FYcsnw6AptKZ1xKJ5BMmpfbNeS/Uws3fTKYGdmXEASgQuJmoFE8ipc472fq+FqRpLdzcOajxDHhp0bna1WLpUjLG+x7rjQwM6m1sbPaNba9shgITJtfxb75zexBWeS9gZbYZAO6bi+Z5Jk6q5c+fe8cnmnypra/jrv/djf5Qdb9TXMANK4suAYCrDNlcDFnGwBGAURO6qUEDIU7qIE4mER4ceCpzJpWIQB2cREXiIyZLAsiC2+UMF9Gxd30Em4/J0WSvJE4RnDs0P6cxxniSJAbzAMwgJgSYzAUOEggoOJiMQ5rxRCIm4xlPeNPFuSGQPONJlgqE0ypFnYkkUDf3fMNTO4Jf+4vF9kTkUcqlV84S719ybQB72ASfTxFuvnVBSU1dKHnL9x9OH6hyKNLeS5+5/6nkFxZd6CtsKqr+F5fryLCqeAZm1XOw+tTbHEZ4nh8DgDjqRqWwn6UpjrpRqcyHm0nh+QcVoGua9sEgDfOYY79t1gCgtrb6oWDAv0RRnF9VFOc15WWhv9fWVj/+aQzuaEOBgJvJVP/A4HwAnAJBWkQmeQHgJFLqgOW8eQGUAxhV+KmEJfwUBBA4CaV7K/GxsbHZC7a9shkKLH3ihlIAIVh2PAigBJbYEQcAM2fXK5deOWvPwPqgUbwKsVoMfSQ4B6w2URSW82oQDsxVQVKighSj6DOz6GAMuikQykRiUJHoVOYI48CDQ5w6uAwAwnjipQ7OCcs5HoiajXQPuQDyaORoslcnCKc6QRlHBaITyhJM5tLgkEdxTTYFz3iigpA044kKDnEmEMN08x7GE8I4YjIeeb2E79KDQg7WunR3cPxEpfqMOXsTCLMZ4vj8Cn7+myt82D04J7AmCUsABC+9clbdjq4/lt+/5FrF59/TlOzO0/c9lXvkjof71KRqwDqvij8cCgKoA57OAOhqSj1kO2pz0BgAQLXcfnvOD9j/cZY32FXC+2C/GXRFcVY7ZPlSAFoikboXgO7zeb7jkOWLPG7XKal05t1PZ5hHBzNRKSsQ9tYmYleLnRAc5jWkntbAUxRO4bG7EnOx32MSgHQBGVHyHuttt7PoNjb751ixV8NrSsjwmpAAAM2tEaO5NWpf+0cRS5fd4K2tC4XQL9YmwHImi7Y9DoCee/4056NLVx5SCfO8q+c7Ci2JeFj3ERf6FbKTAHpgTf5mAPCpNubleKw2cjiRCHDxEhHyJhxUJKVM5tyMEJ3KJMF40st4aAAEEMCUOQbL2Y7DWvuZiTVuSmYjPfY5+Qk5muxVRaie9/IlfDZHdergRMaTOEwmUpkzqIQ8l6eEMPBgAPIsR0zoVCCMOkiQyiQHHlkqkrTh5SP5MrEbBAzW9REAkKqeNcfZvuq1IaG54/cpmHJijQCA2PZ3/5x73jSpoK9RhMAKzBVYPi2FZZOcl145CxMn1wrnz70jub9s+ssPvZR/+aGXouOn13MLvnuhMX56fTE4d8CypcVuACYAV6g6FPrZMz/X1ZSaV5Oqvvyhl3KNaz/e+neb/aPr+gcA9Fy4qRKz5kb29bxcuKnSerpuZ8UHgQOVuI8CQFLpzJ87urpvLmzL+3yenzgc8tihdAMZCown/j1ngb3oXz9ThJ+JymI2hcIK3vesZOBgOUUJBQJ/ARnhepA12OufbGz2z1Flr4aXuvm7rjw5cMGpI3zgCJ9IaZmw7DamnH0SwHHFfq/8+5tbMw89/lbfQ4+vycUT9lq7ocylV84Szz1vmhdWIO7BR0vBFVh2Pz5zdr30g1svdAAgLeFe44Xn1+kHKyA3fnq9DMthrUZ/GT1gnTNuWA5tFlbPZjPTA54TwXlr8EGqHeOzjHg1gTcMN5eHQHJMIBoAkzq4GJjlUFOZGODAYDnDBoCEoWZyjY/8yb4XDQ5Hjb0K+GoFAJC79FSuVvZTieSIwSKMRyWTOJE6SY7LsSCn0RQxWDsxaJ55hXLq5LKmwuWYzGUMhUtQF5dhIhkYPPEAJG/tyE81OD/r9HH85AnDhA2b28z3P2g14gkVfp+Cu3/2JdfVl5zuQv91y95868Pcjf/zePr9za37zR5+XH500WTl6lkj/XWlbhmAuaGlL/NcWNXOmjeVQ2Fy4KHH1uTeeGvrkK1Wqa0rLcYQxcnIclgJJwPWpCTQX4GTnTCpzvnNRfPzv7r9qQN2tGhc20Ab1zao46fXF3vRZ2HZVMAKzosTlFxtfZ2AQvu7qZ+d5lr19MrUkv/+o32zHCR0XU9ks9knEd5+cWrju695Jp2yZ1yD1MZ3tWx4+/hsNrtM1/XEkRjnscaBAnQOADRN6y5u0A29r/DwI1/Q8Y7S7ycVS3y8sJyxPbMkMizDImLfn6NQ+NFnotLxIOwA3cbmABw19urq2aOUP99w9hgoTi8I4UAI8Vc4An5BAOuN9pHSEhUclwNAp0yocUyZcIn3x987L3L2hXfF39/cOmQdtuOdS6+c5YRltz8i/lmAh9U+i/l8Cm6+dcGu7FMioRq3fO/h5MFk1cdbPYOrsHtwXmRX+6GCgysQHgwGwEuI635+fUrlRzKelFInx1GJZEBgMJGkTQ+fZzzpoyKJgUCGde/KAEjFGjclGx/5UzrZsmNQA5XjmKPGXqXVCAUATmOmI6zFtSrJmy8VHCBIcgZTYRIwgTVQRrh8hUj5NGWmi/CmwqdNNydRB5ehTi5LJbLnuW0C4HU186lkqa++5HTx7p99yev3KQN9X/O23zyfvmD+FHnKhJo9lxSSM08f6/z3UzeJZ194V+z9za2feJx+RcJrt86tmFwXKEFxIoDn+SmTR9ROmalwZFhlH1yKeibG6ldfcrr20ONvpb563V+HarDJYCWZXLB83gpY74nBmhwsTrw4Cs/JfPnKWc6DCdAHEC+8Noj+lmvF76FYHi3AmhiIAaBnLJjlibT3mk/fd3ACdTYHpqen5+aampoze59/bFa+u2O9b/qslOALOo1ELJtYu9KTWLtiOqW0p6en5+YDH80im82+qev6xr3t03W9LxqN/kTXd9nE444DisTZHDyN6DPGwV80RAH0f74UlpNTNFYEluNTVKwcKA5non/9hlH4e0j2C7Wxsfn4XHByjfjn//fZ8XA6rGwAx/FQnCWQRA/heS8I5WUh2AAAIABJREFUmQA12wm3KwqgD1bmwPT7FPLvp27iRp6yOBJPqHbZ5RBkwuQ6AVZwVSzh3RteWLa9G5ZzKwPgfD6F3r/kWncioXa88Ny6/QbpjWsblEKQvj88615dl//FlT+PX/D5xUbIPd6pUY5kR4tBM8R1USdHGA+TipwXBCYIsqaTy6G/57MJy8nubXxkSU/zy88d9j7uNkOTvkS4mBElXJ6ZhMEUsjTOCOGKp7rp4V16QMhRB8kTExyAPOMAxhMnyC717eI5a8IKrnIAEGvcdNjV6q++5HTpL/d+xY+PXpf8j793XhksP0tDv1o4KYxT9/sU4e6fXeL+zILffGIxux9dNKlkcl0gBEIIREEgPM/D6ymBy1kCnpdYPJknstQFQcgCSF19yelic2vUvO3Xzx/xYNPnV/DNRfMd554/1TFhUp0Ay1aUwfJnFezeo7xYiaAN2J+trQsddNyRSaomrBMsAyvI1wp/D/xfgPU9FduyqQAw96r5LjtAHzw0TWvr6Og4s6qqalli7YozEmtXGLCWUpUBECil73d0dFysaVrbwR5TVdU39rWvEKDf9slHfvRyUBdKqCT4g1BJ8PrCn3Jh272hkuBvis9p2NoUPAzjO6pQIHCwjBKH3T9bDlZpThr9s4pOWE6Qr/B4ICYsx7yYNbczFjY2B8lQt1f/+9XTygrBOYEoSHA6Q0SWyiAIpeA4LwACw3TBpA7wXAWACKx1xYrfp4h/uecr7KvX/zVil7sPPRJxlfh8igjLtu9NEUmClfWWYDmTA5c38QDE+5dcW3t++I6dmzeE92n3x0+vP5h7N6mrr+MB4J23n07NnflDyajg3aaL8zGeCAByjCAN614lU5lwsEpUi4F4UQDMUT1rjt9TNzKdCu/Q21e9lv+0Mp7HA0PdXgFAWo2w7S0rM6NqZ7mZSDjTzclgYIQxk/GENxXOxwTCTIXTQMAYv6tyMADrfB/YPguFx0W/KGuoGUNUXDhc7dYK5ete7H3SjMASdHTDOudl9IssZguP02eePpYNrylJNbdGD/r/TplQw/m9Cg+AvvHWVnN4qZtcN7++FA5ZJi5FASEcZMlLvO6RIJwIRnVGzSwyqgc+rwkra8xdv3COftuvn+8+wL87rPj8Cnnu5cW+CZPqitUdPPq/1xD6s9oDcWL3THqxKuegWP/quvzli69gsM4hgn5feM/Kofye2xWvIo6fXs81rm2w/edBQlXVph07dkzz+Xznut3u/4Ilat2ZTqdfTCQSL1BK7c96EDnYmayi0NlAXIUfGwCFtmkuWGv/9lXeqMDKmnCFx8W1gnLh76LwRabwmALAe+i1ZwGPEcqnzhCqZ81RyqbOEAGQVMsOvXn5c2r7qteGTL/bY4Ahaa/OrC/nr549yjW80l8HkwYgiX4IgpcQ+AHU9UQzoRXvd1R/GI65GIAxo8ra5pwxamNJqdf3/9m78zg5irp/4J8+anqmZ2bn2Nkrmz2SLEkWCAkJAgIJKHLIw8PlwSM8gFf0pyIgz+OFoogcoo8KyPOoRFEOAVGJoAKCIFkIxBBCQgKbhM0xu9l7du7pnp7urvr90TPZzbI5SPZOvV+vvLLb3dNdvVPznarqOkDkXjiVdPHCDy/y/AafxMWf/L99TtbCTYzVLa1W/RVLAadgL2LvJYEAp2Juw4ntIgYrKqy4DYGA6nnwsetCx8/7yv5qAmbx3/5mMM6pflUEgJ5Yq736jRXJ4y/9clWxcg7BZJogCm4GiFQRbCYJQQzOlpyC011UAODy18+O+OtnKzjtzHzTxZfZrb9bkeIxa9RMyng13Gubf5fzqRVSpP7ovco3tlcMU7egWmWSDgFhOI06eTjlmBSceRIs7J1XBTjlngCA3Q1nXxCoXHyy+sZdtybHYgjFhecucgVLk5nZtszSWRV5QwJjAJEjMK16wTTLwRgYkdNCoGwAPm/pSXoWTpmONdZFpF0dA/sdYlRWVnayoigfWnRsQ83p72/Kn7i4fs2HljatTaa03Ibn1tlwK17BqwZAGQGoBy5yHCirgMgkCKIpECkFy3bB+RvmAZQFA6p+xinzBl58ZeuETX724GPXeYdUzgEnlglwKuA+DObXUtm1FNOGll+FzW9GDzpuxDr72bP3P5M5+6pzI8N2DW1oKfX2GQlfDnKUUUppIpH4ayKR+OtEp2W6228F3cgb66IdnYvGKzFT2fFChQInGBTgjJlRsfcTlAKcQCYWfy61CMrF/4d2H5TgBDsVQP45tntsmpW5cbVg+XWe2tPO3Ktw46+fLS1Yfp073Lwgt2nFnXyegcMwmePVT644wX3Nuc1lAGTIYhBEmgdGVQCVO7vSM3/42KaTWjZ0RWzKhhQoWuf/YMUrp1931YmvXXHxohbBo2yDKMYAkAs/vKh6/fM32ovP/P4ROz5rMnrkwZeyn7hiaanyXRrWVJpzhMEprJYq7z4MFnKBwS7lhfqGiOe8C5aQfXV119KaoZapGTjfE8PHK1M4lYpC+5bonsm3tre/VJjb/FVN0qhHKDAZIlSqiDpVhHYmCZVDXl+A04NLHJImFK+Vl1WvtGD5dSE91jcQ37KJz4dwiCZzvBpJwdTw99W3pY+xP2o1LbzMLRpMtlXRb5WJLqqIueLM7IBTpvFhcBxyFk6esuGUaUrrXBtw1rAmAPKeSCU58Zu3hVbfeE18tFcJaKwvlwEQGEaIJdIhCIICWVKQNyqRTFcDLMRkmQHIIm8QltUCKPMPCNUVnRAEs5hWP4Y9JQ6qLlzz4Wb3VUtn+9sGzKXf+9vO78ay5nEAEO1M44E/rscDf1yPo2ZH1t15879/7YwzF3az6O6ZKJgCKPXB550FQpoZ4BLAnCXFKAuCMQanQbaneO1Sz8wJUd8QEU5d1jy8R1Bp6GXpfQ5gsIcQLW4bXna177hl5Xsqz/7utod0APGzrzp3zzKVcPKSiMGJMDFkewmLdcb4E11uytpvBd2mNKtp+sbxSsxUVg/f0L+lBSewlsY7MQwG9qEFNV9xm1Y8VsbehbjQI+ydHVuQ4IWgKa5q8cmkWDkvdW8tFX4NAHbtaWd69Vif2bbyYd5b4hBN1nh14Ql1pFg5FyBJAhSlHpQGIAjy2q2xuV/82avLcnlLBgCPIuOYeVWMEFn41xvt0PImue2Xq0/Z0Npb+cUrT3yq6ajqGESxHwAWHVsX/O5X/12bDGMTOcfqlla7PRrrr2+IVGJwrpGhlew8nM9/aRmioSQ4lYAcALrguIZ9VtBfXtmSPvuqc0NwKj86Bp9OluYwYQD0lx5/aU/Db1n9bIFJgmX5Ja14PBmStlKDQek8pHie9JDLlp4o2gCE5ss/61t947V8tt5DNFnj1YHsaGvR62quSmNwcrB9zU0gwcnnKTh5x4CT9xNwKux5OPl0z1wKsuqVmy6+zLNpxZ2jOn4nWKYSWHaYJdJBSKIfgkBgFLzIaXNBmQpJkAHmAgUDZTYEZiOdEZgsFYSK8kIx7bShrlwupr002VvFwoZQZMXLXR+98/mO77Ih8wU1zYrAo8hQFBnrNu4+4T+vfvShR++44Np6QnwoFEy43VVQPY0QRC8EQWQMLgHIF5vrImAsBkEoAOiGUzacsMrmqcuaSw+Uhiqlp7SiBIXT+8bGYM+h0tBOAGC/uOfvsaeefP099wL43W0P5eaf2Nxd39xQqqQXhlxzqD15ccva1nyss58PxeGmrAlrkZtuYsgPDxQFFAta2LvVdejsk8ni7244hSGz+LrSRBg58In8phSietF08WXuk75+W/DDv/lL+anfvyvUfPlyteniT/jhtDAH4RRaSpMJhou/o/HsCyZVl0ZudHznkoV7lu0R/N4IREEBBOzszVV85RdrTylVzgFANyyse7NTePX1KChlOGZeFWqqyvBUS1vThZ9/9IsP/mH9xzAYE6SrLj1lpHHO3AS6evm9cQADcCrOpYk+S+uTx+F89vfXzdMNJ7/ss3vmynse12KdsQE43xM2BrvDmsVt1opv3Ns9tICqx3pZ8Rgb7+5SjWJ6S2uo6/u49J4yg79+tkJUHrKONHqsj3W+/HwOTh7aV6Wx1LXZjcGyTAqDwyd0DJaF9iqHVi0+efiwkMPSWFcuXPu5M90sl/NAFFQIgtMIlc1VwaJO/BSEYsWaeSBABIMEytxIZ8MwzVJ6tMb68j0V8Pu+8sGKhUfXNrywM7v0zhc6vsOGTebbtjOGTVt6sCM6gFPf14hEUp/xo9/862q4FQkyqYXbdRQY9cO2CKjtBqOEMXggiAoIoTCtAJyu9QSAseGtjgnr3h4IqiPFogKc984L5/0uxQwdzvtcKG6XAeQfefCljm/990OH3PCy4pv3prS0phXPn8K7G4ayKE6wrKU1+3e3PcR7JHJTGq+gj5ItLDnSU6zSl1IpUNEY8tlH2DtJOMGl1L29VHDrh1Owy8H58vLNF4K8gj5FeCKVwin/7/s1CxZe0jhTOmqmZ4cxM0Lq5sw6/fxj/PWzTxEs1iDqtEI0qF+gexW+vQA8suolZfWz+ZipaWZhQ8h5QiSKAgiJQBCyb+1O+//jf169NJ4x3Ios4hv/Ngd/vvp4rPx/C/A/H2nCaU1OL9CBuIa7b7kAN1zzAbgVWbr1nhf/82/PbzkdxScqjc4THW4SWd3Sal+9/N5YKqUl4TwtjMP5HjBSKS1b/Hl/T3Y0AO5UKrfPArmW1nDjRTf0b1nb2oW9x3gi1hnL3PWln+58eWXLXgVYU8tBj/WVGgqGx5lSLy2zuF+E82SzDM5EX+VwGheVIa8V/PWzeRniCBT9+5ManMrQSHmUwamgUQxWWk0MPlEPwslPITjffXt9FmTVO6p56sIPL3KGgGQ1wKZ+mJYCo+CBafr2HMTAQGkBjGFPBZ4xGbalIF9wwSmTGcmURhvryoX1z30reOFZC+a9sDO79NqH3r6bMcifPKUGT3x5MVZ/8yQ88NkFWHZUEAAQT+rYtKUH1RV+vLA2elrcQhXcrgoIgg82I7CYAJtJsKgHtu0BkW3IMmBTBc5nUFr1yrbeiZwUdNPG9pHe59K2UjzIw6kks+K+HJw5l1JP/eX1vquX35t+9ykOXntrlN5+5a3xLWtbS407aTiVdXPIz4h1xozbr7w13t6670k2OW4qGLfCHSHE7fd5/82tKIsJkcsBmHnDaNc0/blMNrdhvNIxVt5Av9WObL4evuGtv4XiP2xFUr+DrU8DwHwEjeNRccDHDyrIgZbS4SYBF1Fx6vn/XRtw1wVhMsFWRR91i+VMYBE5bpcJ1KoVbBiSTvuYiAwjQtYKSJ1mUM4Wv95K4ztLQyK46cN5h2VZhG2TfAGua1as/3haNz2SKOCnHzsK7z8qDC1vgjEBteUqakNuzK/y4leru/DAY6/jY/9+HH707fNy3/mf56Qf/OzFL5y97KgXCJH4JHGT1CMPvmQ99ZfXB8779yWu+oYKGQDb9GbUrG+ISLf+6D/74Dx9HD7B2561g1MpTXjqydcLw887lJbW2O1X3Jqob25I1s9vcKllqrhlbWuhvTW6zyFRbSsfzi1Yfp0Lg/OkEDiV8SycyndprpQQnO6pQ3uA0SHpTgGwM3xN9COSqWUBpxJeWpasND7bhFNRK+ULA8WlyuBUzIf23ChNomjDKYtaAKDH+kY1T53x/rmE9cV8SCRVeFUXRJHAstywLC8kWYAgMlBqgrECGDUhyxSMOZ8hizJmmgnB+UywVa9sKzz+2y8GFzWEg7Gt0Yqv/2bDTQKg/M9HmvDBoyNOi4Qg4ISGMhx92dG4+9mdeODVbqQzeWh6AZQy6a0diblL55cnYFkAZTYYM0GZCFGQIMs2kyRBABMxOIlkz1du/P3BTx0/BopDdwr1DZHh5VEdTtmltL3UK5TAee8zAIzNG9v3G8sOVntrlN5+xa3pSG1FJlIbEQH0xTpjbP6JzUQtU4Uta1tNXjHnpotxqaC7FaWhvr72n9SmnXnDeE3T9W5RFCW3ojSHQ8FrNU1/LNrRef14pGUs3cPeTF8tHCfUwzd80h60I5v/GXtzzxqaW1hSP16oONApNfA10Ce9OfVLyezZy8pDZfUNSNtuu0zyMyKoYCwoFphbzLMZkFDGJMFkRBBFnRJmQyaGJQsmOgqVcumJlhjfsol/uUwzSa1gB1WXBEoZKHU99a/OxV1xvRIAltT7EfBI2BhN7jne6zYR9is4rSmA51rjeK7lHVz04WNQVel3XfnR47t+cu/LjWvWt89detKsHRs2d/CZtCepVFLDIw++tKeBFgA+ccVSEU5FpANABQYnCy3NR0EBsM0bo6n2aOygGuraW6OsvTV6UPMQdL78fCHcvCBde9qZATiF66Fd2X1wKlI2nAL30LlTSt2UgeIkYPEtm/rGalksbnLTY33M0nKW7Axx6IFTVnlXucfScgkAaVn1llYIGL7ygAUn3wfg9DRh8S2b9jWm/ZAc5zL9SGc9EEQblp2DyAgoFcFgoFDwgBALYAYEQQQDBaV5WLbzuFqWM4IgxAHgrnufz55+ylzXomPrXEhnlIde3HVRLm+WnTYngJkhBW+2D07HoBARNUEPlp9ehxe3JtAez8OynK92PW+GIaAPiisNChNgGiRJgSxLEAQRtm3CFi0QWQfQc/En/691w+aOCZ+H6IqP35le9a9bwti7522pt4QM570txTADI/euGBWxzn4W6+zf8zd5eWX/qDQAcNxkMi4V9Egk/FVN0+/f3dn9rkXnJVF0z2qsf0VVPUdrmv72eKRnrMSQZzextcmzUOdaLFS4AYgaTHs9ixmr0b1XAFEF2YTzhRTCuyvhpaBXwHtYM5IbXy6i4uxTbygLB+o9VBGPoUl7LpMElYlCxKVRL1UEi0mCCAEKKGTYcDEJNnMJBVGnlq2INhmwyi2/qFOPaHa+/Dyf7GsaenJdh37lsjk+WJYAy7ZWvd0/v7SvqcIDSRRQVaZAFICCzTCQNdEdd8qHi+v9WLmhHx2dKdRU+q0TFtYJALBtR6x66UmzcNe9z/Ma0hSyuqW1NEa8AGestxd7V1hMANn2aGzM+rNuWnGnnonu2D3/8uU1xWuXuqSWxseHMfjkE8V9Wew9Q7Ir+vcn9zVOnTsC7Hr2Sa3pok+UViVJw2nUcWNwItzCrmef7Nb7e+mC5deVjtPgDJsQ4OSx0nA+EYBiabls28qHRy1fXbVsjqvBIzgVSkl0eoZQasKmWQhCL0xbBgOBSwZMKwdZFsBYqZGcQRA6obh0trMdX56ruoVTFgVYe6cKUYy81tq3GADqw27YFKgJKnATCaZFMZAzsas/h5DPhfc1lqE9PtjmMKva1wfAAGMEoigCogwGE7YtgzECWzCYbYuAkIhuaHt719vRCRt7PtTmjVH79JO+Hf/fFct9xeXWSpNgEjjv/z7LL3/7y/57A3Ec927jUkGXRLE8mc48P9I+m9K8Te0dcCqq08Jz6Cg8xzr2G5C2sqQNAaXHZgr2noW31AqZwRi2QnKH5wMnXucLB+pVAHMEyhYLFkJMEkJigbqZLHgkm8mAINuqkGSyYIEyCYLgYgxBAFnBZi6BwSWnbW/BI/bHWzdlDnRNbuq5/sF12gUn1ClB1eVhhpHIFeieCtncKhXN1V6IogDGGERRQG1QwTt9GtK6tefxpU0pwgFPoacvawGAKAr5+3//Sub+37/CG/CmkPZojK1uadWLSxaZcLqElrqz0tK/H97y+JgOON317JOav2F2b+1pZw4fZlWamLQ0GZOKkcerp/0Ns8Xe9WvGMpncJNa28mG9avFJxF8/u9Rtfa/eIn3r12htKx/ON118mQeD8xoQOOWa0kSIe1hajq29/YbEaC6xduWyOW7oeRNGwYTiAijNQRB8kCQTspQFkRMwrTAozcHpmm/AsjRQRkHpADyeLfSNzRb0PAAQFAohxJN+eDyh3rgWAICagIJja32QRQFC8Y5qggraB/LoTRuw6eDtzJpR1nNUtXc3RDED0yoAlhsyCQCQUbCB0sz2zJWDFu+s19O51287v/wzv3wleX/L9gmP9Zs3Ru3TT/x2qr4hItQ1VEiBoIp7VnwuGCitMT+C1S2t+uaNvNs5x71X41JB13T9karKyM/ciiuSyebWoNhVThLF2kCg7EJCyALTtF4fj7RMFu3IWBosQ4WcwOAYntIXVmkGeGs96+dPVSehcKBerIrM9wGoAbAAFAEw+C1bCyU6t8zSMl1VDLboKatJl9Uc3SlGQp0CgwmbGRBAGIQQBFAmCGkpS5nlz8X71q/hrczTUFIrsBNu+Fvi+W+d5W6o8GW8EouX9nUk8hAEFJe9df6XRBFNlSreaE+jPZ6HKAqYN7vCCgY8xgurtysA8Ic//+sfz7ywceJmDeIO2RUfvzP75LM3SMWnUMBgxRwA6NXL700ebPf2w7FpxZ1ZPdZn1p52puqJVBIAgqXlTDhdkkvXH7okG+A0GPM4xQEAVt94bbrp4ssKxTwkA4Ae67M6X34+N8KSoaWuz6Un5y4MmUQu3b4jnR79OQ0kAGD9A2mhqiIAl9wHQRRg2ypkogHohwAdRsHpaq/nNWaaWQDJVa29W09vrtpzDgAemJYLiguQRcLgzPyezlsQiyU3Z445pyDXUO5GrmBjS4/TyUkQBPr1K5Y8Awg6IJhgtMuZGI7qAIIQoEIQKSQxg2x2B+uLtRcnsnP9+IoTsKq1N7arPzsp5qdpj8ZYezRmAcC3Ag8l71nxuRBGmHS6PRorXL38Xv7ggeMOwbhU0Ptj8ScBWD6f90vhUPBmOGPv8gA6MtncM+3tnctM0xzVcUeTnQYLT7CdmU8IRwUxuNasiCHj/tqRNVajm1fQJ6G6miWlGY5nAvAINlP7+95Y1NH13Im2bQxOpNL1BoStzzRH5i3dVD73/RslUVFE0SWIBnUxKshMgsl8ohzfssng4zmnr139WfbE6x2pa85tJsfO8L727NvxD7tkEY+t68PHllRiRsCpqzEG0OITl90JA+vbMzh5cR1ddHSVtqm1J/3YkxsbZtWHX96y9Z0pPRzoSJZKauyCs29L/r+rz3X/xxVLPfUNEQkAe+ovrxd+8bO/51a3tI7b06a2lQ8bwytSzZcv9zacfYFvyKbha7kDAOKtmyb8iR438dpWPpxvW/nwPstvmeiOfc30vle+i2/ZPBb5yfksWRZl3b1J+H1uQfUUoLp9IMQDUxRgU8DvjbNk2oBp0qRWyN/8pzcHGiu8wunNVaWu+SKAMlYoWILiUtr7cuWdsdwMAHh0XS8+urgStSE3BDgxHAAgAD3pArb0OO2on7rgmFfPOKmxjdl2BqYVh2VrMC0LRHaDsjoIIKAWUCikWTLVC5v6UOxRGVRd7IIldZm7n2mddOXkRx58yVrd0jrwtW9fop53wRIlEFCF9mjMfvTBl/K/uOcZPZXk7cgcdyjGbRb3/lj8qf5Y/Knxut5U8Bw6DBVy8kLMKsNgF0cAzqRy97A30xrv4T5ZlcbbBQBU9CbWHb+r55nTJFFBY825dkBtZKJI0J9+E7t7W+T+LasW9m9ZtRAA3GXV8arqE18Pz1zSwogQYJK8u3Lxyd7a087Md778PH86NU2tau01rjm3WRYg2ABw4cIIVr7Rj+UPbsUPLp6NBbVOnci0KVa9k8RdL+wGAFxx0cLcH/662f7xL1+qp5Tlb/n6Od+451ci2fX7V3hemaJSSQ133PJ4/o5bHp90Be7Ol57XG86+wIv9rMOux/qM+JZNEz5xFTf5xbdsMvVYn1nspbEvtPOlf4z6Z+GJdR3G6c1Vzso6lDKk0jpLpXUACQhwVtagzD7ze0/HTz+6Wlr1do+9MZqwkloBz3/77ACcp+delJaHMwqAq+BLJK0gACyo9bHW7pzwuYe24LaLZmPhTD+EYj/3NTvTuPmvOwEA8xuC8S9fcuwbcLstgcjdrKNLZ30DPYJfVcFogBWsFCiVYZom8kYajOXgfP7k4j/XhSfUDUzGCjoAtEdj9Orl92axHHztcY4bJXwN3Qn2BNtprEZ37HhUKKogSwBYO8sW3kA/r5lPbiKcz4/boJmaaO8/TpNEhc2v/w8qCrJoWjkhm+9CV/8rEAQBrnANxDIfaDKVL6QHAtFtT56ViG+pmbXoE/cIJvNIGq1pvny5O9O+MzoG3fy4SeCJdR1WtD9rlPuICQBBj8y+dMZM4Rctnbj8vrdxVKUHtUEFbf06dicGHy5d/d2/+RljqK7wd91z+0VfWHJcbVtjfTlff5obE+n2HXTL71ak5l++PIARKumWlrPfuOvWw1rTmDtymFoOm1bcmVp87bdCsuodaawy27TiztRojj0veaBle/47HzlODaqudzcOMACmRe9+pjWzqrXXWtXaO7zMJcNZGq603JnzMk3PewVJA4AF9YHMh+aH/P+3qlO4/L630VThQV3YjZ0xHbsGBuvS135s4XqPz52DR8lBlrtXbe7avsxP/WygkIcABggyGHPBGUJSWmN+KBJUXUE4EzlyHHcE4IW8SSCGPHsOHfkn2M7cE2ynxivnU4IJAEwSXP3ZN4+jzJKrw+8DY1SizBYYKNvd3wJRVlDefBogCdB3boOR6HVT23TGrsW2Hdux/alzBcpkOWX7ZdUbPPGbt1WX1c/mn8tp6onXO1Knzg5sl0TBbu3RtA81hwo/umQOzl9QDstm2NSZQ0gl7FNLZ1qyJKAi6Ml/5Kx5G3/0Xx9c8fcHr/rKyYvrNwHAZFh2h5u+dj37pPHGXbfGM+07DAxZB71v/Rpt9Y3XxHkjIvdexLdsstfefkO8b/0aDYM9BVmmfUd+7e03xMeq51hSK+DMW55NRvuzI52fPdCyPXP9g+tGnDW+MeJVMVhGHoy3tm01hVzRqoDS1bItTi59f23sl1c0F85fEIFNGd7cnYUsCvjUqbVWY6VakCXRPmFJwxtwK10oFLoB9F5/zwuxputW7nqgZXu5Q+ajAAAgAElEQVRHtC/bDcZ0OOPy09jHHA8LG0KuxgrfPnu1cNx0QQjxhkKhj5eXl39GUZSaiU7PRBmXJ+gNdbU/UVXPJ/d3TOvWtvB4pIXjRkNWi5kAdKoIalbrLAOAgHeWIEse5lEqsKvnaYFSE7RgIfZ2CwDAXTcLJFwBUXZZerQNhViPnNi17vzy4z/4jOQL9woWk2TVixO/eZur9Xcrunh39+knqZl2hZ8MnH9cZPUTG/qXndYU7L/8xCr55DlBFwDJ7ZIYRJHd8MR20bKZfP1Hj1138XnHPg3FtRuq0g9AT6Y0e9Ur2/j4X25M9a5fY/WuX5Mkqhf++tlCfMumSTFBFTc1pdt30PV33ZoBkAnPXyBk2new8Zh3ZWM0QZfc8LfEhSfUuS5YUqcEvS5hYzRuPbGuw1jV2jtiQ+eFJ9TJDRW+oftKSyM688uYVuGzZzQ8fesT2z5z29PR+C0Xzupf3BhQ8wVbhiAIHpdk/+bVbntXnxa+7Ky5a/0espOZVgyW3Rsd0BMbNncwAOzTv3wl952PLKTfueS4KjhD5vaFAtAbIl55V3+Wx35uQvl8vmZFUc6BM8QzZRjG37PZbOtonFtRlJl1dXUtALK2bcfLy8vv7OnpuSCdTv9zNM4/lYxLBT3a0Xn9jOqq7YGA/0vRjs7zx+OaHDeWCqZGAXgYEdyUWgQAREGGR6kAoxYG0qVYxeCum4XIWRdDcLth5zJgti0HTjq9kFrboqfWveSJvd1yftXpF/1GYBAZQGXV612w/LoQUb3JXc8+yScJnBosAO1Dft4XCiD1rX+b/ei2Xn3OD57ZVfvK9lTygoURY0ZQIdv6dDy8tkfZ1pNzXXRK/TsXnTxzM4AkVDUBQdgFgH3vf/6STqb4xDvc+DC1HHjlfNo52Hg1JsY7PyW1Au5v2V64v2X7QTV6L2wIy3CemmcBlCZMLC1BKDLD0C97/4x/vrYzecyTr3efvLUrk/3ESTXaUZUeuzNVEFdu6Jdf3ToQXtAY7Prvjx79V1YwOwGkobgyX7nx93uN036gZbvxnUuOK60jP1LvOQZnaTpeFuAmlKIo4erq6hWKolwyfJ9hGI/39PQsNwzjsIZhVFZWfse27Wg0Gv0QpdSeMWPG7ZWVlT9Lp9PHHs55p6JxG4MeTyQfCAT8n9U0fcd4XZPjxkI4UC994MRrQxAgMEEwFTWSQmoHbFpgAgTs7H1WsGyn15yoeFDzkU9BVL2gjEL2B1AY6IOZTrgCJ52eyW7ZgOS2f51TfspZTzEpYMMZf6YBkOZfvrxMj/UN9K5fw7uTTnKapre3bm1rONBxG6NxC4Duk/HWQ8sX/OaHT++84E/re49peSexZ2xmUCXGNy5pfuPKc+a+CiL3w+/bAVnaAiD3lRt/n77rXt6zguO4Q3ew8epIlcwVSg0IOpxGVRVOeTkFQAVjgpDPR3/6H/N/uqLW/7FfPr/rvJv+/E5l6fVEFq3LljX8678vmv+0R0QbKM2A0sL3fvK3XU88vWGvBpFd/Vn6xLqOgQtPqCPF6wxfclcDkABAo7EcH/7ITQhCSFldXd2LoigeU6Y2vFwTPjEpSe4B286Xd8fXBtOIXlRXV3dUNBo9zTTNg5qjpKys7GJFUVQAZaFQ6LPbtm1b4vF4LhwYGPgKpdQGgGQy+XOfz/cNVVXnaZq2dUxvcpIZtwp63jAyrVvbjh+v63HcWDnjxOv8ACQwiJCh+6uaOvq71x7Xl9oAValAf2ojSHklzHg/fEcvgqh696xzDVGEK1LFjK52wcplvN75C5Fe9zLJdrUdpdYs2QZnveEwnNZ7evy13/IA0PvWr8m3rXwkx8d+Tm3FieK0hgqf4BHZyzddMIdef07jqjei6bredME7K+JJLqzzR12ysBPxRA9cru13/fWtHbv6M4Unnt5Q2NUxwJ9kchzHjaFVrT1Du5GX1m6X4FSc+wAAjHkEw0h87qSK6FUnRB5dH00fszOm18+MeLXjZwc7ylRXFhISYKwfmWx/Ykdn791/fD010vUeeGl7/MIT6txwejOIGFzVx4bz5N5Y1dqrT5Z10LkjT3V19a2iKC6oCZ/4REPV2QE4PUt8ABDyz7Oivc/+pTu+9sLq6upbOzo6vnww5ySEHBcKhS41TTM1MDDwsKqqFQAqdV1/q3SMpmntALKKosznFfRx4laUOaZpdtiU8qdB3JQRDtSLPjWioNTKTSHL3kAKAAZSbwlGIQnGKNSZc5Aa6IOrvBKMMTAMfq8KkgRJ9TErmxZJKAIAKBgpojpfzgTOWLQQnFZzCYBYufhkNTx/gbL29hv4BE1T3M2Pv9n968+fMhuMdTLLetUvC7OXNQXaBUBkjBlg9gAz0QPT6lz1RnvHV259li9dw3EcN042RhN0VWuvdnpzlTpk817j1ZNaIRPtzw0sbAjNUETo759V1v3+WWUBAG4I1GaFgg7bTiCTjSVTWvrMW55NJrWRi7tPrOuwHmjZHrty2RwDThlAQHGe+eK1zM/88hX+PcBNCEKI6vF4PqWQ4Lpi5fxdGqrODsQz29YB+BQh5OumaR7sOLxANBo9nlJqqKo6u7gtM+yYuCiK5Yea/qlqwmaLrqqM/EpxKzMn6vocdyhcZM8yMTIAt2Azw11WtVsiHg0AsnonAICEnd5uhURsr8p5iSA7bWNU12wAkFSvAMAPoAxAHs4XtILBGZQhq17p+Gu/VTYW98WNjBCiqKqnWlU95ZIojsoMuve3bM8/0LK9Hc77vBuMvQVKdzJKd4KxHQDaALStau3d/ZGfvMgLZRzHHZSxiFdHqo/85MXsxmhixBp1UivQ/3pwXXLJDX9N3/z4m+1JrVCagb0DQBso3Q7T2rnqjfbo9fet6W66dmV8YzSx34b1T//yFe36B9clov3ZHJyu7SYA+4GW7bmma1cm+NNzbqIQQo4D4K0IHrffWR2L+73F4w+KrutrKKXD51cYHrtcGNZAdiTg66Bz3HtT+pJVAAhgSEmmkKloeP+GnrYXThElF6V2QbSNHEh5JXJb3kTw1DMhebzFkCMwZpmwdU0ABBi9nXkIgqo2zO2Es+ZqAYOVchnFFvQST6RSCc9fIMW3bDrigtV4qqqMfD4QKPuSJIrHYLAhM5s3jH/09sW+rmn6tsM5/6d/+Up2VWtv9Mplc9TTm6v6AIirWnsFAMbGaNwozjDMxxtyHHdAYx2vjkRJrcDOvOXZxDUfbnZfuKTOs7AhJCW1An1yXUfh5sff1Hb1ZykA3PynjdrdT7duv/CEOvfpzVWuhgofW9Xaaz3Qst0oHXOw7n6mtXD3M62Fxgqf0BDxYlVrL6+Uc5NBaYWBGID9PcmODTv+YOx5CGHbdg8AJopiJYDtACCKogQgbJpm13s457QwYRX0vGG8TG069mtscNwoiqeiNpyWPAEAk/I0Yful6qrGpeuy8V312fiOmQBgdXTS8KLT0fv8H8TY039Exb99HIIkA4DALBPUyCO1tsXW3nnL629evJl4gyZ1xrmZALxwWuMNjDDDbrh5AeEV9LFTEQl/MhAo+3JvX+w6I2+8ljeMlCSKouJW6vw+7wUNdbXPtu2ILjBNc3g3rPfk/pbt5v0t20cck8hxHHcwxiteHYmSWgE3/2lj/uY/bcwf6Lj7W7bn72/Zvt/jDtau/izb1c87T3GTRjsAGIVUHZxJC0dU3L/n+PfKMAyNUrrZ4/GclM1mXwUAVVWPByAZhvHGoZxzKpuwLu69fbEb84bRO1HX57hDUTA1bNy6stSwlBMKLCtqNCMIkjDr5CvWVDQt3SmIMs10tIrpXZtEde6x0N55G10P/C8yb66zCz2dVuq1l2nsb49B27ZZcoUreys+eNGzTMDQlvbS5zIxzrfHAVA9nvN6+2JfS6XS/8gbRgoAbEqppunR3r7YzzRNf5UQmU94yXHchOPxiuO4saRp2g5K6Yb+1MYGLd874jJqWr433p/a2EAp3aBp2iGv1pXNZn8VCASuVVW1nhDiKy8vv9UwjMcNw4gd+NXTy4RV0Dluqtq4ZaWe1WKlJ9xZSadJiFAESRaqF5zT2nze11/yhGcm9e1bYPX3mkp5TcGM92PguT9LXb/7uZx85XnRSsVp2THv21B36ZdXSMSTgCQU4IxJLsCZtTVb/P1d4q2bzJG2c6PDpnSX3+c9TxLFd8VHt6JUKm5lEYAjrrsVx3GTD49XHMeNtYGBga8DqNi6+w9L0lp0YOi+tBYd2Lr7D0sAVBSPOyimaUbz+fyWodv6+vr+1zCMv86cObN11qxZA6Io0p6eni+Ozl1MLZN2EhFV9cxuqKvdDgDRjs45fP10bjJZOP9iz8J5F4epIpSZ5fLxEIR6WxUrba/oY0RwsUKB9G149thE29r51DJdpdeRsnAs9L4PvOprOvZtyePTAFhMQpK6xSSAbQA8xUN74ay5updM+w5j9Y3XJsflJo9QbkUpr6+v/ZskijM0TV8Lp8EEoiTWuhVlSSqVua2rp/dHE5xMjuM4Hq84jhsX1dXVXygrK/sZgILXXdUmie4em+ZrcvneOQBc6XT6yz09PT8fjWsRQlyiKBLDMI7YodC8gs5xh8BFVJx96g0h9dimauYSFoLBQ4lQxYhQBgkykwTJVkSD2QWST/XWWHrGJYcjHUpFzW4mwC0wZoDCYC4hb3tEDc4T8+1wJqEjcCrne81saWk5a+3tNyT4MmtjTxJFQVU9J7jdymIAFXDei/ZMJvfPvGH0TXDyOI7j9uDxiuO48eDz+Y4vLy//hqIoZ8OZ2DhpGMazAwMDP8hms0fcOPGxxCvoHHeIyuccK5722Ztnkbj1PjD4mEtQQUGYDJm6RAoRFAIMJsBgRBBtn9QDIA+GMiZCYETQmSzocCaG6wcwAACWlkO6fUc+PH+BC4BoaTmrd/2afNvKhzU91sdndeU4juM4juMmjKqqXk3Tjtgn3GONL7PGcYeIEUEwZpAscwmdcspuYDLARMGGsPfC5wIFA0WKEiHGiFAA0A1nvXMJzpJqaQCliTfophV3JnrXr+FLbHEcx3Ecx3GTDq+cj61JW0HXNH1H69a2SfuEn+P0WC8FwMyg1CfqtArCyJ8nJgl5AJ2MCANweq3YcMYJEjhdEXUAVt/6NUbbykdyvAv7xAoEyv7NReQ5+zumPxa/e7zSw3Ecty88XnEcx00/k7aCznGTnR7rY3qsr+CJVGYKFfJukrRnCRZzoTR0RACjsmBAFOJWmdQLwNjyuxWJdPtOCkDQY70W77I+KeVDoeAnJVFs0DT9rxOdGI7juP3g8YrjOG6a4U+oOe4whOcvkE785m3lAFyiyeaIORoQLUYAgImwmSxotkfcxoigW1rOWvVfnxkwea+gSc/v886fWVvzaOvWtkUTnRaO47j94fGK4zhuepEmOgEcN5XpsT6Wad9pVBy3RBLcrgx1izZVhDz1iDnbI/VRt7gTkmBaWo6uv+vWZK57N+++PgXYlj3AwJim6a9PdFo4juP2h8crjuO46YU/Qee4UUBULyoXn+zyVFS5wvOPVcPzF0hwlkyjfevX5PnYco7jOI7jOI7jOI7juMNUVRn5vltRqiY6HRzHcQfC4xXHcdzUJk50AjiO4yY7t6KcJkqid6LTwXEcdyA8XnEcx01tvILOcRzHcRzHcRzHcZPAtBmDrqqe2ROdBo4bbaZp9ZummZnodBzp3IoyxzTNDpvSwmicj8crbjri8Wpy4PGK4wAjb3TZlOYnOh0cdyimTQW9eV4TX0+am3ZiA/HP9Mfi9010OrjRxeMVNx3xeDU98XjFTUXRjs4PaJr+4kSng+MOBe/iznEcx3Ecx3Ecx3GTwLR5gs67YHHTEe8yOj3xeMVNRzxeTU88XnFTEe/iznEcx3Ecx3Ecx3Ecx3Ecx3Ecx3Ecx3Ecx3Ecx3Ecx3Ecx3Ecx3Ecx3Ecx3Ecx3Ecx3Ecx3Ecx3Ecx3HcqJAmOgFTlap65hJC6kzT6iltk0RRCAUD5/l86lkAUqZpDUxgEg9bIFC2uMzvvYQQohqGEZ3o9Iymikj445qmvzV023S530CgbGmZ33uBW1EaCoVCG2PMBqZf/uQOHo9XUxuPV1M/fx5pRsqzquqZGQyUfcKtKLMKhcI7jDE6UekbC0daniWEeP1+39mGYbwzdPt0iU0j4fGKGy98HfRDIImiWFUZudWrepYN3V5fX3tvOBT8LpFJY0Nd7cuBQNkZE5TEw1ZVGfnPqsrIn4lMZlZVRu6dUV31zYlO02gJBMqOj5SH/3fotulyvw11tbdVVUZ+I4pidTgU/PycOY2rJFF0AdMrf3IHj8erqY3Hq6mfP480I+VZv897dENd7etEJs3hUPDz9fW1f5FEcdos9QsceXm2ojz86WCZ/zNDt02X2DQSHq84bhILh4KfbJ7X1Nk8r4lVRMLXlLarqmdO87ymBCHEDwAVkfBVTbMb/zlxKT08zfOaooFA2TIAUFXPrOZ5TWlCiHei03W45h41+7XmeU2F5nlN8aHbp8P9EkK8zfOaDFX1NAJOxWzuUbM3h0PBi6db/uQODo9XUxuPV9Mjfx5J9pVnG+pq759RXXUjAEiiKM89ava26VSJOZLyrFtRFjbPa2ptntfEGupqHx+6bzrEppHweMWNN/4E/T3KZHOPRjs6j80bxqNDt3tVz9K8YbxqmmYGAHKa/ndC5PdLojjlhhEUA1A4m8m+DACapu8E0OVWXIsmNGGjYHdn99m9fbETh26bLvcriaJb0/SfaJq+CwBsSim16YAkicp0yp/cwePxamrj8Wrq588jzUh5FgBU1bMsk83+HQBsSi0jb7zgVlynjn8Kx8aRlGdN03wr2tF5SiqVuXro9ukSm0bC4xU33ngF/T0yTTOvaXqC2rQwbFc1tWl/6RdN03sAKIpbCY1vCkdFNYC4Teme8WGapne73UrNBKZpVGiansgbRnrY5mlxv3nDGIh2dH4TcMZDzaiu+oIoiQ3JVOYpTK/8yR0kHq+mNh6vHFM8fx5R9pFnAaDGpjS25zhd73YrUyvPHsARk2dtSi1N0xOmZerDdk2L2DQSHq+48SZPdAImM0JIVTDg/38AUDCtLalU+vf7OXxfY6mmYgvadLqXgzGt7tfv8x5XVVlxDwC0t3d+0DTNNKbZPXLvxuPVu0zFezkY0+p+ebyaugKBsktdRJ4PAMlU5hemafbu5/CR3tPp9H7yPHsE/A14vOLGC3+Cvn8mgI7iv9gBju0FECj9oqqeCgCWaVrxfb9k0trrXgDnfgqm1TdB6Rlr0+Z+KyLhj82srXkylU7/rG3HrmV5w9hR3DWd8ic3Mh6viqbq5/cgTZv75fFqyothMOaYBzi2B0PfU4+n0pyCeXY/eJ6dRrFpJDxeceOJV9D3wzTNeH8sfl9/LH5fKpV+fn/H5vPGalX1nFKa0dGtKKeZpvWaaZoH+tKadIy8sRNAzu/zHg8AbkWpANCgafrrE5uysTFd7pcQ4o+Uh38e7eg8pz8W/8PQfdMpf3Ij4/Fqan9+D9Z0uV8er6a+VCr9fCnmmKa538pI3jBe8fu8Z5R+V9zKqTldXz3miRwnPM9On9g0Eh6vuPHGu7iPkkw2tzVvGP+sr699IJvN/TMUCn4tFot/aaLTdShsSlk8kbylpqbq1+5E8n99Pu/lmWzu7mJXnmlnutyvqnpOAuBvqKvdq2tzbCB+fX8s/sJ0yZ/c4ePxauqaLvfL49WRJRaL/3Bmbc1fAJiEkGOoTdPZTPYfE52u0TKdYuqhmi6xaSQ8XnHjbVqtQTme3Ioy36Y0a5rm7tI2SRRJOBy8EsCMgmk9l0ql10xgEg9bOBQ8V5LEk2ybvp1Kpf9oU8omOk2jQRJFt+JWlmja3q33U/1+CSERQuRjh283TWuzaZqx6ZY/uYPH49XUxePV9MifR5J95Vm/z7vA7VYuAJBIpjIPmKaZnaAkjokjLc8SQqolUYzkDWPz0O1TPTaNhMcrjuM4juM4juM4juM4juM4juM4juM4juM4juM4juM4juM4juM4juM4juM4juM4juM4juM4juM4juM4juM4juM4juM4juM4jptI0kQnYKqRRNHn8bhPIYQ0DvnXQAixTNPKDD/erSg1oVDgfK/qWUoIKaeU7qaU2sOPU1VPAyGkito0ztj+V6RwK0qDS3Eda5pWx5DXVxNCGgFBo5QWRnhN2KW45gBCjlJq7u/8hBCX263Mk0RJtGw7t9/E7P881W63cgK1aQc70E2NEUJIxO1WTpREyWfZdt+BjncrSpNLcR0HCBaldEzW7hyPa3AcwOPVe8Hj1cRdg+OAsYtX78VI8eoAaRY9HncTIaSC2jR5oNhRjFdzCSE+07SSh5pOHq8m7hocNx7kiU7AVKO4lTkNdbX/HGEXyxvGI+3tnVfZlFqSKAo1NVW3+H3er2Hvv3NXV0/fFalU+gUAcCvKzJm1NQ8TIi8FAJvSd3r7Yv+RSqXXj3R9SRRd9fW1zxh5Y01U61ztVpTy4uvPLh5iaJp+9+7O7m/YlFJJFF0za2t+oaqeqwCIALRUKvPtrp7en450/qrKyMfDoeDPAYSL9/RYe3vnp21KNUkU1Zqaqtv8Pu85ecN4JxaLfzWTzW0FAEJIsGl2w5vRjs7TNE1vd7bJQkNd7VPxRPKG3r7Yne/9r334ggH/0kh5+HFN01dGOzovOdDxVZWRq1XVc21sIH5tfyx+91ikaTyuwXEAj1c8Xh0+Hq+48TLa8aqEEOJtmt3Qpmn676Mdndft6/rD49XQfU2zG58gRG5o3dq2qLQtECg7o6oy8ltJFBuGXP8zqVT6mZHOX1UZuSIcCt4FIAQANqWbevtin0yl0ut5vBodPF5x04U40QmYwvTYQPym2ED8pkw293MABbeiXBaJhD8JAJFI+Hq/z3sDADuVyvwkNhBfrmn6wwBmzKiufFJVPbMAYGZtzX2EyEs1Tb8vk83dIYni7KrKyBOSKLpGumhVZcUXJVGcF08kf+z8HrmDEPls07RejCeSX7Up7VJVz1fD4eBni8d/X1U9n8obxtPF/fFAwP9jv8977PBzq6pnQTgU/B0AO55IfiNvGH92K8qlM2tr7iim9S6/z/sZTdNfJITMq6mpepYQohTT8fVMNvdY6csDADRN79Y0/bfhUPC7hJDgKP7tx0wynXkmNhC/Kafp/5rotHDcKOLxiscrjpsqRiVeqapnfjgUvLRY6a8+0EWHxyu3otQGAmXnNtTVPkKIfMHQYwkhyozqysckUaxJpTK3ZbK5nwComVFd+QghRB1+blX1zAqHgvcBUFKpzM2apv9KEsUFVZWRxyVRJDxecRw3FH+Cfujy/bH490q/zKiuygYC/q+6FeUUQsiD4VDwWwAQG4h/uj8Wf7h42K+aZjeqhMgX+X3eT1Cb3keI/CGb0m3Rjs7PAMCsxrq5bkW52Of3nZFKpZ8dekFJFKVAwP9fpmm9mMnmNgOAqno+DMCOdnReZJpmKm8U2mZUV65UPZ7zCSG/DQT8nwfQ2d7eeZFNqaVp+t/cbmW+PULXn2BZ2ccAyPFE8tu9fbF7JVGU5h41e5eqej5NCPlvVfVcmEpl7uzq6b3R7/OeNLO2Zg0h8kJC5C6/z/vJth3Ro4efM55I/kJVPV+oKA9/oaun9/aR/pCq6mkMh4LXSKI426a0J5VK/zKTzb1BCPHPqK78KQASTyR/ksnmNlZEwpeqHs95NqVtuzu7v99QV/tzAGomm7vT7/NeA8Cv6frK/lj8d/t64wKBsmXBMv+VACIAspquP9kfiz8GAG7FVe5WlNkF0woAOqoqI59zK8qpmWzuQbeinESI/D6b0u5YLH5H3jB2AQAhxFdRHv4yIfL7AOiZbO4P8UTyz6XruRWlJhIJf00SxdmmabWIksg/d9x44/GKx6tdAI9X3JRw2PFK0/TbGupqHwWw8GAuOFK8qqqMfFVVPdeOdDwh8lEAKvKG8XhXT++3AKB5XtM5AI4hRJ5tmubmocd7Vc9ZAORiTPouADTNbjyBEHmR4laO4/GKxyuOG4pn5NFTGidJVdVzEpwuTH3xePLRoQdFOzqvIUS+i9p0t+JWjgUgGHmjtbTfyBda3YpysYvI8wHsVeD1+X0fBDAzk83+oLQtNuB04TFNM1U6RWkfIfIxAAKapq8OBMrO8fu8H8wbRjSeSN1nmmZ2+A0QItcAQN4o7AIAm1I7bxi73IoykxB5nk3pLlX1nOn3eR/x+3wfA2BRm7ZXVUZuSaUyPzZNMzH8nJls7k2b0m0+v/cK9OBdXyCq6pnTUFe7BkAQQCuAc/0+76e6evrOT6XSz9mUdvh93psIITPyRuGjkfLwPQDKe/tiZxZffymAkKp6zgPgBuBVVc9HFEVp2t3Z/b3h1wsEyj4wo7ryueKvnQBmqKrncgD+/lj8125FeZ+qeq7UdP11AM+6FeVUVfVcWaxYBOH0OpFU1fPB7dt3NQOQZzXW/UMSxZMAvAMgrKqey9yK8t2unt6bCSHh+vraNZIo1gPQAJwP4JDHyXLcKOHxiscrHq+4qeI9x6vi7+cDcAXLyj4aCPjv2N8FRopXvX2x74uSeLdbUeqqKiMvDj3eyBtt0Y7ORdSmCcCZNwPOU/oB07R2DD9/TtM3AvGbcpo+tPv7nvk3eLzi8YrjhuJd3A+dUhEJX1MRCV8zo7rqu4GA/0sAkMnmnncReRYA5A1ji00pHfoi0zQ7NE1/MW8YbS4i+4ubU3v2W2as+GNo+AW9Hs/pxWvsGRvVH4vf0R+L3wEAkiiqFeXhbwCApuuPE0KqAEBVPcuqKiN/VVXP9eFQ8K5ZjXUvS6LoHn7+vGFsAQC/z3tR8XVNbkVZXNwd7u2LfYkQee7M2pq3AgH/dalU5huiJFYobuXM/oH4/+3rD2XkjdWSKDarqqdq+L5iS3gkNhBf3rq17bjevtg5AFxVlZEfA0B3d+9tNqWbCHa13D4AAAsMSURBVJE/1FBX+ycAEU3TfxlPJPcap5bJ5v647Z0dga6evjMAWH6f92uEEN/w6wXL/B8BkI4NxC9r3drWEBuIXw4Aqsdzzr7SX/zbvNq2IxqKdnTWAUhKojhXcStNgUDZJyRRPEnT9N+2bm2bu3NXxzwAsUDAf6NbUaoqysPXSqJYnzeMldve2RHa3dm9CICwv2tx3Bjg8YrHKx6vuKnisOMVAGiavlvT9B1D4tQ+jRSv8oYxoGn6jrxhvGvCOJvSvKbpG/OG0T6ztuaOWY11OwCEYgPxT5imqQ0/XtP0f/XH4t/Tit27w6HgBYTIJ9iUthl5YyOPVzxecdxQvIJ+6NRIefiuSHn4rkDAfxOAQN4w/pRKpR8DQA/w2v3Z52ychMhNxR93vysxqmfOnDmNLxEiL9M0/Sf9sfhvXUT2lPb39sXe17YjGjJN65+SKC4MBMo+Nvwc8UTqfgB9fp/3C83zmqINdbUbMdjLopBKpV9r2xFtjHZ0Ltu5q6Oxq6f3xzOqq25NJJLfU1XPCXOPmv1687ym/oa62ofcihIonTdvGKX0zhp+Tb/P+wEAIDI5tiIS/q4kiWcAMCVRXOBWlJBNqdnd3ftpABYh8pk2pe1dPX1ff3fakz+0KaWpVLrFNK2XAaiq6nnf8OOiHZ1X79zVsQCAMrO25kehUPCm4q799ibJZnNPmKaZ0zS9G0C0uNnj93nPKN1KRST8Xb/fe7VpWgkAsqp6TlFVz4cAIBaL/9CmtJDJ5jZpmv6X/V2L48YAj1c8XvF4xU0VYxWv9ml/8epATNPcljeMFwAIoVDwXkJI+b6OlURRmFFd9d9VlZE/2ZTu7u7uvdim1OLxiscrjhuKd3E/dNnYQPzK0i/5vLE7k829Vvy5DQDcijJfEkVxaCuvqnpO8aqeswqmtaFgWnpx856WSFEUS4F9pG46pYmY9lpGpCISvjRSHl4BALGB+H+Wxgfl88YAAGia/kw8kVwHAKl0+leR8vAHVNVzXDyx9wofpmnGdnd2LwuHgjcpbmVWNpP7tap6PlScsbm7eEzWNM2XAGeskSiJjZlM7i+zGut2mKa1IZVK3x0OBf+rqjKiRTs6PwcAQ5Y9eddTMAAB51z+szBY2H8bAERJ9AJI5I1CG4AEgApq062maY60dMaemzFNs4cQGS4iB4YfVBEJXxIpDz8CwMobxmojb7ypqp7mEc53sAIAoKqe41XVU/qC1wBslCSRESJ7AcCmdGgL/sBhXI/jDgWPVzxe7Uk/j1fcJHfY8SqVSj/xHq85Yrw6GL19sV8D+HVDXe19qur5VDDgv3ykGcTdihKuqal82K0o5+QN4+/d3X1X5g1jz/JkPF6NnH4er7gjEa+gHzqzPxZfOdIOTdPXAegHUBkOBy/rj8UfAva0nN5MiHxmPJH8Wiab+xMAKG5lTwBTVc9cALBtum2EU5cCjw9OQEVFJHxxpDz8iE1pV3d371mZbG7P+FCb0u0AQAipKG0TRbEGAEzT7JNEUVTcTkuspukJVfXMcbuVo/sH4t/RNP0dSRRdc4+a/WUA/UbembCjRBJFoaI8/KNYLP5NVfUsAVDW1dP7OU3TtxJCAn6fd8+SG4qieIs/xoffkE1ppySKoWhH53mapncW/waNAEQjb3QCwIzqyjsBVACIEyKfVREJf6o/Fv/N0PMQQhYAeoskioLidrq55jR9h1f1zBl6XCgUvAGAa3dn9+mZbG5NOBT8aHGc1SGxKe0CgNhA/MbSZDVuRakSJdFrmlaPz+f9tFtRFnpVzxJN09uKaT2oSWs4bhTxeMXjFY9X3FRx2PEqNdKL9+9d8Wp/wqHgOVWVkUfyhvGnnbs6lgOAaVo9xd3e4fFKEkW5vr7275IonpDJ5n7c3d37teFd9Et4vNqTfh6vuCMWr6CPAZtSM55I3hgOBX8RKQ//WvV4Ttd0fVugrOxcQuQPAhiIJ1K/NU3z/7d3b7FRVGEcwD/O7OmZa2dndqCsLaXWDbWmaiCg8RKNxiiJIUp4MAQCkuCL4IMmJhqDGIiJCReNDxq8RV9IJMYXAygJJoIR5JooiSDXQst2t2vb6ezOzOnZHXzolmyXQluklobv97Znz5xzZjf5z24y35zuUhQdlwiZ21Cf/DgqRV0yY4sBoNfLF36uHtcPguOqqgCltBUguAQAYFnxDTBYc1OyrfgntjW420bI+a+ZbG6dEMV9lMaebKhPbhVCdNhWfB0AlLx84Qcms6bZs+rPAgD8derMNJmxOU7C/t6yopO9vX0f6rr2NAA4Xr7wUSmKht3KatvxJQDg9/T27ZQZm12ef61EyHeqqiwSonj1BzulsTYACIQonqo+p7xX2GGaRlvdDGdTJpvbYlvxZwxd+0CI4oEz5y48alvx51RVWVmKovZ0OrOkoT550EnYWzyvsDvkfOhiCHUznG01NLZRVZSnJELmlKLoBA/5n9UXECiXdZhm7QuU0jrHsTeM+wuu4OULOwxdW2NZ8TfDkHfIMpvuJOwvAICcOdd+Tz5f2CEztsiy4psJIRqldO7QHtII3Q4wrzCvAPMKTRFjzavxjjtSXt2wvx8cAwBZZmzFXTPr2ktRiQ/Vyg+I4t7qvDLN2hclQuYDAFAaW9BQn9w7NFYmm1sdcn526DXm1SDMK3Qnwz/oEySTzW0DAN224u+pqrJaVQfLK0tRdCKTzb0shOgGAEinMyuSybqdhq69Vj60/3JXdvlITy0u+MFuJwFXDF171nX798iMqRIhbQAAEiGNqqo0VnTvBQDo6Eyvamys32Xo2uvl9gHX9db6fnBSVZXmyvFdt/9H0zQ+lxl7xUnY2wAAhCjuy2Rz6yv7SYTELCv+fiabWw4AEHLe7rreG6ZpbDJ0bW0pik51dKZfBQCglMoyY4+HnO8RQnCoksl2b2JyzQKZsaWzZ9UvLX9GRzs60y9RSmsdx/4MAKC3t+8dL1846vvB16qqrE4mZ3x6/sKlxRVDtTsJe2jrj0wmm1tZ/SO9fI5bbSv+jaFrbxm6Vgw5/1ZirJXJ7MGR9i4djev279MUZb1pGu821Cd/KTd3Xe7KLhNC5Hp6+rbrurZQZmyZbcW/BIBCKYoOS4RcU7+F0GTBvMK8wrxCU8VY82o8qvNqtP4h5925f3pWOQn7K9M0NpabB1zXe9t1+w9V51VlzbbM2BOV7xGJXC0bwrwaNh7mFbpj4dMOJxil1FBV5aEaGjMKfnCeh/yP6mCjlMqqqjwmEVLj5QsHhBB91xsv1dy0i9LY3L9Pn2ssRZG4Xr9KEiEx3dAfkQgxfT84HHKeuVF/Q9fulWXWGoY87fvB79XrlQipZTK7z/eDg5XtMmNxIpGZPOSnS+W6qOmOvdRJ2Nsvd2UXum7/TzeY835ZZqkw5Bd9Pzg2UviPpLUl1QMAVkdn2iGS1CYRonn5wn4hhHe9Y1RVSWmq0uZ5hWMh5xfHMs9oZMYaDEObNyCKnu8Hv1VfLA1de0CW2d2eVzgUcp6+FXMidKthXmFeAWBeoalhLHk1HjeTV5RSy9C1hwGA+H5wpLKm/GZgXl0L8wohdNszzdr5rS2p0nTHXjbZaxmLVHPToVRz0/6JGr+1JdXT2pK6oqrKNds8IYQmF+bVcJhXCN2+MK+Gw7xCaPLgNmtTjOv2H/H9YLNlxZ+f7LWMxjRr51Eaq8lku9dM9loQQv8/zCuE0FSBeYUQul3gLe5oSlNVJQ4A03jI+/7LrW0IITTRMK8QQlMF5hVCCCGEEEIIIYQQQgghhBBCCCGEEJpc/wKECL4s5TOl0QAAAABJRU5ErkJggg==\n", "text/html": [ "
r0-0r0-1r0-2r1-0r1-1r1-2r2-0r2-1r2-2r3-0r3-1r3-2r4-0r4-1r4-2r5-0r5-1r5-2r6-0r6-1r6-2r7-0r7-1r7-2-1001020PC0 (26.9%) explained-1001020PC1 (13.2%) explainedr0-0r0-1r0-2r1-0r1-1r1-2r2-0r2-1r2-2r3-0r3-1r3-2r4-0r4-1r4-2r5-0r5-1r5-2r6-0r6-1r6-2r7-0r7-1r7-2-1001020PC0 (26.9%) explained-1001020PC2 (12.4%) explainedr0-0r0-1r0-2r1-0r1-1r1-2r2-0r2-1r2-2r3-0r3-1r3-2r4-0r4-1r4-2r5-0r5-1r5-2r6-0r6-1r6-2r7-0r7-1r7-2-1001020PC1 (13.2%) explained-1001020PC2 (12.4%) explainedr0r1r2r3r4r5r6r7
" ], "text/plain": [ "" ] }, "execution_count": 93, "metadata": {}, "output_type": "execute_result" } ], "source": [ "tool.run(nreplicates=10)\n", "tool.draw_panels(imap=IMAP)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Imputing without a priori groups" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### What happens if you don't impute\n", "Even a small amount of missing values (e.g., 5%) will cause significant bias if you arbitrarily set it to the ancestral state (0) instead of imputing. This example uses the `impute_method=None` and you can see the difference from the results above. I do not recommend using this setting." ] }, { "cell_type": "code", "execution_count": 95, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Samples: 24\n", "Sites before filtering: 10033\n", "Filtered (indels): 0\n", "Filtered (bi-allel): 303\n", "Filtered (mincov): 0\n", "Filtered (minmap): 7\n", "Filtered (subsample invariant): 0\n", "Filtered (minor allele frequency): 3151\n", "Filtered (combined): 3458\n", "Sites after filtering: 6575\n", "Sites containing missing values: 5116 (77.81%)\n", "Missing values in SNP matrix: 9513 (6.03%)\n", "Imputation (null; sets to 0): 100.0%, 0.0%, 0.0%\n" ] } ], "source": [ "# init a tool with the missing data dataset\n", "tool3 = ipa.pca(\n", " data=\"/tmp/test-pca-cov75.snps.hdf5\",\n", " imap=IMAP, \n", " minmap={i: 1 for i in IMAP}, \n", " mincov=0.5,\n", " minmaf=0.03,\n", " impute_method=None, \n", ")" ] }, { "cell_type": "code", "execution_count": 19, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Subsampling SNPs: 6647/55885\n" ] }, { "data": { "text/html": [ "
r0-0r0-1r0-2r0-3r1-0r1-1r1-2r1-3r2-0r2-1r2-2r2-3r3-0r3-1r3-2r3-3r4-0r4-1r4-2r4-3r5-0r5-1r5-2r5-3r6-0r6-1r6-2r6-3r7-0r7-1r7-2r7-3-505101520UMAP component 10102030UMAP component 2r0r1r2r3r4r5r6r7
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "tool.run_umap(n_neighbors=6)\n", "tool.draw();" ] } ], "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": "3.7.9" } }, "nbformat": 4, "nbformat_minor": 4 }