diff --git a/examples/diagnostics_and_criticism/sampler-stats.ipynb b/examples/diagnostics_and_criticism/sampler-stats.ipynb index a9fb9f01a..db25574a2 100644 --- a/examples/diagnostics_and_criticism/sampler-stats.ipynb +++ b/examples/diagnostics_and_criticism/sampler-stats.ipynb @@ -21,11 +21,12 @@ "name": "stdout", "output_type": "stream", "text": [ - "Runing on PyMC3 v3.11.0\n" + "Runing on PyMC3 v3.11.2\n" ] } ], "source": [ + "import arviz as az\n", "import matplotlib.pyplot as plt\n", "import numpy as np\n", "import pandas as pd\n", @@ -37,6 +38,16 @@ "print(f\"Runing on PyMC3 v{pm.__version__}\")" ] }, + { + "cell_type": "code", + "execution_count": 2, + "metadata": {}, + "outputs": [], + "source": [ + "az.style.use(\"arviz-darkgrid\")\n", + "plt.rcParams[\"figure.constrained_layout.use\"] = False" + ] + }, { "cell_type": "markdown", "metadata": {}, @@ -46,7 +57,7 @@ }, { "cell_type": "code", - "execution_count": 2, + "execution_count": 3, "metadata": {}, "outputs": [], "source": [ @@ -57,15 +68,13 @@ }, { "cell_type": "code", - "execution_count": 3, + "execution_count": 4, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ - "/Users/CloudChaoszero/Documents/Projects-Dev/pymc3/pymc3/sampling.py:465: FutureWarning: In an upcoming release, pm.sample will return an `arviz.InferenceData` object instead of a `MultiTrace` by default. You can pass return_inferencedata=True or return_inferencedata=False to be safe and silence this warning.\n", - " warnings.warn(\n", "Multiprocess sampling (2 chains in 2 jobs)\n", "NUTS: [mu1]\n" ] @@ -88,7 +97,7 @@ " }\n", " \n", " \n", - " 100.00% [6000/6000 00:14<00:00 Sampling 2 chains, 0 divergences]\n", + " 100.00% [6000/6000 00:03<00:00 Sampling 2 chains, 0 divergences]\n", " \n", " " ], @@ -103,179 +112,567 @@ "name": "stderr", "output_type": "stream", "text": [ - "Sampling 2 chains for 1_000 tune and 2_000 draw iterations (2_000 + 4_000 draws total) took 36 seconds.\n" + "Sampling 2 chains for 1_000 tune and 2_000 draw iterations (2_000 + 4_000 draws total) took 11 seconds.\n" ] } ], "source": [ "with model:\n", " step = pm.NUTS()\n", - " trace = pm.sample(2000, tune=1000, init=None, step=step, cores=2)" + " trace = pm.sample(2000, tune=1000, init=None, step=step, cores=2, return_inferencedata=True)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "NUTS provides the following statistics:" + "- `Note`: NUTS provides the following statistics( these are internal statistics that the sampler uses, you don't need to do anything with them when using PyMC3, to learn more about them, [check this page](https://docs.pymc.io/api/inference.html#module-pymc3.step_methods.hmc.nuts)." ] }, { "cell_type": "code", - "execution_count": 4, + "execution_count": 5, "metadata": {}, "outputs": [ { "data": { + "text/html": [ + "
\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "
<xarray.Dataset>\n",
+       "Dimensions:             (chain: 2, draw: 2000)\n",
+       "Coordinates:\n",
+       "  * chain               (chain) int64 0 1\n",
+       "  * draw                (draw) int64 0 1 2 3 4 5 ... 1995 1996 1997 1998 1999\n",
+       "Data variables: (12/13)\n",
+       "    n_steps             (chain, draw) float64 3.0 3.0 3.0 3.0 ... 3.0 3.0 3.0\n",
+       "    perf_counter_start  (chain, draw) float64 7.234 7.235 7.235 ... 9.779 9.78\n",
+       "    perf_counter_diff   (chain, draw) float64 0.0002823 0.000284 ... 0.0002742\n",
+       "    energy_error        (chain, draw) float64 -0.5087 -0.4354 ... 0.04225 0.1197\n",
+       "    tree_depth          (chain, draw) int64 2 2 2 2 2 2 2 2 ... 2 2 2 3 3 2 2 2\n",
+       "    max_energy_error    (chain, draw) float64 -0.6245 0.5775 ... 0.3449 -0.1726\n",
+       "    ...                  ...\n",
+       "    energy              (chain, draw) float64 20.89 20.49 18.01 ... 17.55 16.76\n",
+       "    diverging           (chain, draw) bool False False False ... False False\n",
+       "    step_size_bar       (chain, draw) float64 0.9604 0.9604 ... 0.9419 0.9419\n",
+       "    lp                  (chain, draw) float64 -15.3 -13.86 ... -13.54 -13.95\n",
+       "    acceptance_rate     (chain, draw) float64 1.0 0.8879 0.927 ... 0.9095 0.9688\n",
+       "    step_size           (chain, draw) float64 0.9213 0.9213 ... 0.9988 0.9988\n",
+       "Attributes:\n",
+       "    created_at:                 2021-04-06T16:24:24.465349\n",
+       "    arviz_version:              0.11.2\n",
+       "    inference_library:          pymc3\n",
+       "    inference_library_version:  3.11.2\n",
+       "    sampling_time:              10.891047954559326\n",
+       "    tuning_steps:               1000
" + ], "text/plain": [ - "{'depth',\n", - " 'diverging',\n", - " 'energy',\n", - " 'energy_error',\n", - " 'max_energy_error',\n", - " 'mean_tree_accept',\n", - " 'model_logp',\n", - " 'perf_counter_diff',\n", - " 'perf_counter_start',\n", - " 'process_time_diff',\n", - " 'step_size',\n", - " 'step_size_bar',\n", - " 'tree_size',\n", - " 'tune'}" + "\n", + "Dimensions: (chain: 2, draw: 2000)\n", + "Coordinates:\n", + " * chain (chain) int64 0 1\n", + " * draw (draw) int64 0 1 2 3 4 5 ... 1995 1996 1997 1998 1999\n", + "Data variables: (12/13)\n", + " n_steps (chain, draw) float64 3.0 3.0 3.0 3.0 ... 3.0 3.0 3.0\n", + " perf_counter_start (chain, draw) float64 7.234 7.235 7.235 ... 9.779 9.78\n", + " perf_counter_diff (chain, draw) float64 0.0002823 0.000284 ... 0.0002742\n", + " energy_error (chain, draw) float64 -0.5087 -0.4354 ... 0.04225 0.1197\n", + " tree_depth (chain, draw) int64 2 2 2 2 2 2 2 2 ... 2 2 2 3 3 2 2 2\n", + " max_energy_error (chain, draw) float64 -0.6245 0.5775 ... 0.3449 -0.1726\n", + " ... ...\n", + " energy (chain, draw) float64 20.89 20.49 18.01 ... 17.55 16.76\n", + " diverging (chain, draw) bool False False False ... False False\n", + " step_size_bar (chain, draw) float64 0.9604 0.9604 ... 0.9419 0.9419\n", + " lp (chain, draw) float64 -15.3 -13.86 ... -13.54 -13.95\n", + " acceptance_rate (chain, draw) float64 1.0 0.8879 0.927 ... 0.9095 0.9688\n", + " step_size (chain, draw) float64 0.9213 0.9213 ... 0.9988 0.9988\n", + "Attributes:\n", + " created_at: 2021-04-06T16:24:24.465349\n", + " arviz_version: 0.11.2\n", + " inference_library: pymc3\n", + " inference_library_version: 3.11.2\n", + " sampling_time: 10.891047954559326\n", + " tuning_steps: 1000" ] }, - "execution_count": 4, + "execution_count": 5, "metadata": {}, "output_type": "execute_result" } ], "source": [ - "trace.stat_names" + "trace.sample_stats" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "- `mean_tree_accept`: The mean acceptance probability for the tree that generated this sample. The mean of these values across all samples but the burn-in should be approximately `target_accept` (the default for this is 0.8).\n", - "- `diverging`: Whether the trajectory for this sample diverged. If there are many diverging samples, this usually indicates that a region of the posterior has high curvature. Reparametrization can often help, but you can also try to increase `target_accept` to something like 0.9 or 0.95.\n", - "- `energy`: The energy at the point in phase-space where the sample was accepted. This can be used to identify posteriors with problematically long tails. See below for an example.\n", - "- `energy_error`: The difference in energy between the start and the end of the trajectory. For a perfect integrator this would always be zero.\n", - "- `max_energy_error`: The maximum difference in energy along the whole trajectory.\n", - "- `depth`: The depth of the tree that was used to generate this sample\n", - "- `tree_size`: The number of leafs of the sampling tree, when the sample was accepted. This is usually a bit less than $2 ^ \\text{depth}$. If the tree size is large, the sampler is using a lot of leapfrog steps to find the next sample. This can for example happen if there are strong correlations in the posterior, if the posterior has long tails, if there are regions of high curvature (\"funnels\"), or if the variance estimates in the mass matrix are inaccurate. Reparametrisation of the model or estimating the posterior variances from past samples might help.\n", - "- `tune`: This is `True`, if step size adaptation was turned on when this sample was generated.\n", - "- `step_size`: The step size used for this sample.\n", + "The sample statistics variables are defined as follows:\n", + "\n", + "- `process_time_diff`: The time it took to draw the sample, as defined by the python standard library time.process_time. This counts all the CPU time, including worker processes in BLAS and OpenMP.\n", + "\n", + "- `step_size`: The current integration step size.\n", + "\n", + "- `diverging`: (boolean) Indicates the presence of leapfrog transitions with large energy deviation from starting and subsequent termination of the trajectory. “large” is defined as `max_energy_error` going over a threshold.\n", + "\n", + "- `lp`: The joint log posterior density for the model (up to an additive constant).\n", + "\n", + "- `energy`: The value of the Hamiltonian energy for the accepted proposal (up to an additive constant).\n", + "\n", + "- `energy_error`: The difference in the Hamiltonian energy between the initial point and the accepted proposal.\n", + "\n", + "- `perf_counter_diff`: The time it took to draw the sample, as defined by the python standard library time.perf_counter (wall time).\n", + "\n", + "- `perf_counter_start`: The value of time.perf_counter at the beginning of the computation of the draw.\n", + "\n", + "- `n_steps`: The number of leapfrog steps computed. It is related to `tree_depth` with `n_steps <= 2^tree_dept`.\n", + "\n", + "- `max_energy_error`: The maximum absolute difference in Hamiltonian energy between the initial point and all possible samples in the proposed tree.\n", + "\n", + "- `acceptance_rate`: The average acceptance probabilities of all possible samples in the proposed tree.\n", + "\n", "- `step_size_bar`: The current best known step-size. After the tuning samples, the step size is set to this value. This should converge during tuning.\n", - "- `model_logp`: The model log-likelihood for this sample." + "\n", + "- `tree_depth`: The number of tree doublings in the balanced binary tree." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "If the name of the statistic does not clash with the name of one of the variables, we can use indexing to get the values. The values for the chains will be concatenated.\n", - "\n", - "We can see that the step sizes converged after the 1000 tuning samples for both chains to about the same value. The first 2000 values are from chain 1, the second 2000 from chain 2." + "Some points to `Note`:\n", + "- Some of the sample statistics used by NUTS are renamed when converting to `InferenceData` to follow [ArviZ's naming convention](https://arviz-devs.github.io/arviz/schema/schema.html#sample-stats), while some are specific to PyMC3 and keep their internal PyMC3 name in the resulting InferenceData object.\n", + "- `InferenceData` also stores additional info like the date, versions used, sampling time and tuning steps as attributes." ] }, { "cell_type": "code", - "execution_count": 5, + "execution_count": 6, "metadata": {}, "outputs": [ { "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAAArAAAAEiCAYAAAABAuFjAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAA9hAAAPYQGoP6dpAABUR0lEQVR4nO3de3wU1d0/8M/sPZvrBhIC4aoQQG4aICCXogiCYKmiWJ4i4r1quTzUyg+sL+/PEysPtQhWq21FIbWQVqCiVEVaICCEaygCRrlESIAk5J6978zvj7Ozm80Fkk1CstnP+/VqJbOzM2fOOfOd78ycmZUURVFARERERBQiNG1dACIiIiKipmACS0REREQhhQksEREREYUUJrBEREREFFKYwBIRERFRSGECS0REREQhhQksEREREYUUJrBEREREFFKYwBIRERFRSGECSyFv6dKl6N+/P/bt29ch1nOtfPzxx7j33ntx0003IS0tDY899hgOHTrU1sUiIi/Gtqa5cOECMjIysHTpUtxxxx0YMGBAh9guqp+urQtARNfe//zP/+DDDz+EyWTC2LFj4XA4sGfPHuzevRtvvvkmJk2a1NZFJCJqks8//xzp6eltXQy6RpjAEjXSL3/5Szz22GPo1q1bWxelWfbs2YMPP/wQcXFxWL9+PXr37g0AOHz4MObOnYtly5YhLS0NMTExbVtQIromOkps69GjB+bNm4chQ4ZgyJAheOWVV5CVldXWxaJWwgSWqJESExORmJjY1sVotvfffx8A8OSTT/qSVwC46aabMHv2bKxduxZ/+9vf8PDDD7dRCYnoWuoose22227Dbbfd1tbFoGuEY2Cp3bpw4QJeffVVTJkyBUOHDkVaWhpmzpyJ1atXo6qqqt7v7N+/Hw888ABuuukmpKam4vHHH8f3339fZ76KigqsXbsWjzzyCG699VYMHjwYo0aNwiOPPILdu3fXu+yGxolNnDgR/fv3BwBkZmbixz/+MYYOHYqxY8fi+eefR0VFRTNrouXY7Xbs3bsXADB16tQ6n6vT/vWvf13TchGFE8Y2ouZjAkvt0oEDBzBjxgysXbsWLpcLt956K1JTU1FVVYVVq1bh3Llzdb7zr3/9C/PmzYPdbseECROQkJCAHTt2YM6cOSgqKgqY98iRI3j11Vdx5swZ9OnTB5MnT0afPn2we/duPPLII/jb3/7W5DK//vrreOmll5CYmIgf/ehHUBQF69evx5NPPglFUYKui5Z05swZOJ1OxMfHIykpqc7nN9xwAwDg22+/vdZFIwoLjG1ELYNDCKjdKSsrw4IFC1BRUYElS5bgoYcegkbjP9c6fPhwvbe7PvjgA6xatcr3AJLH48HixYvx+eef4y9/+QsWLVrkm7dPnz5Yv349brzxxoBlHD9+HPPmzUN6ejruuOMOREZGNrrc//jHP/CPf/wD1113HQCgpKQEs2fPxoEDB7B3717cfPPNjVrO3LlzkZ2d3ej1AkB6ejpmzpx51fkKCgoAoN7kFQDMZjNiYmJQXl6OqqoqREVFNakcRNQwxrbWi20UfpjAUruTmZmJkpISjB8/Ho888kidz2+66aZ6vzd9+vSAp+e1Wi0ef/xxfP755zhw4EDAvD169ECPHj3qLOOGG27Az372M7zzzjvYt28fJk6c2OhyL1q0yBfgASA+Ph6zZ8/Gb37zGxw4cKDRQX78+PFITk5u9HoBoGfPno2az2q1AgBMJlOD80RERKCiogLV1dVMYIlaEGNb68U2Cj9MYKnd+frrrwEAs2fPbtL3xo0bV2ea+pBSYWFhnc88Hg++/vprHD58GEVFRXA6nQCAs2fPAgDy8vKatP6xY8c2uP7at/mu5PHHH2/SeokoNDC2MbZRy2ECS+3OhQsXAKDeqwhX0qVLlzrT1CuILpcrYPrFixfx85//HCdPnmxwedXV1U1af3235dXbdOoBpK2ZzWYA4mGuhthsNgBo0i1GIro6xjailsMEljqMmmPJrubXv/41Tp48iSlTpuDRRx9Fnz59EBkZCY1Gg/Xr1+P5559v8sMJTVn/lbz77rs4ffp0k75z7733YsSIEVedT33P48WLF+v93Gq1oqKiArGxsRw+QNROMLZdPbZR+GECS+1O165dcfr0aZw7d873CpeWZLVasWfPHnTu3BlvvPEGtFptwOf1PQV8Le3atavJDzqkpaU1Ksj36dMHBoMBJSUluHTpUp0rO8ePHweAVql3onDH2NZ6sY3CDxNYanduvvlm7N69G+vXr2+VnzStrKyELMtISEioE+BdLhe+/PLLFl9nU6xdu7bVlm0ymTB69Gjs3LkTW7duxYMPPhjw+T//+U8AwK233tpqZSAKV4xtrRfbKPzwPbDU7syaNQsWiwU7d+7EmjVr6tzuOnLkCC5fvhz08jt16oTo6Gh89913OHjwoG+6x+PB//3f//kedOioHnroIQDA22+/HbCthw8fxvr16xETE4N77723jUpH1HExthG1HF6BpXYnLi4OK1euxJNPPon09HSsXbsWQ4YMgd1ux+nTp5GXl4dNmzahU6dOQS1fp9Ph0UcfxRtvvIG5c+di9OjRiI2NRU5ODi5fvow5c+YgIyOjhbeq/RgzZgweeOABfPjhh7jrrrswZswYuFwu7NmzB4qiID09HTExMW1dTKIOh7GtdRUWFmL+/Pm+v9Xxti+99JJvTP+ECRPwi1/8ok3KRy2LCSy1S6NGjcLmzZvxxz/+Ebt27cK2bdsQGRmJ5ORkLFy4sMlP8db2xBNPICkpCR988AEOHToEo9GI4cOHY+HChb5xoB3Zr3/9awwcOBDr1q3Dnj17oNfrcfPNN+Opp55CampqWxePqMNibGs9TqcTOTk5daafOnXK9++a77Ol0CYp/B04IiIiIgohHANLRERERCGFCSwRERERhRQmsEREREQUUpjAEhEREVFIYQJLRERERCGFCSwRERERhRQmsEREREQUUvhDBo1QWlrapPljY2NRXl7eSqWhUMA+QE3tAxaLpRVLU7+mxjaAfTvcsf0pmD7QGvGNV2BbgUbDag137APUUftAR90uahy2P7WXPtA+SkFERERE1EjtdgiBw+HAb3/7Wxw7dgx5eXkoLy9HTEwMevTogVmzZmHGjBnQ6/WNWpYsy8jIyMCGDRuQl5cHs9mMMWPGYPHixc3+3WkiIiIiurba7RXY6upqfPTRR5AkCbfccgseeughTJo0CYWFhXj22WfxxBNPQJblRi3r+eefx6uvvgpFUTB37lyMHz8eX3zxBe69916cPXu2dTeEiIiIiFpUu70CGxcXhwMHDsBgMARMd7vdeOihh5CVlYWdO3filltuueJy9u7di8zMTIwcORJ//vOffcu788478fjjj+OVV17Bn/70p9baDCIiIiJqYe02gdVoNHWSVwDQ6XSYPHkysrOzkZeXd9XlZGZmAgAWLVoUsLwJEyYgLS0NWVlZKCgoQLdu3Vqu8F6XL8u4cBEwGoCqaiAuFujcGbDZAKdTzON0AnYHEBMNJCRIcDoVlJQC8RYgOlpcIHc4FNhsQEQEYDRKvr+1WvFfQHzm8UiIiBB/5+crKC4GzJGAyQhYLIBWKz53OhVcvARoNYDBIMpQWSW+1yleLKukFPB4AIcDcLkAsxno0R0wGCQUFSm4XALo9WJadLQGlZUyzp0X83aKB2JjRXlql7GoCCgvB3r2BDp1Et8rKQWiIv3lMxolXx06HArKyxVUVYky6vVAl0RRtqoqQJaBLl1EucrLFd961HXGxkq+5anLami60+lvD5NR1I3BINZ17ryor8goQKcV6wSA02dEuZK7irZV28DhUFBYKMPpFMuIjRXrKipSYLcDkZGA0+Xfbq1W8X23pERB3g9Ar55AfLwUUDZ1e6OjNQH9oGbbq+3jcon6MhpFmdV1xlv89VVV5e9/sbGibez2wDpV+6rBIOq2vByoqPTXUVU1cPEC4HYDiYlAcnJgW1ZWyrh4SSzXZgVi4wBLnL+dai+/ZtsB8PX9mv1aXUbteYxGybffdU0Sfaymmm1dc/8zGYGoqMAy1azvq9WFWn9qn/d4JJjNypXCQ8iqqJCR94Ps60c2G+DxKKiq9vet2v3coBf/hQLkfgeUXAYSEoBOnUX/VOOMyyXaI8II6PT+v2OiAIMRqCgHTCbA7QGKCkV/M5pEuYxG0V4mI1BtBQoKgE6dgD69RZtevCSWldwViIkR/ba6CvjhPGAyAA4noNMBGi3QyQL06SP6wKnTYt+3WALjpRr/ABEfrFYRJ9X4VFIqyhsbB+h1/hir0YjpsgxIWlEejQY4mQtoJBFHoPhjpMOhID9f8cVNAPjhnFi/RuPvu+XlwDcnxL7eNQmQNCLe6nSARwaizGKbtFqxDK1W1AEUUbauSUBUlD8ulJSK7/W9TpTj8mUZp04DMbF2GHSyry7O5wMXC4HYaBEfEzqL7ZVlQFFEvVviRN3YHaJubFbR/t26Sb7jkXq8sVpFG8THAfkFQGUF0LWbvx1rxh5VYRHwQx4QYRZlSO4q+snpM0BFhYhNOq3oD0Dd+FEzFmo08P07obPYBrUd1X/rDYDL6W9bWRZ9vapa9BXA397RUf44WVEujsl2u/g8Jsrfr9SY4vFI8HgUXCoUdQGItlaXA/jjlscDuF1AUpL4t3pctsSJOql5bFLjqs0m+nREhOirQGBbaSTRDhaL2LdMRrHvXi4W8xlMwPCbPNC1g+yxHRShaWRZxq5duwAAKSkpV51/3759MJvNSE1NrfPZ+PHjkZ2djezsbNx1110tWs4du2T87WPgwgURWAx6wOTtMBERQFkZYLMDdm+nssQD1/dR4HKLnS0mGpg6VUZCZwk5OeLgEBUJdEtWUJAPFFwAfvjBf1AwGIGePRVERgDfnwIOHRE7riyLA2vnTsDgwQrMJuC70yLhqKoWO6vDLjqsRisCodEbzCsrxY4iSWLnvK4PkNBZwYlvgZISccDo3w8YPVrG13vFgcnhEOvr1UOUteSyv4wlZUBxESArojzjxsq4eFHsaHYHkNJPQUpfYNgwIKmLhIuXFOzcqWBvNnDmrCin3iDqJiLCH8iSugJJiQoqq0R9ehQRRCJMQL9+Cib8SNTxzp0Kvv0OkFB3+pGjwPnzInB5PKIuYqLFNhYWAlabaBe9XgSDhASgqlIEWKcTiDQD118PDB0q2sDlrsY3x4HSUiA+HujTWwTIk7kiILvdIkDoDUByNwVaDdA5QcHFAuDIUVH/BgPQP0WBxyPKZrOJdujTB7hlggy7TfSD4mLx3cgI8ffBw6LMbrfYPoNBHDzUxFucYCj44Zw4EMoeICpa9C+1vTt1FnVaWARcuiSmxcaKA1ppiQiMiiKmV1aI8iqKqK+uXYG0kaItTREK/v1v4PgJEVzV8nTvLurF5RT7grp8vcHfdklJCqKjvQdgt79fV1WJfnvddcCNQ/3zREUCdoeCbdvFPFFRwL0zZUwYL44oan86clRsk7r/eTxivUldRD24nOLAr9a32r8aqgurN3lO7ir6vN0m2qPv9Q7066sgqYv/hCzU/ecbGdv/ZUVxsehPycniJCr3e28iFQno9Qp+OO/v55GRYv/WakT/dDj8y9PrRdvJsvifx+PvV5Ik/g2I9pa81ajOo35mMIhlSJJYnscjEgTFGwdiYwFIIg4DIrGJjxcH/kuFor1rnmpoNKI9k7qI/aG0TCTSJpPYFofd+7d3HkDEAatV9H9LnChPYZGYDxDJoiKLxNvj8W+vTifK7/bGfY9HrD86GuieDEy7Q8bZM8DX2f4TUkkS/Vtdn7pvXyr0r0/djpqj7PR6sdxePcXxqKpazF98GTBHiGSsV09xgvj9KcBaLU4iunUFRoyQsf+AOJ653dUwGkVdVJR7k+Aa6zQaRCKpkfxJHCCmAyJWaLXiJGLgAHHMO/eDSJgdDlFmtexqDFOTMlOEv20jIkTyWFYOlHmTbbWuzWbxb5vNvwzA30dMJn/8AER8+uGciAd2h7/eJMnf92TZ/1+3W7SdViuWFWHyT3O6RL/xuEW/U/unzQq43P72V8saESGOh507izoxGhV8mytiuJp0qvNFRYm/7TZRD06X6AN674mXw+Hfb9QT8EiziLcWizjp++G8/zOTSSxXkUVbeTxiOyRJ7DumCLFup1P0FVkW6+vTuwJzfuaPrW1FUhRFufpsbcfpdOIPf/gDFEVBWVkZvv76a5w+fRozZ85Eenr6Fb9rtVpx0003ISUlBZ988kmdzz///HMsXLgQTz31FBYtWtTgcpr6rkS3OwaLnynD5csi0bNaRScw6EUAs1hER6ioEB3NZBKd3OUUZ3yDhwKXL4uzswH9AadbJB2XCoH880CXJODSRZEM6b2nIC430O96oLQc+OaYWI/De4am04nAldxVBI+ycjGtvNy/g6g7quwRZ+4ajb/D6rRifp1eJH+St89qNIBBB+i8Z+I1g3BUFNAtCbDaReByOv1XDeLixHpl79k9AFwuBSyxwPDhQGICMOZmYOcu4Ot93is2Jf4Dm7pNkZGi7pxOQK8FUvqLg8blyyIgqBfVhw4W/z16TOyUiiS2U51+OAe4UACcLxBBQaMR69LrRL26XGK7FYjp0dEiONkd8J2FejziIHDzaFHHJSUSZFmB2yPa3uUSB52ICJHUOhwi6JkjAa0EJHUTZ9jZ+0UdJiSIZEmWxYGsqNh/MmI0inmHp4plFZeI4F5SCpz6XiT2Nm9ipratTudPvrVacRBWAFRX++vU7QbiYoBOCSIp1kpATJzoJ2rdV1X5l2m1Bh4cVEYj0D8FuOlGkYyXlYnAafcmLlqtv/2MBv+BR6MR5UjoLBLtS5fEf0eOAHZ/DfznqJjX7RZ1oSYZycmi3gsuANu+EsvvkgAUXRYHhBefE1eWtn2lIPuA2HeKLovtkmUxv1Yr+rlBL/p5VbXY5pgY0XY6Tf11IUniYGmziat43bqJK0CJicB110UgwmjDLROkgLsKDWnv74GtrJTxzrtAVbUOsTFufHNC7BcRJqCiSvTByyUimbBYRJyx20Uio9H675i0BrV2lVrT1L8lSbRxzfWr/f5KtFqx37hc/uRSo/HvT2oyriZctZevKP59pGZCfjVqshlhErHKo4hpFy+KZahXP9V5a54UXIkkidgcFeW9y2AH3N4kR/aWNcLkP9nUe2O+rHjjoiySpqvVW83trbnd6kmITitOGtUrl263/3h4JQaD/4q9RhLHIvUY1hR6vVhvYgIASVyJL68QMa2+PtpQ26knWuqdOq3WH0vVN035kmGIeqxNoxExJDZaXN10uURct9v9V0bV47P6v4AkWCOO95Ik2s5m958E6PXeCy86cWHhwgURh2tuj3pyVV/Z6qP2oT69gFdeqnuXqyGtEd/a/RVYl8uF1atX+/6WJAkPP/wwnn766at+t7KyEgAQpZ621KJOV+drSGxsbJPee3Ykxwm7TYO4OBmXL3tvbXmv8ni8ncug95/ZmYyio9ntgMkkISbKgCgzcD7fA6tVh+uv13kPgDJOnXLCEqfDxYtuxMTIkD0KJEiIiFDg9mihyB64PTIMev8OJctiB7PbAVmRoNMp0OtEuex2sWNptYBWBzhkf7nUTdbpxGfqDmP07vweDyC7RRJkMvmvglRUiO/a7BKiIiVvMiyLbTUBMdEiEpaUAtHRWjgcErp2VVBdBcRbDJAVCbKsh93hhE7nAhS3L5B6ZACK2OHiYiUoANxuBR4ZiDTrYDIqMBg8MJm06NxJC5tdQbVV3DOLjJRhiRMbVVrm8U03Gj3QGzzQ6WTUfLGFJIn1aLX+M2/1yo+siHKoZ9fV1eJgabVpYTJJcHs8MJu0iIoWt9DtdhmyDMRES6isUKAxieVZLBKKi4EuCToUFXvg8cgwRwARRnELurLS23f03rN9owYREQqsVglRkQZUVMro1UNCeYUMjSTD5fZA6w1cgCi/esIRYQIUDxARKaGqSlz1jYgQ22G3i4Ok0QTERGtRVuaBywnoNKINPR7ZdztLp/Xf3q19dUM9QDkcGhgNelirnTDoFd+VCLfbH+zVec0RItVweE8ITCYtYmM1KCoW7WjQ6yHBAbfHA5Mx8Iqd2y3BZNIhOsoAs1mGw2FDn14axMbpoDN4cLlYgc0WiU6ddLA77DAa3TCaZJiMHlR6DwZajbiD4XaJupJlEfB1esCgF7fz3J7660KvE8FcPWA5nBr0TRIdJiZKgsNphslkRFxc+3xetimxrazMDbvDim5dNZA0OsTEuFBe5oHDqUHXrhKcDsBk9MDj8d5OrRKxzW4XyVhFRStuiLfvSahxYK/RT6D4TzbVq5RXSsK0msArepIktgMQiZPRIPqH0+5PLk2mGieEbsAY4T+hVxR/4tHIzYFeL+pQbwASOktwuhRoJLE+jXf4l8vlj9ONWabkPXmHIrbBahV3mmQFiNCLk021L+u0/v3CYRcn6BqdWJB6FfRq1LLV3Hb1WGPQi6Eeatl8iRhqnYjUTIAV8b3aJwtNpfcuw+USJ1fw3j2qmQzXvgNQs+1qJudqHNPrRbzV6QJPdtQTbo1WxN+a9aaeeGk04g+TUbS5tsYxWL3yW/NqsE7nPbGusS71Dph6oUDNLaze5NTtEfWqlktRvN+VgHquQ9RLPRnVaIAqqwY2WyQslrpDPa+Vdp/ARkZG4ttvv4UsyygsLMT27dvxxhtv4MiRI3jvvfcaTE5bUlN/cSK5WwxMESJ5lbzjSbRa760qb/BxuvwJkd0hOphWA9jtCiqqHL4rsGazG3nn/FdgTUagtMwFCeKAUPMKbJdEGZI38Dhd/uCp04mzbVMnwOFUfEmHw+FNyCB2LDUpq3kLp+ZtHPUKrMfjT1IN3itpNa/Aqt+LMCmoqlZgNPgTP7sdqKiUYbOLna2yUkQF9QpsSakNiQn+sV1u720Yl6vGFVhvkCgrV2DyJoJ6LVBtdcPu8I4PsntQfFksO9IsjljV1eLWi3oFVp3ucIi2cbv9BwT1Cqyk8SY28F+BlSSxI0MKHCspATBHeMRYOq0Eu8ODKqv/Cp9GA1RUKr6r4xEmoLRUgUEHXCpyITpKzGe1ATaHAqvVuy7vFVyHA9DrZSgAoqMUVFXb4XEDeYXi6pfsLbPN5r9iBACSrJ5QiKuDNrs4gVEgbhOqdapI4kBVUenxjRlzy6INFcWfCLs9Yv6atyuBwNu6RqMMh9MBc6Q4KEIJHPet0/nntdrEPzQa75Vtuwfl5R7IHsBul+F0uaBAlMfhDLwCq9MpsNtdqKxy+W6pXiqSATh9V2AjIirhcEgwGUWS7LD7bxPKMqBoAcUhlq/IgFYv9ienSyxfvQtRX1243OKA43KJK9ZGg4wLF2UkJgIVVVpEGG2w220oLW2fV2CbEtu0WhkmI1BwQVyBVWOA0SDjwgXRB+3eWFddLfqJOqaw2lr3lnaLUvyxTOVx1/hb8vcb//Y0nFB6aszn8QTGH1kR/VBWxD6jkQLnUfuH0xm4TzRl2xWI5UVFiVh1uUQRSVeNZan7U2OXq3j/T6MV9eFwijsLFZXiCqzNLhJL9QTRKYvvqMm0evfD6Wr8leSat91VHm+S6tT4x+Mq8Nd57UXXXJciifW7XP4rsMFwucQJgF4Pb2bvPx7Wt97adax+VnMel8vbp9zeixxyYKIte+rWm1q/6kmW3SFimNXqX0bNdanHZvWkyOXyn2xJtbbBl1uoJ+Ra/7oC+qVS96ShIbLizRVkIMosIyKiEqWlvAJ7VRqNBklJSfjZz34Gi8WC//7v/8bbb7+NZ555psHvREdHAwCq1PuetajT1flaSkKCFvfOBP72sXcsjKf+MbAmU+0xsOKAWOZN5qZOhW8MbFmZeLBgyGCgIF/sJD2tgWNguySJcaqx0WKsoHp7VB0DO/AGBIyB9XjEAb5pY2DhGwNrMKhjYOEbA6soYnyZGAOLgDGwRpMYA+t0inoYNxa+MbAxMcD1fcUtnWHDJERHSxg+XEF1tUjkXK6Gx8D27g0kJYoxPJ3jvQ+saUSA69cPGD1aJA9Op3cMrCJucdecrj4s0pgxsBZL3TGwMdFiDGz3HvCOgTXgm+OOGmNg4RsD63aJ5NU/BhbeMbAiET9yVNR9XJwop3q7yGYT2yvGwPr7jscjhgb06iFunV9pDGznTv6H7BoaAyt7RL0lJYq2UeT6x8AajQ2PgR04EOjZA0hJAf79b7HNTRkDq9cBI4aLK3cOB3DTUCA60j8GVq8X7X7jUDFPdbUYIvPTWcC27eKEKCYGuHem/xbX8OFAdbVoa5fbe7XiKmNgI0yivq9UF1abeDhEjIEVy7RYgE7xGvTr27jhA6EgOlqDqVNlbP+XFsXFbvTsLoZvqGNg9XoxJEivF+PsPB5Rf+oY2E7xbT8GVh2z2NQxsIoiYmPtMbB9eovvqGNgzea6Y2ANhqaPgTWbxUNV0+5AwBjYrl39w1fU/a0lxsCaTOoYWDF8yOUWJ7darToGFr4xsHr466L5Y2DhGwMry40fA6uO3WzOGFg1fgBiDKzHe6La9mNg0awxsDGGwDGwsTH+MbARppYZA9ujuwY//anc6OEDraXdj4GtT2VlJUaMGIGhQ4f63jLQkHHjxqG6uhoHDhyAVj3d83r33XexYsUK/OY3v7niQ1xNHQNrsVhQWlqKy5dlXLokdhirFYjxJpJ2u9iJAdExHE5xtbVzZ/FEZlmZSF5qvoVA3WnVtxCIsZrivwq8Y6VkCSbvE7kFBeJJdLWDxsUCOp3kHTOqoLDQO0bGO5aqqlokdvHx4jtlZd5xtE7A7RQBITlZPGFcXKzgcqm4+pqc7H8Lwfl8cVCIjxc7oixLdcpYXCwSz+7J/rcQlJV5g7C3fLXfQlBRIR7QqqoW60xIEDuS+iBRYoIoV0WF4luPus7YmMC3Dajz1Dfd4fS3h8kg6sZgEIEoPx+QtCLJ0+m8Y6cgHi6rrhIP+nTu5G8Di8WCvLwS38NYsTFiXcXFCuwOcTvO7fZvt0aj+L5bUqLg3DmgRw/xFoKaZVO3V30LgdoPara92j5up+h7Bu9td3WdcXH++qqs8ve/mBhxZd3uCKxTta8aDN7xjhVizKNaR1XVQOElEUwTE0UiV7MtKytlEYzt/tvJajCtuS+oy6/ZdoD/gFWzX6vLqD2P+hYCdfxsfW8hqLlNNds7MiqwTDXr+2p1odaf2udlWULXrhZYrWVNihvXWlNjGwBotbE4e7bU14/sdjGMx2r1963a/Vyn818B/f57MW47obP4n8HojzPqnQaDSezr6t+R0SIBqqwU83vcIvHq3EkcZCXvSbysiLassoqT404WoFcv0aYXC8Ut/6Su4oTTahXLO5/vvR3vFlcktVogLl6M8XM6gTNnxL4fHxcYL9X4B4hl2LwJrBqfLpeKk9zoaO9wAG+MlTRivWJYiiiPRgJyc8W6LfEiNqsx0uFQUFCg+OImIB7q1HqHdql9t7wMOHFSJDJdvcssvizWIctARKTYJp337otW4x/zWV0t9peoKH9cKC0TV9v69PG/heDMGSA6JhIGfbWvLvILgOJCwBwl9oP4Tt7b6Yo4GS4sFP0iwizaWdxlESd93bpKvuOReryxVfsvvFy4AJRXipiitmPN2KMqLBJ1YjSKMiR1FSf7Z88CZZX+txB0jhf1Xzt+1IyFGkm0ldMutkWBvx3Vf6tX79W2VRTR163eOw0K/O0dFeWPk1WVIk45nEB1pejXar9SY4osS3C7xUOjthpvIVCXA/jjlsebECfW6HMGnai7ouLAY5MaV21W0acjIkS8Vu8sqm0lSeK5m7gYMQzGZAAcLnGy7vaIN4TceGMcdLqmjQlqjfgWkgnsqVOnMG3aNKSmpuKjjz664ry//OUv8emnn2LdunUYOXJkwGdz585FdnY2tm/fjmQ1EtUj2ASWwhf7ADW1D4RKAsu+Hd7Y/hRMH2iN+NY+nywA8P3338NWz+OFNpvN9/aBCRMm+KaXlJTg1KlTKCkpCZj/vvvuAwCsXLkSTnXQEIAdO3YgOzsb48aNu2LySkRERETtS7sdA7t161a8//77GD58OJKTkxEVFYVLly5h586dKCsrw4gRI/Dggw/65s/IyMDq1asxf/58LFiwwDd99OjRmDVrFjIzMzFz5kxMmDABRUVF+OyzzxAXF4fnnnuuDbaOiIiIiILVbhPYW265BYWFhTh8+DCOHDkCq9WKqKgo9O/fH9OnT8c999wDXSN/CuLll19GSkoKNmzYgA8//BBmsxmTJ0/G4sWL0VP9aRMiIiIiCgkhOQb2WuMYWGoq9gHiGFjqiNj+xDGwRERERERBYAJLRERERCGFCSwRERERhRQmsEREREQUUpjAEhEREVFIYQJLRERERCGFCSwRERERhRQmsEREREQUUpjAEhEREVFIYQJLRERERCGFCSwRERERhRQmsEREREQUUpjAEhEREVFIYQJLRERERCGFCSwRERERhRQmsEREREQUUpjAEhEREVFIYQJLRERERCGFCSwRERERhRQmsEREREQUUpjAEhEREVFIYQJLRERERCGFCSwRERERhRQmsEREREQUUpjAEhEREVFIYQJLRERERCGFCSwRERERhRRdc75ss9lw7NgxFBUVwel0NjjfXXfd1ZzVEBERERH5BJ3Arly5EmvWrIHdbm9wHkVRIEkSE1giIiIiajFBJbDvvfce3n77bWi1WkyYMAF9+vRBZGRkixbs0qVL2Lp1K3bu3InTp0+juLgYsbGxSE1NxaOPPophw4Y1ajn79u3DAw880ODn6enpmDlzZksVm4iIiIhaWVAJbGZmJkwmEzIyMjBo0KCWLhMAYO3atXjvvffQs2dPjB07FvHx8cjLy8O2bduwbds2rFixAtOmTWv08tLS0pCWllZn+sCBA1uy2ERERETUyoJKYC9cuIDRo0e3WvIKAEOHDsXatWvrJJ0HDhzAgw8+iBdffBGTJk2CwWBo1PLS0tKwYMGC1igqEREREV1DQb2FICEhARERES1dlgC33357vVdMR4wYgVGjRqG8vBzffvttq5aBiIiIiNqfoK7ATps2DX/7299gtVphNptbukxXpdPpAv7bGGfPnsWaNWvgcDjQpUsX3HzzzejSpUtrFZGIiIiIWklQCeyCBQtw+PBhPPnkk3j55ZfRq1evli5XgwoKCrBnzx4kJCQgJSWl0d/bsmULtmzZ4vtbp9Ph/vvvx5IlS6DValujqERERETUChqVwNb3FL8sy8jOzsa0adPQrVs3JCUlQZKkOvNJkoQPPvig+SUF4HK5sGTJEjidTvzqV79qVOIZHx+Pp59+GrfeeiuSk5Nhs9lw+PBhrFixAmvWrIEkSVi6dGmLlI+IiIiIWp+kKIpytZkGDBgQ/AokCSdOnAj6+ypZlvHMM89gy5YtuO+++/DKK680a3lFRUWYMWMGKioqsHPnTnTq1OmK69Zo+KNlRNSxMLYRUahq1BXYr776qrXLcUWyLOPZZ5/Fli1bMGPGDLz00kvNXmZCQgJuu+02ZGZmIicnBxMnTmxw3vLy8iYt22KxoLS0tLlFpBDGPkBN7QMWi6UVS1O/psY2gH073LH9KZg+0BrxrVEJbHJycouvuLFkWcayZcuwadMm3HnnnXjttdda7IqBWqE2m61FlkdERERErS+oTHD16tWNuiq7fft2rF69OphVAAhMXqdNm4bXX3+9RR+4ysnJAdC2CToRERERNU3QCey2bduuOt/27dvx1ltvBbMK37CBTZs2YerUqVi+fPkVk9eSkhKcOnUKJSUlAdOPHTtW7/wffPAB9u3bh969e2PIkCFBlZGIiIiIrr2gXqPVWB6PJ+jb/W+99RY2btwIs9mM3r174+23364zz6RJk3w/BZuRkYHVq1dj/vz5Ab+4tXDhQuh0OgwePBhdunSBzWZDTk4Ojh8/jpiYmKsmxkRERETUvrRqAvv9998jJiYmqO/m5+cDAKxWK955551650lOTvYlsA2ZPXs2srKysH//fpSVlUGj0aBbt26YN28eHn74YSQlJQVVPiIiIiJqG416jRYALFu2zPfvjRs3olevXkhNTa13Xo/HgzNnzuDYsWOYNGkSVq1a1TKlbSPBPG3HpzTDG/sAhcJbCILpo+zb4Y3tTyH1FgJAJK0qSZKQl5eHvLy8K36nf//+WLJkSfClIyIiIiKqpdEJ7IcffggAUBQF8+bNw/jx4/HYY4/VO69er0diYiKf7iciIiKiFtfoBDYtLc3377vvvhvDhw8PmEZEREREdC0E9RBXenp6S5eDiIiIiKhRmvUWAqfTic8//xwHDhxAYWEhACAxMRHDhw/HlClTYDQaW6SQRERERESqoBPYPXv2YOnSpSgqKkLtFxls2LABy5cvx2uvvYaxY8c2u5BERERERKqgEticnBz8/Oc/h8vlwrBhwzB9+nTfA1sFBQX49NNPceTIETzxxBNYt24dhg0b1qKFJiIiIqLwFVQCu3LlSrjdbrz44ouYPXt2nc/nzp2L9evX44UXXsCbb76JP/3pT80uKBERERERAAT1O685OTkYPHhwvcmr6qc//SmGDBmCI0eOBFs2IiIiIqI6gkpgNRoNevbsedX5evbsCUmSglkFEREREVG9gkpghw4ditzc3KvOl5ubi6FDhwazCiIiIiKiegWVwC5atAhnz57Fm2++CVmW63yuKArefPNNnD17FosWLWp2IYmIiIiIVEE9xHX69GncfffdePvtt7F582bcfvvtvrcQ5Ofn48svv0R+fj5mzZqFM2fO4MyZMwHfv+uuu5pdcCIiIiIKT5JS+yWujTBgwABIkhTw/ld1rGt901SKokCSJJw4cSLY8raJ0tLSJs1vsVia/B3qWNgHqKl9wGKxtGJp6hdMH2XfDm9sfwqmD7RGfAvqCuwvfvELPpxFRERERG0iqAR2wYIFLV0OIiIiIqJGCeohLiIiIiKithLUFdiaTp48iaNHj6K0tBR9+/bFbbfdBgBwOp1wOp2IiopqdiGJiIiIiFRBX4E9ffo0Zs+ejbvvvhsvvPACfve732Hbtm2+zz/55BOMHDkSO3fubJGCEhEREREBQSawFy5cwP33348jR47g1ltvxTPPPIPaLzO44447oNfr8cUXX7RIQYmIiIiIgCCHELz11lsoLS3Fq6++invvvRcA8PrrrwfMYzabMXDgQOTk5DS/lEREREREXkFdgd21axf69+/vS14bkpycjMLCwqAKRkRERERUn6AS2MuXL6NPnz5Xnc/tdsNmswWzCiIiIiKiegWVwMbFxeHChQtXne/MmTNISEgIZhVERERERPUKKoFNTU3Ff/7znyv+JGx2dja+++47pKWlBV04IiIiIqLagkpgH3nkESiKgqeeego7duyAx+MJ+Pzrr7/GkiVLoNPpMG/evBYpKBEREREREORbCIYNG4Zf//rX+N///V888cQTMJlMkCQJX3zxBbZt24aqqipIkoQXXngBAwYMaOkyExEREVEYC/qHDObMmYOMjAzceuutkCQJiqKguroaTqcT48aNw9q1a/HTn/60JctKRERERNS8n5K98cYb8fvf/x6KoqC0tBSyLMNisUCr1bZU+YiIiIiIAjQrgVVJkoT4+PiWWBQRERER0RW1SALbGi5duoStW7di586dOH36NIqLixEbG4vU1FQ8+uijGDZsWKOXJcsyMjIysGHDBuTl5cFsNmPMmDFYvHgxevTo0YpbQUREREQtrVEJ7G233Rb0CiRJwrZt25r8vbVr1+K9995Dz549MXbsWMTHxyMvLw/btm3Dtm3bsGLFCkybNq1Ry3r++eeRmZmJfv36Ye7cuSgsLMTWrVuxe/durF+/Hr17925y+YiIiIiobTQqgc3Pz2/ygtUHu4I1dOhQrF27ts57ZA8cOIAHH3wQL774IiZNmgSDwXDF5ezduxeZmZkYOXIk/vznP/vmv/POO/H444/jlVdewZ/+9Kegy0lERERE11ajEtiTJ0/Wmfbqq6/i448/xpw5czB9+nR0794dgEh2P/30U2RkZODuu+/Gc889F1TBbr/99nqnjxgxAqNGjUJWVha+/fZbDBky5IrLyczMBAAsWrQoINmdMGEC0tLSkJWVhYKCAnTr1i2octbmcCj497+t+McWGVYrEBMDSAAkLWCtAqwOwKgDDCbA7QQUCTCZANkNVFYBkgQY9IDLDSgK4JbFqyJ0OkCBmK4BoNEA0AARBjHN6QZ0GjGPRwbMRiA6FoAM2GxAtVUsT0b969dpAI9HzCcDiIkCZI9Yh1EvPistBzxuwGgAomIASRHTXE7AYAAsFkCRgYpKf3kk73suZFmULTYK0GoDvxcZCVhtYtmyWCXcsii7RiPWFxEplud0Aja7qCNJEvPbrOLfESZ/XRh1gFYvPlNfU6zT+eexO/31KMNfVoMBiIgAHDYxDyCmRUUBHhfgcgEaHVBRIbY1NkbMb7OJ7Vbr2Gy6DJfbv+0eWaxDp/eu0w0YTYHbrbafTieWW1EOuD3+OtRqAb0ecNj95dfpxOey7O8b5gh/2zud4jONTtQFFMDh9NeXJIm+6XaKOtdqRflq1qkse9sDgeVQ+09kJGA2AZDE96qrRRu43KIMBkPgfuB2is8ddrEOnd6/XWrbGUz+vzXw14/RFNg+at2q82ggtjUyAqiqFnWr0fj7k9oH1G3SakX/lxTR92uuX/HU7V8N1YXDLqYbtGK7ExOAWbMc6NG9OdGk/Tl+XEbWnlLkF8i+ejHoRRtoNf6+FRUp5q+9r8puILELkNAZKC8HysoAux0wRog4o1L3WYMBvn4ly0BkjGh7RQbi4gG7Vex7Ho/YN3UG0ZaS5J+3ssLfT3UawGoFXB7RL3QaQP2xSHUfkj1iv6moFGW2xAIaSUyz2/0xU+0zbhmIihD90uMByitF+Ywmf9yw20XZdBrRZ9TnnD0eUR5JC8R3AjxOwOEAJB1QUiz+rdEA0VGAOVpMUxQRa2uWyekGzGagbx+gygaUXhZ1EBcPuBxiPU6XqHOtRmynWxbHD50uMKapccEYIbaxvNK7H+lEXZgiSlBVofjqwmIBEuK9dVYl6lv2tp3HLerd5q3ziAixX0la0TZWu4h5EQb/Os0mUdaqKqBTJ8ASB1y8CBRfBnRacSyQFLHd6j5ssQCWGBFPbTZRpzKA2FjR/4ze7SwpFdteO37UjC1upzh26DSi/tV/q+0UFQ24HWL5Dpt/W21WUbeS5G9rtV9KWhGb9EaxfINJxKhqq79/ajQijigKAI04Vqp9yun0L0eNu3o9oNcCehNQXiK+px6Xq6pFOxoMotzqsUyWAaMR6NpF1FNpaeCyTJGA4gZ03nLabaKeNTp/XUaageEj7OjdU0Z0dNAvsmoRQY2B/fDDD/HXv/4VH330UZ0Esn///ujfvz8mT56M//qv/0L37t3x4IMPtkRZfXQ6XcB/r2Tfvn0wm81ITU2t89n48eORnZ2N7Oxs3HXXXc0u18VLCl56ScF/jtuavSwiCm2bPqnCj+8Envll2wb5lrL8tzL+8QmgKPLVZ6Y28cU1WUvwd1apY/jzB9XonwIsXCBjyKC2i29BrXn9+vUYNWrUFa9+DhkyBKNGjfJdAW0pBQUF2LNnDxISEpCSknLFea1WK4qKitC9e/d6X+3Vq1cvAEBeXl6zy+VwKNj8DwX/Od7sRRFRByDLwKefiauWoe74cTV5beuSEFFbk2Xgu++BzEygsrLt4ltQV2DPnTuH/v37X3W+2NhYHDhwIJhV1MvlcmHJkiVwOp341a9+ddX3zVZWVgIAoqKi6v1cna7O15DY2FhoNFfO9UtLZZzNK4e4oUBEJG7/nT0bgbFjzW1dlHo1JrYBwPenq6Eo9mtQIiIKBR4PUFKihaJEwWJpmxdaBbXW2NhY7N+/Hw6HA0ajsd55HA4H9u/fj5iYmGYVUCXLMpYuXYr9+/fjvvvua5Fb/o1VXl5+1XkcDgW9eynYlXUNCkREIUGrBXr3tqG01HHVeS0WyzUoUaDGxDYA6HudDEniFVgiErRaID7eA0kqR2np1U+CWyO+BTWEYPLkySgqKsLChQtx/vz5Op+fP38eixYtQnFxMSZPntzsQsqyjGeffRZbtmzBjBkz8NJLLzXqe9HR0QCAqqqqej9Xp6vzNYfRKOEnMyQMuaHZiyKiDkCjAaZPA264IfTHwN5wgwYzfiweUCGi8KbRAP36ArNmoU0f5ArqCuyiRYuwd+9e7NixA1lZWRg8eLDvKf6CggJ88803cLvduO6667Bo0aJmFVCWZSxbtgybNm3CnXfeiddee61Rt7wAwGw2IyEhAefPn4fH46kz5EAd+6qOhW2upC4SfvcGcPJbEz79zIaqavE0uSSJ/1mt4ilJo0H8z+0Ww+FNRvEkZGUlAEl85nJ530LgEd/Vq28hcIm/tRoxb4QJcLsAh0s8oalAXNo3RwDR0eJvu008Ga4ogKzUv36tzvtEbLWYJzpajHORABiM3rcQlInvmIziqVhIYprT+zaBeIv4blWFvzxqU3lksf7YGO9bCGp8LypSPBHpdovvaySx3Yoivm8yiu3R6rxP3doAvUHMJ8uiTiGJedS6MBrE+q02sSxAPIWqzmNz+OtRVvxlNRjEE8YOh/gf4H9TgtsjniiWNN4nlGUgLlY8KWpVn3b21nGkWbSVuu0ej1iHXu8vt8EYuN1q++l1ov6rKr1vFqj5FgKdKJdafr1OfO6R/X0j0uxve5dTPFksabxPUCve7fLWl8bbN91usX1arZhWs04V2V+HNcuh9p9IMxDhvUNutQG2atEGah82GAL3A7dblMuuboPev11q2xkN/r81kr9+TKbA9lHrVp1HI4nymc3iSVy3W7Sx2p/UPqBuk1Yr+hck0fdrrl9R6vavhurC4RBPWRv0YrsTE4F7ZkahR3dri8SW9uCZX2owfaqMvXu1OJ/v8dWL0SCe3Ndo/X0r2jtqq/a+KstAQqJ4ury8QjwV77SLJ6MNRtGfJHjjHrxvIYC3DWSxH9odABSx71XbxZPbHo/Yr3U6sQCN5J+3qtLfT7U6EYfdbjGPTgfEd/ZuoPoWAlm8TaGqQvSpuDix/7ic3qfEEdhnXB4Rw9QnxisqxDJMJn/ccNrFurQ6/1sIFIh6s1rFPhxvEfM6HKJsJSViP9dogJhoUf6SEsCjiHlrlsnhEp/36SOejC8qAbSSqCO7S/Rbp1OUTaMRdeXyiOOHXhcY09S4YPRuY0WFdz/SirowmSRUVSm+urDEAZ07e+u6ShzHFEXs17Is6sVu9x47TOIzSRLbabWJvyNM/nVGmEVZq6uB+Hjx9PvFS0DJZf/+C8n75g/vPmyJE2WrsorpVu9bd2JiRB8yGkT/LC3zHiNqxY+ascXtFn9rvfu1+m+1naIia/QTh39b7Xbxudqv1GOq2xuXNd630Li9b/Ixm8VbCNT+qZFEHSrenUB9w43bI9pZXY4ad/U6b+zUi/1IVvzH5epqsS3qWwjUY5lHFn030fsWgoqywGWZTGI5ep33bT3e/Uuj9ddlpBlITY1Er57VofkWgtjYWPz1r3/FihUrsHnzZuTk5CAnJ8f3uclkwsyZM/H0008jNjY26MLVTF6nTZuG119//arjXmtLS0vDp59+ikOHDmHkyJEBn+3atQsA6kxvDqNRwi0TzBg29Oq3DKnjslgsKC0tbetiUBuyWIwoLe04CSwgrsSOHRvHvh3GGNvIYjGhtLTt37YU9MjbmJgYvPTSS1i6dCm++eYbFBYWAgASEhIwaNAgmM3Ne2hBHTawadMmTJ06FcuXL79i8lpSUoLS0lJYLBbEx8f7pt9333349NNPsXLlyoAfMtixYweys7Mxbtw4JCcnN6usRERERHTtNPvRsYiICIwYMaLR82dmZuLQoUNIT0+/4nxvvfUWNm7cCLPZjN69e+Ptt9+uM8+kSZMwcOBAAEBGRgZWr16N+fPnY8GCBb55Ro8ejVmzZiEzMxMzZ87EhAkTUFRUhM8++wxxcXFB/9ACEREREbWNa/7ug0OHDmHTpk1XTWDVn6+1Wq1455136p0nOTnZl8Beycsvv4yUlBRs2LABH374IcxmMyZPnozFixejZ8+eTd8IIiIiImozkqJc2xejqGNaT5w4cS1X2yxNHe/DMULEPkBN7QNt8RqtYPoo+3Z4Y/tTMH2g3bxGi4iIiIiorTCBJSIiIqKQwgSWiIiIiEIKE1giIiIiCilMYImIiIgopDCBJSIiIqKQwgSWiIiIiELKNU9gr/FrZ4mIiIiog7nmv8T1+OOPY+bMmdd6tURERETUQTQrgf3++++xYcMGHD16FKWlpbjtttuwZMkSAOInY48dO4YZM2YgLi7O953rrrsO1113XbMKTUREREThK+gE9v3338eKFSvgdrsBAJIk1flpsfT0dBgMBsyePbt5pSQiIiIi8gpqDOy///1v/OY3v0FSUhJWr16NPXv21Bnbmpqaivj4eHz11VctUlAiIiIiIiDIK7Dvv/8+IiIi8P7776NHjx4NzjdgwACcOXMm6MIREREREdUW1BXYb775BjfeeOMVk1cAsFgsKC4uDqpgRERERET1CSqBdblciIyMvOp8JSUl0Gq1wayCiIiIiKheQSWw3bt3x8mTJ684j9PpxLfffovevXsHswoiIiIionoFlcBOnDgR+fn5eP/99xuc549//CNKSkpw++23B104IiIiIqLagnqI69FHH8Unn3yC119/HTk5OZg8eTIA4PLly/jyyy/x5Zdf4pNPPkH37t0xZ86cFi0wEREREYU3SQnyt13PnDmDhQsX4rvvvoMkSVAUBZIkARA/F9u3b1+89dZb6NWrV4sWuC3Ufr/t1VgsliZ/hzoW9gFqah+wWCytWJr6BdNH2bfDG9ufgukDrRHfgv4hgz59+mDz5s3Yvn07du/ejfz8fMiyjKSkJIwZMwZTpkzhA1xERERE1OKa9VOyGo0GkyZNwqRJk1qqPEREREREVxTUQ1xERERERG2lWQlsVlYWfvGLX2D8+PEYPHgwli1b5vts165dSE9Px6VLl5pdSCIiIiIiVdBDCF599VVkZGRAURSYzWa43e6AzxMSEvDBBx+ga9euePDBB5tbTiIiIiIiAEFegd20aRPWrVuHQYMGYePGjTh06FCdeQYMGICuXbti+/btzS4kEREREZEqqCuwH330EWJiYvDuu+8iPj6+wfn69++P3NzcoAtHRERERFRbUFdgc3NzcdNNN10xeQWAqKgoFBcXB1UwIiIiIqL6BP0Ql/qjBVdSWFgIk8kU7CqIiIiIiOoIKoHt3bs3vvnmG7hcrgbnqaqqwsmTJ9G3b9+gC0dEREREVFtQCezUqVNRVFSEFStWNDjPb3/7W1RWVmL69OlBF46IiIiIqLagHuKaN28ePv30U3zwwQc4fPgwbrvtNgDAuXPnsGbNGnz55Zc4ePAgbrjhBsyaNSvowm3evBkHDx7EsWPHkJubC5fLhfT0dMycObPRy9i3bx8eeOCBBj9v6vKIiIiIqG0FlcCaTCasWbMGS5cuxc6dO3H06FEAwIEDB3DgwAEAwNixY7F8+XIYDIagC7dy5Urk5+fDYrEgMTER+fn5QS8rLS0NaWlpdaYPHDgw6GUSERER0bUX9A8ZxMfH491338XJkyeRlZWF/Px8yLKMpKQkjB07FkOHDm124V599VX06tULycnJePfdd684ZOFq0tLSsGDBgmaXiYiIiIjaVlAJ7Pz585GQkIAXXngBAwYMwIABA1q6XACAMWPGtMpyiYiIiCh0BZXA7tixA5MmTWrpsrSqs2fPYs2aNXA4HOjSpQtuvvlmdOnSpa2LRURERERNFFQC2717d9hstpYuS6vasmULtmzZ4vtbp9Ph/vvvx5IlS6DVatuwZERERETUFEElsNOnT8ef//xnFBUVISEhoaXL1KLi4+Px9NNP49Zbb0VycjJsNhsOHz6MFStWYM2aNZAkCUuXLr3iMmJjY6HRNO2NYxaLpTnFpg6AfYDaex8IJrYB7X+7qHWx/ak99IGgEtif//znOHr0KO6//3786le/wi233AK9Xt/SZWsR/fr1Q79+/Xx/m81mTJo0CcOGDcOMGTOwdu1aPPbYY+jUqVODyygvL2/SOi0WC0pLS4MuM4U+9gFqah9oiwNCU2MbwL4d7tj+FEwfaI34FlQCO3XqVCiKggsXLmDhwoWQJAnx8fEwGo115pUkCdu2bWt2QVtaQkICbrvtNmRmZiInJwcTJ05s6yIRERERUSMElcDWfh+roigoLi5ukQJdS+oZQaiN5yUiIiIKZ0ElsCdPnmzpcrSJnJwcAEBycnIbl4SIiIiIGqvpo/cBLFu2DH//+9+vOt/GjRvx7LPPBrOKJispKcGpU6dQUlISMP3YsWP1zv/BBx9g37596N27N4YMGXItikhERERELSCoK7AbN24EANxzzz1XnO/gwYPYuHEj/vd//zeY1SAzMxMHDx4EAOTm5vqmZWdnAwCGDx+OWbNmAQAyMjKwevVqzJ8/P+AXtxYuXAidTofBgwejS5cusNlsyMnJwfHjxxETE4Ply5fzNVpEREREISTon5JtDJfL1azkUE2Aazp06BAOHTrk+1tNYBsye/ZsZGVlYf/+/SgrK4NGo0G3bt0wb948PPzww0hKSgq6fERERER07UmKoihN/dKAAQNw9913Iz09vcF5FEXBjBkzUFZWhl27djWrkG0tmNdF8DUj4Y19gELhNVrB9FH27fDG9qeQe43WAw88EPD3rl276kxTeTwe/PDDDyguLsZPfvKT5pWQiIiIiKiGRiew6rhTQLzbtbi4+IqvztLpdLjllluwZMmS5pWQiIiIiKiGRiewX331FQAxNGDSpEmYMmVKg8mpXq+HxWJpt7/ORUREREShq9EJbM13pc6fPx8DBw7k+1OJiIiI6JoL6i0E8+fPb+lyEBERERE1SlA/ZEBERERE1FaYwBIRERFRSGECS0REREQhhQksEREREYUUJrBEREREFFKYwBIRERFRSGECS0REREQhhQksEREREYUUJrBEREREFFKYwBIRERFRSGECS0REREQhhQksEREREYUUJrBEREREFFKYwBIRERFRSGECS0REREQhhQksEREREYUUJrBEREREFFKYwBIRERFRSGECS0REREQhhQksEREREYUUJrBEREREFFKYwBIRERFRSGECS0REREQhhQksEREREYUUXVsX4Eo2b96MgwcP4tixY8jNzYXL5UJ6ejpmzpzZpOXIsoyMjAxs2LABeXl5MJvNGDNmDBYvXowePXq0UumJiIiIqDW06wR25cqVyM/Ph8ViQWJiIvLz84NazvPPP4/MzEz069cPc+fORWFhIbZu3Yrdu3dj/fr16N27d8sWnIiIiIhaTbseQvDqq69i+/bt2Lt3L2bPnh3UMvbu3YvMzEyMHDkSH3/8MZ555hksX74cb731FsrKyvDKK6+0cKmJiIiIqDW16yuwY8aMafYyMjMzAQCLFi2CwWDwTZ8wYQLS0tKQlZWFgoICdOvWrdnrIiIiIqLW166vwLaEffv2wWw2IzU1tc5n48ePBwBkZ2df62IRUQMcDgVlZQocDqWti0JERO1Uu74C21xWqxVFRUVISUmBVqut83mvXr0AAHl5ede6aERUj4uXFOTkKKiqBqIigWHDgKQuUlsXi4iI2pkOfQW2srISABAVFVXv5+p0dT4iajsOh0heyyuBuDigvBLIyeGVWCIiqqtDX4FtKbGxsdBompbrWyyWVioNhQr2gaYpLZUhKw706qGB0SghKlJBSakMk8mIuLjQPNdu730gmNgGtP/totbF9qf20Ac6dAIbHR0NAKiqqqr3c3W6Ol9DysvLm7Rei8WC0tLSJn2HOhb2gaZzOBRoJAV554BO8cDlEiA2GrDbbSgtDb1hBE3tA21xQGhqbAPYt8Md25+C6QOtEd9C87JGI5nNZiQkJOD8+fPweDx1PlfHvqpjYYmo7RiNEoYNkxAbDZSVieR12DAJRmPoJa9ERNS6OnQCCwBpaWmwWq04dOhQnc927doFABg5cuS1LhYR1SOpi4RbJkiYNFH8lw9wERFRfTpMAltSUoJTp06hpKQkYPp9990HQPyql9Pp9E3fsWMHsrOzMW7cOCQnJ1/TshJRw4xGCbGxvPJKREQNa9djYDMzM3Hw4EEAQG5urm+a+t7W4cOHY9asWQCAjIwMrF69GvPnz8eCBQt8yxg9ejRmzZqFzMxMzJw5ExMmTEBRURE+++wzxMXF4bnnnrvGW0VEREREzdGuE9iDBw9i48aNAdMOHToUMBxATWCv5OWXX0ZKSgo2bNiADz/8EGazGZMnT8bixYvRs2fPFi83EREREbUeSVEUvmTxKoJ52o5PaYY39gEKhbcQBNNH2bfDG9uf+BYCIiIiIqIg8AosEREREYUUXoElIiIiopDCBJaIiIiIQgoTWCIiIiIKKUxgiYiIiCiktOv3wIaao0ePYtWqVTh8+DDcbjdSUlLw4IMPYtq0aW1dNGrA5s2bcfDgQRw7dgy5ublwuVxIT0/HzJkz652/qqoKq1atwhdffIGioiIkJiZiypQpmD9/PiIjI+vML8syMjIysGHDBuTl5cFsNmPMmDFYvHgxevToUe86du3ahT/84Q/45ptvIEkSBg0ahKeeego333xzi257uLt06RK2bt2KnTt34vTp0yguLkZsbCxSU1Px6KOPYtiwYXW+E67tz9gWehjbwls4xDe+haCF7N27F48++igMBgOmT5+OyMhIfPHFF8jPz8f/+3//Dw8//HBbF5HqMXHiROTn58NiscBsNiM/P7/BIG+1WvGzn/0MJ06cwLhx4zBw4ECcOHECWVlZGDJkCDIyMmA0GgO+89xzzyEzMxP9+vXDhAkTUFhYiK1btyIyMhLr169H7969A+bfvHkzlixZgvj4eF9y8Nlnn6G0tBS/+93vMHXq1Fari3Dzf//3f3jvvffQs2dPpKWlIT4+Hnl5edi2bRsURcGKFSsCErRwbX/GttDE2BbewiK+KdRsLpdLmTRpkjJ48GDl+PHjvukVFRXK7bffrgwaNEg5f/58G5aQGrJ7925f2/zhD39QUlJSlL///e/1zrty5UolJSVFWb58ecD05cuXKykpKco777wTMP3rr79WUlJSlDlz5igOh8M3/d///reSkpKiPPzwwwHzl5WVKSNGjFBGjRqlXLhwwTf9woULyqhRo5RRo0YplZWVzdpe8vv888+Vffv21Zm+f/9+ZdCgQcrIkSMD2i0c25+xLXQxtoW3cIhvHAPbAvbu3YsffvgBd955JwYOHOibHh0djSeeeAIul6vOT+JS+zBmzBgkJydfdT5FUZCZmQmz2Yynnnoq4LOnnnoKZrMZmZmZAdPVvxctWgSDweCbPmHCBKSlpSErKwsFBQW+6f/85z9RUVGB+++/H0lJSb7pSUlJuP/++1FaWopt27YFtZ1U1+233460tLQ600eMGIFRo0ahvLwc3377LYDwbX/GttDF2BbewiG+MYFtAdnZ2QCAcePG1flMnbZ///5rWiZqWWfPnkVhYSFSU1NhNpsDPjObzUhNTcW5c+dw4cIF3/R9+/b5Pqtt/PjxAPx9p+a/r9SPas5PrUen0wX8N1zbn7Gt4wvXvh3OOkp8YwLbAs6ePQsA6NWrV53PEhISYDabkZeXd41LRS1Jbb/aY3pU6nS1L1itVhQVFaF79+7QarV15lf7Ss1+caV+VN/81DoKCgqwZ88eJCQkICUlBUD4tj9jW8cXrn07XHWk+MYEtgVUVVUBELfV6hMVFYXKysprWSRqYWr7RUVF1fu5Ol3tC42dv2a/uFI/qm9+ankulwtLliyB0+nEr371K19wDtf2Z2zr+MK1b4ejjhbfmMASEUG8Embp0qXYv38/7rvvPtx1111tXSQiohbREeMbE9gWcLWzh6qqqgavYFBoUNtPPYusTZ2u9oXGzl+zX1ypH13tShg1jyzLePbZZ7FlyxbMmDEDL730UsDn4dr+jG0dX7j27XDSUeMbE9gWoI4PqW/8RlFREaxWa73jPih0qO2njuOpTZ2u9gWz2YyEhAScP38eHo+nzvxqX6nZL67Uj+qbn1qGLMtYtmwZNm7ciDvvvBOvvfYaNJrA0Biu7c/Y1vGFa98OFx05vjGBbQEjR44EAGRlZdX5TJ2mzkOhqXfv3khMTMShQ4dgtVoDPrNarTh06BC6d++Orl27+qanpaX5Pqtt165dAAL7RWP6UX2vRaHgqcF906ZNmDZtGl5//fV6H0oI1/ZnbOv4wrVvh4OOHt+YwLaAm2++GT169MCWLVtw4sQJ3/TKykq888470Ov1HWK8STiTJAmzZs2C1WrF73//+4DPfv/738NqteK+++4LmK7+vXLlSjidTt/0HTt2IDs7G+PGjQt4T+Mdd9yB6OhorFu3DhcvXvRNv3jxItatWweLxYJJkya1xuaFJfW22qZNmzB16lQsX7683uAOhG/7M7Z1fOHatzu6cIhv/CnZFsKfWwxNmZmZOHjwIAAgNzcX33zzDVJTU323MoYPH45Zs2YBEGei//Vf/4WTJ09i3LhxuOGGG3D8+HHfT+2tW7cOJpMpYPm1f2qvqKgIn332GSIjI/HXv/4Vffr0CZj/Sj+198Ybb+COO+5o7SoJG6tWrcLq1athNpvxwAMP+N6JWNOkSZN8L/AP1/ZnbAtNjG3hLRziGxPYFnT06FG8+eabOHz4MNxuN1JSUvDQQw8F/N4wtS9Lly694i8J3X333Xjttdd8f1dWVmLVqlX44osvUFxcjISEBEydOhW/+MUv6n2liCzLWLduHTZs2IC8vDyYzWaMGTMGixcvRs+ePetd586dO/GHP/wBx48fBwAMHjwYTz75JMaMGdPMraWartb2AOr8dny4tj9jW+hhbAtv4RDfmMASERERUUjhGFgiIiIiCilMYImIiIgopDCBJSIiIqKQwgSWiIiIiEIKE1giIiIiCilMYImIiIgopDCBJSIiIqKQwgSWiIiIiEIKE1giIiIiCilMYIlqWbVqFfr374+PP/64rYtCRNRiGNuoI2ECS0REREQhhQksEREREYUUJrBEREREFFKYwFLY+uqrr/DTn/4Uw4YNw6hRo7BgwQKcOXOm3nknTpyI/v37Q1EUrF27FjNmzMCwYcPwk5/8BACgKAq2bNmCxYsXY8qUKbjxxhtx00034d5770VGRgZkWQ5Y3scff4z+/ftj1apVAdPLysowYMAA9O/fH5mZmQGf5ebmon///vj5z3/egrVARB0NYxuFA11bF4CoLXz00Ud48cUXIUkSRowYgYSEBOTk5GDWrFm49dZbG/zeCy+8gI8//hgjR47E9ddfD5fLBQBwOp14+umnERcXh759++KGG25AWVkZDh8+jJdffhn/+c9/8Nprr/mWM3LkSABAdnZ2wPL3798PRVEAAPv27cOsWbN8n+3btw8AkJaW1jKVQEQdDmMbhQsmsBR28vPzkZ6eDr1ej7fffhvjx48HALhcLixbtgz/+Mc/GvzuF198gY0bN6Jfv34B07VaLd566y1MmDABer3eN72kpASPPfYYNm7ciHvuuccX3Hv06IHk5GQcOXIEDocDRqMRgD+Q9+vXD/v37w9Yh3pAYJAnovowtlE44RACCjt///vf4XA4MH36dF+ABwC9Xo9f//rXiIiIaPC7jz32WJ0ADwA6nQ6TJk0KCPAAEB8fj6effhqAuK1X08iRI+F0OnHkyBHftOzsbFx//fWYMmUKLl68iLy8PADiNt7+/fsRFRWFG264ocnbTEQdH2MbhRNegaWwc+DAAQDAtGnT6nxmsVgwduxYbNu2rd7vTpw48YrLPnHiBLKyslBQUAC73Q5FUVBdXQ0AOHv2bMC8I0eOxKZNm5CdnY1Ro0ahrKwMubm5mD17tu9KRHZ2Nnr16oXc3FyUlpZiwoQJ0Gq1Td1kIgoDjG0UTpjAUtgpLCwEACQnJ9f7eUPTAaBbt271Tnc6nVi2bBm2bNnS4HfVYK8aNWoUAP/tM3WM2KhRo3DjjTfCYDD4xorxFhsRXQ1jG4UTJrBETaCO56ptzZo12LJlC1JSUvDMM89g0KBBiImJgV6vx5kzZzB16tQ63+nRowe6du3qGyumjhEbNWoUjEYjbrzxRt9YMQZ5ImpNjG0UajgGlsJOQkICAPHAQ30KCgqavMwvv/wSAPDb3/4WP/rRj9CpUyffmLFz5841+L2aY8Wys7PRt29fxMfHAxAB/eLFizh79iz279+PyMhIDBo0qMllI6LwwNhG4YQJLIWdESNGAAD++c9/1vmsrKwMu3fvbvIyKyoqAABJSUl1Ptu6dWuD31OvOnz55ZfIzc313Xqr+dm6detQWlqK4cOHc4wYETWIsY3CCRNYCjszZ86EwWDAJ598gj179vimu1wupKenw2q1NnmZvXv3BiDewVjTP//5T2zevLnB76lBfcOGDVAUJeA2mjpWbMOGDQD871ckIqoPYxuFEyawFHZ69OiBpUuXwuVy4ZFHHsHcuXPxy1/+ElOmTMFXX32FH//4x01e5qOPPgqtVosVK1Zg5syZePrpp3HPPfdg0aJFmDdvXoPf69mzJ5KSkuBwOCBJUkCQV8eKORwOAAi4gkFEVBtjG4UTJrAUlubMmYO33noLQ4YMwdGjR5GVlYUBAwZg/fr16NWrV5OXN3LkSPzlL3/B6NGjcf78efzrX/+CXq/HqlWrMGfOnKt+F0DAGDGVGvTNZjPHiBHRVTG2UbiQFPW33YiIiIiIQgCvwBIRERFRSGECS0REREQhhQksEREREYUUJrBEREREFFKYwBIRERFRSGECS0REREQhhQksEREREYUUJrBEREREFFKYwBIRERFRSGECS0REREQhhQksEREREYUUJrBEREREFFKYwBIRERFRSPn/S8H+HqrRhUoAAAAASUVORK5CYII=\n", "text/plain": [ - "[]" + "
" ] }, - "execution_count": 5, "metadata": {}, - "output_type": "execute_result" - }, - { - "data": { - "image/png": "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\n", - "text/plain": [ - "
" - ] - }, - "metadata": { - "needs_background": "light" - }, "output_type": "display_data" } ], "source": [ - "plt.plot(trace[\"step_size_bar\"])" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "The `get_sampler_stats` method provides more control over which values should be returned, and it also works if the name of the statistic is the same as the name of one of the variables. We can use the `chains` option, to control values from which chain should be returned, or we can set `combine=False` to get the values for the individual chains:" + "trace.sample_stats[\"tree_depth\"].plot(col=\"chain\", ls=\"none\", marker=\".\", alpha=0.3);" ] }, { "cell_type": "code", - "execution_count": 6, + "execution_count": 7, "metadata": {}, "outputs": [ { "data": { - "image/png": "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\n", + "image/png": "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\n", "text/plain": [ - "
" + "
" ] }, - "metadata": { - "needs_background": "light" - }, + "metadata": {}, "output_type": "display_data" } ], "source": [ - "sizes1, sizes2 = trace.get_sampler_stats(\"depth\", combine=False)\n", - "fig, (ax1, ax2) = plt.subplots(2, 1, sharex=True, sharey=True)\n", - "ax1.plot(sizes1)\n", - "ax2.plot(sizes2)\n", - "\n", - "plt.show()" + "az.plot_posterior(\n", + " trace, group=\"sample_stats\", var_names=\"acceptance_rate\", hdi_prob=\"hide\", kind=\"hist\"\n", + ");" ] }, { - "cell_type": "code", - "execution_count": 7, + "cell_type": "markdown", "metadata": {}, - "outputs": [ - { - "name": "stderr", - "output_type": "stream", - "text": [ - "/Users/CloudChaoszero/opt/anaconda3/envs/pymc3-dev-py38/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n", - " warnings.warn(msg, FutureWarning)\n" - ] - }, - { - "data": { - "image/png": "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\n", - "text/plain": [ - "
" - ] - }, - "metadata": { - "needs_background": "light" - }, - "output_type": "display_data" - } - ], "source": [ - "accept = trace.get_sampler_stats(\"mean_tree_accept\", burn=1000)\n", - "sns.distplot(accept, kde=False)\n", - "\n", - "plt.show()" + "We check if there are any divergences, if yes, how many?" ] }, { @@ -285,8 +682,364 @@ "outputs": [ { "data": { + "text/html": [ + "
\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "
<xarray.DataArray 'diverging' ()>\n",
+       "array(0)
" + ], "text/plain": [ - "0.8115665358193459" + "\n", + "array(0)" ] }, "execution_count": 8, @@ -295,34 +1048,14 @@ } ], "source": [ - "accept.mean()" + "trace.sample_stats[\"diverging\"].sum()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "Find the index of all diverging transitions:" - ] - }, - { - "cell_type": "code", - "execution_count": 9, - "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "(array([], dtype=int64),)" - ] - }, - "execution_count": 9, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "trace[\"diverging\"].nonzero()" + "In this case no divergences are found. If there are any, check [this notebook](https://github.com/pymc-devs/pymc-examples/blob/main/examples/diagnostics_and_criticism/Diagnosing_biased_Inference_with_Divergences.ipynb) for information on handling divergences." ] }, { @@ -336,29 +1069,22 @@ }, { "cell_type": "code", - "execution_count": 10, + "execution_count": 9, "metadata": {}, "outputs": [ { "data": { - "image/png": "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\n", + "image/png": "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\n", "text/plain": [ - "
" + "
" ] }, - "metadata": { - "needs_background": "light" - }, + "metadata": {}, "output_type": "display_data" } ], "source": [ - "energy = trace[\"energy\"]\n", - "energy_diff = np.diff(energy)\n", - "sns.histplot(energy - energy.mean(), label=\"energy\")\n", - "sns.histplot(energy_diff, label=\"energy diff\")\n", - "plt.legend()\n", - "plt.show()" + "az.plot_energy(trace, figsize=(6, 4));" ] }, { @@ -374,34 +1100,32 @@ "source": [ "## Multiple samplers\n", "\n", - "If multiple samplers are used for the same model (e.g. for continuous and discrete variables), the exported values are merged or stacked along a new axis.\n", - "\n", - "Note that for the `model_logp` sampler statistic, only the last column (i.e. `trace.get_sampler_stat('model_logp')[-1]`) will be the overall model logp." + "If multiple samplers are used for the same model (e.g. for continuous and discrete variables), the exported values are merged or stacked along a new axis.\n" ] }, { "cell_type": "code", - "execution_count": 11, + "execution_count": 10, "metadata": {}, "outputs": [], "source": [ - "model = pm.Model()\n", - "with model:\n", + "coords = {\"step\": [\"BinaryMetropolis\", \"Metropolis\"], \"obs\": [\"mu1\"]}\n", + "dims = {\"accept\": [\"step\"]}\n", + "\n", + "with pm.Model(coords=coords) as model:\n", " mu1 = pm.Bernoulli(\"mu1\", p=0.8)\n", - " mu2 = pm.Normal(\"mu2\", mu=0, sigma=1, shape=10)" + " mu2 = pm.Normal(\"mu2\", mu=0, sigma=1, dims=\"obs\")" ] }, { "cell_type": "code", - "execution_count": 12, + "execution_count": 11, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ - "/Users/CloudChaoszero/Documents/Projects-Dev/pymc3/pymc3/sampling.py:465: FutureWarning: In an upcoming release, pm.sample will return an `arviz.InferenceData` object instead of a `MultiTrace` by default. You can pass return_inferencedata=True or return_inferencedata=False to be safe and silence this warning.\n", - " warnings.warn(\n", "Multiprocess sampling (2 chains in 2 jobs)\n", "CompoundStep\n", ">BinaryMetropolis: [mu1]\n", @@ -426,7 +1150,7 @@ " }\n", " \n", " \n", - " 100.00% [22000/22000 00:15<00:00 Sampling 2 chains, 0 divergences]\n", + " 100.00% [22000/22000 00:04<00:00 Sampling 2 chains, 0 divergences]\n", " \n", " " ], @@ -441,8 +1165,8 @@ "name": "stderr", "output_type": "stream", "text": [ - "Sampling 2 chains for 1_000 tune and 10_000 draw iterations (2_000 + 20_000 draws total) took 28 seconds.\n", - "The number of effective samples is smaller than 10% for some parameters.\n" + "Sampling 2 chains for 1_000 tune and 10_000 draw iterations (2_000 + 20_000 draws total) took 11 seconds.\n", + "The number of effective samples is smaller than 25% for some parameters.\n" ] } ], @@ -450,27 +1174,35 @@ "with model:\n", " step1 = pm.BinaryMetropolis([mu1])\n", " step2 = pm.Metropolis([mu2])\n", - " trace = pm.sample(10000, init=None, step=[step1, step2], cores=2, tune=1000)" + " trace = pm.sample(\n", + " 10000,\n", + " init=None,\n", + " step=[step1, step2],\n", + " cores=2,\n", + " tune=1000,\n", + " return_inferencedata=True,\n", + " idata_kwargs={\"dims\": dims, \"coords\": coords},\n", + " )" ] }, { "cell_type": "code", - "execution_count": 13, + "execution_count": 12, "metadata": {}, "outputs": [ { "data": { "text/plain": [ - "{'accept', 'accepted', 'p_jump', 'scaling', 'tune'}" + "['accept', 'accepted', 'p_jump', 'scaling']" ] }, - "execution_count": 13, + "execution_count": 12, "metadata": {}, "output_type": "execute_result" } ], "source": [ - "trace.stat_names" + "list(trace.sample_stats.data_vars)" ] }, { @@ -480,6 +1212,39 @@ "Both samplers export `accept`, so we get one acceptance probability for each sampler:" ] }, + { + "cell_type": "code", + "execution_count": 13, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "az.plot_posterior(\n", + " trace,\n", + " group=\"sample_stats\",\n", + " var_names=\"accept\",\n", + " hdi_prob=\"hide\",\n", + " kind=\"hist\",\n", + ");" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "We notice that `accept` sometimes takes really high values (jumps from regions of low probability to regions of much higher probability). " + ] + }, { "cell_type": "code", "execution_count": 14, @@ -487,14 +1252,373 @@ "outputs": [ { "data": { + "text/html": [ + "
\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "
<xarray.DataArray 'accept' (chain: 2, accept_dim_0: 2)>\n",
+       "array([[  3.75      , 269.57489704],\n",
+       "       [  3.75      , 147.77852694]])\n",
+       "Coordinates:\n",
+       "  * chain         (chain) int64 0 1\n",
+       "  * accept_dim_0  (accept_dim_0) int64 0 1
" + ], "text/plain": [ - "array([[1. , 0.01449865],\n", - " [1. , 0.63938711],\n", - " [1. , 0.28670672],\n", - " ...,\n", - " [4. , 0.09272909],\n", - " [1. , 1.1761186 ],\n", - " [1. , 0.98494351]])" + "\n", + "array([[ 3.75 , 269.57489704],\n", + " [ 3.75 , 147.77852694]])\n", + "Coordinates:\n", + " * chain (chain) int64 0 1\n", + " * accept_dim_0 (accept_dim_0) int64 0 1" ] }, "execution_count": 14, @@ -503,31 +1627,59 @@ } ], "source": [ - "trace.get_sampler_stats(\"accept\")" + "# Range of accept values\n", + "trace.sample_stats[\"accept\"].max(\"draw\") - trace.sample_stats[\"accept\"].min(\"draw\")" ] }, { "cell_type": "code", "execution_count": 15, "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAABoMAAAIwCAYAAABN6RLXAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAA9hAAAPYQGoP6dpAABzDklEQVR4nO3deXxcd3kv/s/RZlmyvMZxEmeFxGFJQhIghDVAWAoFLkuhtKzlRyml9HajFCjQ0lugvS30UmjZLpASoIXAZS1bA2UJJAGSsCSQfbcdZ/FueZGl8/vjzEhy7CReJI90zvv9es3rjGZGM1/raKw585nneYqyLMsAAAAAAABQS12dXgAAAAAAAADTRxgEAAAAAABQY8IgAAAAAACAGhMGAQAAAAAA1JgwCAAAAAAAoMaEQQAAAAAAADUmDAIAAAAAAKgxYRAAAAAAAECNCYMAAAAAAABqTBgEAAAAAABQY8IgABrlxBNPzIknntjpZQAAABx0jocAmksYBAAAAAAAUGPCIAAAAAAAgBoryrIsO70IAA6On//85/n617+eH/3oR1m9enU2bNiQ+fPn55RTTslLX/rSPOpRj9rj991www35t3/7t1x44YW57bbb0t3dncMOOyxnnHFGfvu3fzsrVqzY5fYbNmzIueeem29/+9u56aabMjIykqVLl+akk07Kc5/73Jx11lm73H7nzp35/Oc/ny996Uu56qqrMjw8nEMPPTSPfexj8+pXvzqHH374Lre/+OKL89KXvjRnnHFGPvShD+X9739/vva1r+W2227LggUL8rjHPS5/9Ed/lGXLlo1/z3vf+968733vu8efzbe+9a0ceeSR+/ojBQAAZgnHQ46HAJpMGATQIC9/+ctz8cUX5/jjj8/hhx+euXPn5pZbbskVV1yRJHnTm96Ul73sZbt8z5e//OW86U1vyo4dO3LEEUfk5JNPztjYWG655ZZcddVV+YM/+IP84R/+4fjtr7zyyrzqVa/KmjVrMjQ0lNNPPz2Dg4NZvXp1fvWrX+WUU07JueeeO377zZs35/d///fzox/9KAMDAznppJOyaNGiXH311bnhhhuycOHCfOxjH8uDHvSg8e9pH/ycdtppKcsyV111Vc4444zMnTs3l1xySe64444sXbo0n/jEJ3LssccmSc4///ycf/75+fznP58kec5znrPLv/P1r399Fi9ePKU/bwAAYOZwPOR4CKDJhEEADfLd7343D3zgA3PooYfucvlll12WV77yldm+fXu+9a1vjX+C7PLLL88LX/jC7Ny5M3/5l3+ZF73oRenqmugwunLlyqxbty4nnXRSkmR4eDhPf/rTs3r16jz72c/OW9/61gwODo7fftOmTfnFL36xyyfu/uzP/ixf+cpX8oQnPCFvf/vbs2TJkvHrzjnnnLzzne/Msccem69+9avp7u5OMnHwkyTHHHNMzjnnnBxxxBFJku3bt+fP//zP841vfCOnnnpqPv3pT+/yb20PS73qqqsO7IcJAADMKo6HHA8BNJmZQQANctZZZ+124JMkp512Wl70ohdlZGQk559//vjl73//+zMyMpIXv/jFeclLXrLLgU+SLF++fPzAJ0nOO++8rF69Og984APzjne8Y5cDnyQZGhra5cDnuuuuy3/+53/m0EMPzT/+4z/ucuCTVJ/cO+uss3LjjTfme9/73h7/Ta9//evHD3ySZM6cOfmrv/qrzJ07Nz/96U9z6aWX7sVPBgAAqDvHQwA0WU+nFwDAwbVu3bp897vfzdVXX52NGzdm586dSZIbb7wxSdUPO0lGR0fzwx/+MEnyghe8YK/u+/vf/36S5Dd+4zfGP7V2b7773e+mLMs87nGPy7x58/Z4mzPOOCPf/e53c9lll+UJT3jCLtfNnz8/Z5999m7fs2TJkjz2sY/NN7/5zfzoRz/K6aefvlfrBwAA6s3xEABNJQwCaJDPfOYzeec735nh4eF7vM2WLVuSJOvXrx+/3XHHHbdX979q1aokyf3ud7+9uv0tt9ySJPnsZz+bz372s/d627Vr1+522fLly1MUxR5v3x5+etttt+3VWgAAgHpzPARAkwmDABri8ssvz1vf+tZ0d3fnda97XZ74xCeOD00tiiKf/vSn89a3vjUHc5Tc2NhYkuSBD3xgHvCAB9zrbR/ykIfs12MYjQcAADgeAqDphEEADfH1r389ZVnmxS9+cX73d393t+vbbRHaFi5cmLlz52br1q254YYbsmLFivt8jMMPPzzXXXddrr/++l16Yd/b7ZPk9NNPz1vf+ta9+4dMsnLlyvu87rDDDtvn+wUAAOrF8RAATdd13zcBoA42bNiQJLsMF23bvn17vvnNb+5yWXd39/gBzGc+85m9eozHPvaxSZLPfe5zGR0dvc/bP+5xj0uSfPvb38727dv36jEm27hxY7797W/vdvnatWvH+3WfccYZu1zX29ubJOO9wQEAgPpzPFRxPATQXMIggIa4//3vnyT5whe+kM2bN49fvn379vz1X/91br311t2+59WvfnV6enryyU9+Mp/85Cd3azGwcuXKXH755eNfP//5z89hhx2WX/7yl3nzm9+8Wy/uzZs3jw9hTZIHPehBeepTn5rVq1fnta997R7XMDw8nC996Uu588479/jv+vu///td+mDv2LEjb3vb2zI8PJxTTjklD33oQ3e5/bJly5Ik11577R7vDwAAqB/HQxXHQwDNVZSahwI0wsaNG/PsZz87K1euzMKFC/Owhz0s3d3d+clPfpJt27blec97Xj7+8Y/nOc95Tv7u7/5u/Pu+8IUv5M1vfnNGRkayfPnynHTSSSnLMrfcckuuvPLK/MEf/EH+8A//cPz2v/zlL/OqV70qd9xxR+bPn5/TTz89AwMDWb16dX71q1/llFNOybnnnjt++82bN+e1r31tLrzwwvT29uYBD3hAjjzyyJRlmZUrV+bKK6/MyMhIvvrVr44fwF188cV56UtfmtNOOy1jY2O5+uqrc+aZZ6a/vz+XXHJJbr/99ixZsiSf+MQndhve+vd///f56Ec/mkWLFuXMM8/M4OBgkuR1r3tdFi1aNJ27AAAA6BDHQxXHQwDNJQwCaJC1a9fmve99by644ILcdtttWbhwYc4888y89rWvzSWXXJI3vvGNux38JNWnxj72sY/loosuyu233545c+Zk2bJlOfPMM/Nbv/VbOf7443d7nI9//OP51re+lVtvvTVjY2NZunRpTj755Dz3uc8db5/QNjY2lq9+9av50pe+lCuuuCIbNmzI4OBgDj300Jx00kk5++yzc9ZZZ423NGgf/Jxxxhn54Ac/mPe97335xje+kTVr1mTBggV53OMel//5P//neA/uybZv355//ud/zn/9139l1apVGRkZSZJ861vfypFHHjmVP24AAGAGcTzkeAigyYRBAMw6kw9+Jn+qDgAAoO4cDwGwP8wMAgAAAAAAqDFhEAAAAAAAQI0JgwAAAAAAAGrMzCAAAAAAAIAaUxkEAAAAAABQY8IgAAAAAACAGuvp9AIAaI6vfe1r+dSnPpUrr7wyIyMjOfroo/PMZz4zL3/5y9Pb29vp5QEAAEy566+/Pj/4wQ9yxRVX5Iorrsh1112X0dHR/NEf/VFe85rXdHp5ADSEMAiAg+Ltb397Pv7xj6enpydnnnlmBgYGctFFF+Uf//Ef89///d/56Ec/mv7+/k4vEwAAYEr9+7//ez7+8Y93ehkANJwwCIBpd/755+fjH/94BgYG8olPfCIPfvCDkyRr167Ny172slxyySV5z3vek7/4i7/o8EoBAACm1ooVK/KKV7wiD3rQg/KgBz0oH/zgB/PFL36x08sCoGGEQQBMuw984ANJkle96lXjQVCSLF68OH/1V3+VF73oRfnEJz6R17zmNRkaGurUMgEAAKbc85///F2+7uoywhuAg89fHwCm1Zo1a/KLX/wiSfKMZzxjt+sf9rCH5fDDD8+OHTvy3e9+92AvDwAAAABqTxgEwLT65S9/mSRZuHBhjjrqqD3e5qSTTtrltgAAAADA1BEGATCtbr311iTJ4Ycffo+3Oeyww3a5LQAAAAAwdYRBAEyrLVu2JEnmzp17j7cZHBzc5bYAAAAAwNQRBgEAAAAAANSYMAiAadWu+tm6des93qZdEdS+LQAAAAAwdYRBAEyr5cuXJ0lWr159j7e57bbbdrktAAAAADB1hEEATKsHPehBSZL169fnlltu2eNtLr/88iTJgx/84IO2LgAAAABoCmEQANPqsMMOy8knn5wk+cpXvrLb9T/5yU+yevXq9PX15ayzzjrYywMAAACA2hMGATDtXv3qVydJPvShD+WKK64Yv3zdunV529veliR58YtfnKGhoY6sDwAAAADqrCjLsuz0IgCov7/927/Nueeem97e3px55pkZGBjIhRdemI0bN+b000/Pxz72sfT393d6mQAAAFPqiiuuGP8QXJLcfPPNWbduXQ477LAsW7Zs/PL3ve99OfTQQzuxRAAaoKfTCwCgGd785jfn9NNPz6c+9alcdtll2blzZ44++uj87u/+bl7+8penr6+v00sEAACYcps3b87Pfvaz3S6/7bbbctttt41/vWPHjoO5LAAaRmUQAAAAAABAjZkZBAAAAAAAUGPCIAAAAAAAgBoTBgEAAAAAANSYMAgAAAAAAKDGevb2huvWrZvOdTTSggULsmHDhk4vg9gXM4l9MTPYDzOHfTEz2A8zh30xM8y0/bBo0aJpf4yZdjw00/YB3BO/q8wGfk+ZLfyuMhv4Pe2MvTkmUhnUQV1dfvwzhX0xc9gXM4P9MHPYFzOD/TBz2Bczg/3QefYBs4XfVWYDv6fMFn5XmQ38ns5c9gwAAAAAAECNCYMAAAAAAABqTBgEAAAAAABQY8IgAAAAAACAGhMGAQAAAAAA1JgwCAAAAAAAoMaEQQAAAAAAADUmDAIAAAAAAKgxYRAAAAAAAECNCYMAAAAAAABqTBgEAAAAAABQY8IgAAAAAACAGhMGAQAAAAAA1JgwCAAAAAAAoMaEQQAAAAAAADUmDAIAAAAAAKgxYRAAAAAAAECNzdowqCzLlGXZ6WUAAAAcdI6HAACAfTErw6CxsbE89/llfvslDoAAAIBmKcsyz/gf6/Pc55fZsWOs08sBAABmgVkZBq1aldxxZ3LLrcnwsDAIAABojm3bkhtvGssddyaf+WynVwMAAMwGszIMWrBg4vwdd3RuHQAAAJ30iU8lGzf6gBwAAHDvZmUY1NNTjJ+/6ZYOLgQAAKCDNm9O/v0zwiAAAODezcowaLJbhEEAAECD/fSnnV4BAAAw0836MOimmzu9AgAAgM659tpkdFR1EAAAcM9mfRikMggAAGiqvr5k67Zk5apOrwQAAJjJZn0YdLPKIAAAoKGOO67aXnNtZ9cBAADMbLM+DNq4KVm/XksEAACgeY6/X7W99lrHRAAAwD2b9WFQYm4QAADQTPe/f7VVGQQAANybWoRBq/THBgAAGuj+rcqga67p7DoAAICZrRZh0OYtnV4BAADAwXfccUlRJHet1T4bAAC4Z7UIg7YIgwAAgAaa219kyZLq/G1rOrsWAABg5qpFGDQ87BNwAABAMy07tNquub2z6wAAAGauWoRBKoMAAICmWras2q5RGQQAANyDWoRBZgYBAABNNVEZpGMCAACwZ7UIg4aHO70CAACAzjj00CKJyiAAAOCe1SIM0iYOAABoKjODAACA+1KPMEhlEAAA0FDtmUG3qwwCAADuQT3CoM2dXgEAAEBntCuD7lqb7NhhbhAAALC7eoRBKoMAAICGWrAgmTOnOn/HnZ1dCwAAMDPVIgwa3pKUpU/AAQAAzVMUxcTcIK3iAACAPahFGDQ6lmzf3ulVAAAAdEZ7btCa2zu7DgAAYGaa9WFQUVTbzVs6uw4AAIBOURkEAADcm1kfBg0MVNthYRAAANBQhxxSbe9aq302AACwu1kfBg0OVtstw51dBwAAQKcsWli1TFi3rsMLAQAAZqRZHwYNzK22W1QGAQAADbVocbUVBgEAAHsy68Og8cogYRAAANBQixZWW2EQAACwJ7M+DJqrMggAAGi4RYuq7br1HV0GAAAwQ836MGiemUEAAEDDLW6FQZs2JSMjZWcXAwAAzDizPgyaO1BtVQYBAABNNW9e0t1dnV+/vqNLAQAAZqBZHwYNjodBPv0GAAA0U1dXkYULq/NaxQEAAHc368OggXYYpE0cAADQYO1WcWvXdnYdAADAzDPrw6DB9swgbeIAAIAGW9QKg1QGAQAAdzfrw6B2ZdCwyiAAAKDBFi2stuvWdXQZAADADDTrw6CJmUGdXQcAAEAnjVcGrTNPFQAA2NWsD4PmmhkEAACQRYuKJNrEAQAAu5v1YdCcvmq7fXtn1wEAANBJ2sQBAAD3ZPaHQXOq7fZtnV0HAABAJy1aXG2FQQAAwN3N+jCorx0G7ejsOgAAADpJZRAAAHBPZn0Y1K9NHAAAQBYtqrbr1idlWXZ0LQAAwMwy68Og8TZx2x3wAAAAzbVgfrXduTPZurWzawEAAGaW2oRBSbJDqzgAAKCh+vuL9LU6J2zc2Nm1AAAAM8usD4PaBzuJVnEAAECzzW9VB20QBgEAAJPM+jCop6dId3d1XhgEAAA0WbtVnMogAABgslkfBiUTreK2CYMAAIAGGxqqtsIgAABgslqEQf2tMEhlEAAA0GTtNnEbN3V2HQAAwMxSizBojjAIAABgvE3chg2dXQcAADCzCIMAAABqYqgVBm3aVHZ2IQAAwIxSizCoTxgEAACQBfOLJMkGM4MAAIBJahEGmRkEAACQzB+qthuFQQAAwCS1CIPabeK2CYMAAIAGm7+g2gqDAACAyWoVBqkMAgAAmkxlEAAAsCfCIAAAgJpYML/aCoMAAIDJhEEAAAA1Mb8dBm1KxsbKzi4GAACYMWoWBjnYAQAAmmuo1SZubCzZsqWzawEAAGaOWoRB/SqDAAAAMmdOkf7+6vzGTZ1dCwAAMHPUIgzSJg4AAKAy39wgAADgbmoSBhVJkm3CIAAAoOHmt1rFCYMAAIC2moRB1VZlEAAA0HTjlUHaxAEAAC21CoN2CIMAAICGGw+DNnR2HQAAwMxRqzBImzgAAKDphuZV202bO7sOAABg5qhFGNSvTRwAAECSZKg1M2jz5rKzCwEAAGaMWoRBZgYBAABU5s0rkqgMAgAAJgiDAAAAamReq03cZmEQAADQIgwCAACokYk2cZ1dBwAAMHMIgwAAAGqkXRm0aVNn1wEAAMwctQqDtgmDAACAhhtqh0EqgwAAgJZahEH9KoMAAACSmBkEAADsrhZhULsyaHQ02bmz7OxiAAAAOmhoUhg0Nub4CAAAqFkYlCQ7dnRuHQAAAJ3Wrgwqy2R4uLNrAQAAZoZahEF9fRPnt23r3DoAAAA6bc6cIn291Xmt4gAAgKQmYVBRFOPVQeYGAQAATTc0VG03bersOgAAgJmhFmFQMtEqbrs2cQAAQMO1W8VtUhkEAACkTmFQq1WcNnEAAEDTtcMgbeIAAICkTmFQf7XVJg4AAGi68TZxwiAAACB1CoNalUE7tIkDAAAaTmUQAAAwWW3CoD5hEAAAQJLJYVDZ2YUAAAAzQu3CoJGRzq4DAACg04ZaYdCmTZ1dBwAAMDPUJgzq7a22KoMAAICmGxoqkpgZBAAAVGoTBmkTBwAAUDEzCAAAmKw+YVC7MkibOAAAoOHmaRMHAABMUp8wSGUQAABAkomZQSqDAACApEZhUG8rDBpRGQQAADScNnEAAMBktQmDJiqDys4uBAAAoMMGB6vt5i2dXQcAADAz1CcMas8M0iYOAABouHmtMGh4OBkd9YE5AABouvqEQWYGAQAAJJmoDEqSrVs7tw4AAGBmqE0Y1NtbJDEzCAAAYM6cIr2t7glaxQEAALUJg1QGAQAATBgcqLZbhEEAANB49QuDVAYBAACMt4rbvLmz6wAAADqvPmFQqwWCyiAAAIBkcF61VRkEAADUJwzSJg4AAGDcvHZlkDAIAAAarzZhUG8rDBrRJg4AAGC8TdywMAgAABqvNmGQyiAAAIAJKoMAAIC2+oRBZgYBAACMa1cGbdlSdnYhAABAx9UnDFIZBAAAMG5QZRAAANBSmzCot1UZZGYQAABAMjhYJEm2bO7wQgAAgI6rTRikMggAAGDCvHnVVmUQAABQnzCoPTNIZRAAAMB4m7jh4c6uAwAA6Lz6hEEqgwAAAMbNa88M0iYOAAAar3ZhkJlBAAAAE5VBW7SJAwCAxqtNGNQ7KQwaGys7uxgAAIAOa4dBZgYBAAC1CYPm9E2cVx0EAAA0XbtN3PCWpCx9YA4AAJqsNmFQb+/EeXODAACApmtXBo2OJVu3dnYtAABAZ9UmDOrpSYqiOi8MAgAAmm7u3KS7dcRnbhAAADRbbcKgoijGq4O0iQMAAJquKIoMtKqDtgx3di0AAEBn1SYMSpK+1twglUEAAAATc4M2b+7sOgAAgM6qVxjUqgzaoTIIAABgojJImzgAAGi0eoVBKoMAAADGjVcGCYMAAKDRahUG9bbCIDODAAAAkkGVQQAAQGoWBqkMAgAAmGBmEAAAkAiDAAAAamtwXrUdHi47uxAAAKCj6hUG9VZbYRAAAICZQQAAQKVeYZDKIAAAgHEDA0WSZIs2cQAA0Gi1CoN6W5VBIyOdXQcAAMBMMK/VJk5lEAAANFutwqB2ZdB2lUEAAAAZbLWJ2yIMAgCARqtXGKQyCAAAYJyZQQAAQFK3MMjMIAAAgHHtyqBhYRAAADRarcKg3lYYNDJSdnYhAAAAM4CZQQAAQFKzMMjMIAAAgAmDA9XWzCAAAGi2eoVB7ZlBwiAAAIDxNnEjI8n27TooAABAU9UrDOorkpgZBAAAkCQDAxPnVQcBAEBz1SoM6m1XBo10dh0AAAAzQXd3MR4ImRsEAADNVaswyMwgAACAXc1rtYobFgYBAEBj1TIM0iYOAACg0p4bpDIIAACaq15hkDZxAAAAu2iHQWYGAQBAc9UqDOptVQYJgwAAACoqgwAAgFqFQe3KIG3iAAAAKu2ZQVs2d3YdAABA59QqDOrVJg4AAGAXg/OqrcogAABoLmEQAABAjY1XBg2XnV0IAADQMbUKg/paM4N2CIMAAACSJIODRRJt4gAAoMlqFQaNVwaZGQQAAJBkojJImzgAAGiueoZBKoMAAACSJAPtNnHCIAAAaKxahUF9rTBImzgAAICKyiAAAKBWYVBva2aQyiAAAIDKYCsMGhYGAQBAY9UqDOozMwgAAGAX8+ZVW5VBAADQXLUKg9ozg0bHktHRsrOLAQAAmAEGB6qtmUEAANBctQyDEq3iAAAAkok2cdu2JTt3+tAcAAA0Ua3CoL6+ifM7hEEAAADjYVCSDA93bh0AAEDn1CoM6u5OiqI6b24QAABA0ttbjH9wztwgAABoplqFQUVRjLeK0yYOAACgMq9VHWRuEAAANFOtwqAk6WuFQdrEAQAAVAbnVVthEAAANFPtwqDeVvsDlUEAAACVwYFqq00cAAA0U/3CoHabODODAAAAkiSDrTZxw8IgAABopNqFQdrEAQAA7Gpeq02cyiAAAGim2oVB45VBwiAAAIAkE23izAwCAIBmql8Y1JoZpDIIAACg0m4Tt2VL2dmFAAAAHVG7MKjPzCAAAIBdaBMHAADNVrswSJs4AACAXQ0MFEm0iQMAgKaqbRikTRwAAEBlXrtN3ObOrgMAAOiM2oVBfe2ZQdrEAQAAJEkGW23itgx3dh0AAEBn1C4M0iYOAABgV4MD1dbMIAAAaKbahUF97TZxKoMAAACSJIPaxAEAQKPVLgxSGQQAALCredrEAQBAo9UvDGrNDBoZKTu7EAAAgBmi3SZueEtSlo6VAACgaWoXBo23iVMZBAAAkGSiTdzoWLJ1a2fXAgAAHHy1C4PG28SZGQQAAJAkmTs36Wod/Q1rFQcAAI1T3zBIZRAAAECSpCiKDLRaxW3e0tm1AAAAB1/twqC+viKJNnEAAACTzWu1itu8ubPrAAAADr7ahUEqgwAAAHY3OK/aahMHAADNU98wyMwgAACAceOVQdrEAQBA4/R0egFTra8VBmkTx9664cYyX/3a1oyOlnn8WcnQUNHpJQEAwJQbbM0M2qJNHAAANE7twqDevmqrTRx76/VvKLP6tqpXxo03J3/4GmEQAAD1024Tt0WbOAAAaJz6tokTBrEXhofLrL5t4uvrr+/cWgAAYDoNttrEbdlSdnYhAADAQVe7MGi8TZyZQeyFVat3/Xr16j3fDgAAZrt2m7jN2sQBAEDj1C4MUhnEvmiHP4sXVa3hbluTjI76pCQAAPUzb171mlebOAAAaJ7ahUF9ZgaxD9ph0Omn9aa7O9m5M7nzrs6uCQAApsNEm7jOrgMAADj4ahcGtSuDdgiD2AurbquqgI46uivLllWXaRUHAEAdaRMHAADNVdswaMTMIPZCO/g5cnl3Dj9s18sAAKBOBudV22Ft4gAAoHFqFwb1mRnEPmgHP8uXd+WIw6vzq1abGQQAQP3Ma7WJ26xNHAAANE7twqDe1swgbeK4L2VZ7loZdHg1UHf1bR1cFAAATJN2m7gt2sQBAEDj1C4M6muFQSqDuC/rNyRbt1Xnjzi8K4e3KoO0iQMAoI7abeK2aBMHAACNU78waFKbuLLU7ot71g59DjkkmTOnMDMIAIBaa7eJ27492bnTsRIAADRJ7cKg3t6J86qDuDftdnDtEKg9M+iOO5OREQfHAADUy8DAxPnNWsUBAECjCINorPXrq+2SxdV24cKkKJKyTDZu7NSqAABgevT0FOnvr85v2dLZtQAAAAdXrcOgHcIg7sWmTdV2aH617eoqMjRUnd+4qTNrAgCA6TTYahVnbhAAADRL7cKgrq4iPT3V+ZEdnV0LM9umTVUruPlDE5cNzWtf14EFAQDANBtstYrTJg4AAJqldmFQMlEdpE0c96bdCm5oqBi/bH6rSkhlEAAAdTTY+vCTNnEAANAstQyD+lphkDZx3Jt24NMOgJKMt4nbZGYQAAA1NE+bOAAAaKRahkG9fdVWZRD3Znxm0LyJy8bDIJVBAADUULtN3BZt4gAAoFHqGQa1K4PMDOJe7KkyaKJNXHnwFwQAANOs3SZuszZxAADQKLUMg/rMDGIvjFcGDU1c1q4SUhkEAEAdTbSJ8+EnAABoklqGQb3CIPZCey7Q/Elh0Pz5RZKJqiEAAKiTAW3iAACgkeoZBrVmBu0QBnEPtm8vx38/JreJMzMIAIA6G5pXffhpszAIAAAapZZh0HibODODuAcbW1VB3V3J3LkTl7erhNrXAwBAncwzMwgAABqplmGQNnHcl3YbuKH5SVEU45e3K4O0iQMAoI7GwyCVQQAA0Ci1DIPalUHaxHFP2m3ghoZ2vbxdGbRZGAQAQA0JgwAAoJlqGQa1ZwapDOKetNvAzb9bGDR5ZlBZlgd3UQAAMM2GhEEAANBI9QyDzAziPmxqHfzOn7/r5e2vR8eS4eGDuyYAAJhu7cqgTcIgAABolFqGQdrEcV/alUHtT0a2zZlTjP/+tG8DAAB10a6E37Ej2b5dJTwAADRFLcOg8cogYRD3YOOm6sD37pVBSTI0v32bg7ggAAA4CAYGkqKozmsVBwAAzVHPMKg1M2jHiE+6sWebWkHP0NDu180f2vU2AABQF11dRQYHq/PCIAAAaI5ahkF9ZgZxHza128QNFbtd1w6IVAYBAFBHQ+YGAQBA49QyDNImjvvSPvDdU5u49mUqgwAAqKN5rTBIZRAAADRHTcOgqtpjhzCIe7CxXRk0b/fr2pe1bwMAAHUiDAIAgOapZRjU15oZpDKIe7JluNrO20MYNNi6bHjYzCkAAOqn3RZZmzgAAGiOWoZBvWYGcR+GW2HQwNzdr2tf1r4NAADUicogAABonlqHQdrEcU+Gt1TbgYHdrxsYqNoMbhEGAQBQQ+22yJs2q4QHAICmqGUY1NeuDBIGsQdjY2W2bqvO7zkMqrYqgwAAqKN586oPP6kMAgCA5qhlGNRrZhD3YuvWifPCIAAAmkabOAAAaJ5ahkHtyqAdZgaxB+2Qp7sr6evb/XozgwAAqDNhEAAANE8tw6BeYRD3oh3yzB1IiqLY7frxyqCtu10FAACz3tBQtd0kDAIAgMaodRikTRx70g559tQibvLlKoMAAKijoVZl0KZNnV0HAABw8NQyDGq3/tohDGIP2iHPPYVBg8IgAABqTJs4AABonlqGQSqDuDfjYdDcPV/fDom2DidlWR6cRQEAwEEyOQzyehcAAJqh3mGQmUHswX1VBrUvHx0zdwoAgPppt4nbuTPZvr2zawEAAA6OWoZBfa0wSJs49uS+ZgbNnVQxpFUcAAB1M3du0t06EtQqDgAAmqGWYVBva2aQNnHsyX1VBnV1FZnbv+ttAQCgLoqiGG8Vt0kYBAAAjVDLMKjPzCDuxfBw1Rf9nmYGJRNBkTAIAIA6mjw3CAAAqL9ahkHtmUFjY8nOnQaisqt2wDM4eM+3mdsOg7ZO/3oAAOBgG68M2tTZdQAAAAdHrcOgRHUQu9sy3iauuMfbtCuDtmw5CAsCAICDTGUQAAA0Sy3DoL6+ifPCIO5ufGbQvbWJm7vrbQEAoE6GzAwCAIBGqWUY1N2dFK2ijx3CIO5mPAwauOfbDGgTBwBAjakMAgCAZqllGFQUxXiruJEdnV0LM8/WVsCzV2GQyiAAAGpoaKjabt5sxioAADRBLcOgJOlrh0Eqg7ibvakMGhQGAQBQY/PmVa0UVAYBAEAz1DYM6m3NDdImjrtrBzxz721m0HgY5JOSAADUT7tN3KZNnV0HAABwcNQ3DFIZxD3Yu5lB1SclzQwCAKCOhtphkMogAABohNqGQe02cTvMDOJuhs0MAgCg4dqVQdrEAQBAM9Q2DFIZxJ6MjJTjvxP3Gga1WsgJgwAAqCNhEAAANEt9w6DWzCBhEJNNDncG9mpm0PSuBwAAOmFoqNoKgwAAoBlqGwaNt4kTBjFJO9zp60t6eop7vN14GGRmEAAANTQ0qTKoLMvOLgYAAJh29Q2D2pVBZgYxyd7MC5p8vcogAADqqN0mbnQs2eoDUAAAUHu1DYPMDGJP2uHOfYZBrRZyW4VBAADU0Jw5SU9PdX6TVnEAAFB7tQ2DtIljT9ph0Nx7mReUqAwCAKDeiqIYrw4yNwgAAOqvtmFQb7tNnDCISbZtq7YD9xEG9fe3br9dD3UAAOpJGAQAAM1R3zCoXRlkZhCTbG2FQfdVGdS+viyT7dund00AANAJQ0PVVps4AACov9qGQX1mBrEH7eG47cqfezL5egN1AQCooyGVQQAA0Bi1DYN6x8MgLb6Y0A527qsyqKuryJw5re/ZNr1rAgCATtAmDgAAmqO+YVBrZtAOlUFMsnVrFQ7OvY/KoGQiMFIZBABAHQmDAACgOWobBo23iTMziEm2tap87qtNXDIRGAmDAACoo3abuI0bdVMAAIC6q20Y1NtbJEl2CIOYZG/bxE2+zTZt4gAAqKH586tjpo2bOrwQAABg2tU2DOprtYkb0SaOSdrzf+bOLe7ztv0qgwAAqLH5Q9V248bOrgMAAJh+tQ+DzAxism3jYdB933Z8ZpDKIAAAamj+gmorDAIAgPqrbxjUmhmkTRyTtat89mpmkDZxAADUmMogAABojvqGQe3KIGEQk4zPDNqLMEibOAAA6mz+/GprZhAAANSfMIhGGQ+D9qVNnDAIAIAaWtAKgzZtTMqy7OxiAACAaVX7MGjEzCAmabd826s2ceOVQQ6MAQCon6FWm7jRsWTLls6uBQAAmF61DYN6W2HQdpVBTLK1FQYN7EVlUDswMjMIAIA6mjOnyJw51Xmt4gAAoN5qGwbN0SaOPWi3fOvfqzZxRfU9wiAAAGqq3Spu44bOrgMAAJhetQ2DzAzi7sqyHK/yMTMIAACSoXYYpDIIAABqrfZh0IgwiJYdO5Kxser83L2YGaRNHAAAdTe/NTdo48bOrgMAAJhetQ2Denurrcog2iZX+PTvRRikMggAgLprt4nbIAwCAIBaq20YpE0cd9eu8OnrS7q7i/u8fbt6SBgEAEBdtdvEbdImDgAAaq22YdAcYRB3M9wKdfamRVwyURmkTRwAAHU10Sau7OxCAACAaVXbMKhdGTQ6luzc6cCGiVCnHfLcF5VBAADU3fz5VcW8NnEAAFBvtQ2D2jODkmRkpHPrYOZohzr9exkGtW+3VWUQAAA11Z4ZtEkYBAAAtdaIMEirOJKJUEebOAAAqLRnBm00MwgAAGqttmFQT0+R7u7qvDCIZKIyaF/bxI2MaDUIAEA9tWcGaRMHAAD1VtswKJmYG7RDmziSbGu3idvHyqDE3CAAAOpJmzgAAGiGeodBrVZxKoNIJrWJ28vKoN7eieoyYRAAAHU0f1KbuLEx1fAAAFBX9Q6D2pVBwiAyqU3cXlYGTb6tuUEAANTRggXVdmws2by5s2sBAACmjzCIxti2rfqk495WBk2+7VZhEAAANdTbW2RwsDq/fn1HlwIAAEyjRoRBI2YGkYnKoP59CIPat9UmDgCAulrYqg5av6Gz6wAAAKZPrcOg3lYYtF1lEJk0M6i/2OvvabeJUxkEAEBdtVvFqQwCAID6qnUYNEebOCbZ1p4ZtD9t4lQGAQBQUwsXVtsNKoMAAKC2ah0GjbeJEwaRiUCnXe2zN/pbt90mDAIAoKa0iQMAgPqrdRjU21ttVQaRTGoTtz+VQdrEAQBQU+3KoPUbyo6uAwAAmD61DoP6tIljknZlUP8+VAaNzwxSGQQAQE0tWFDN1DQzCAAA6ksYRGNs24/KoP657e/1KUkAAOrJzCAAAKi/ZoRBI51dBzPDAbWJUxkEAEBNLWjNDBIGAQBAfdU7DDIziEn2r01c1TLDzCAAAOpqYSsM0iYOAADqq95h0HibOC2+SLa1wqC5+xIGtdvEqQwCAKCmxsMglUEAAFBbDQmDOrsOOm/nznK8XeA+zQxqBUfaxAEAUFftmUFbtybbt/sgHQAA1FHNw6CqxZeZQWyb1OZt39rEVVtt4gAAqKvBwaSnpzpvbhAAANRTrcOgXjODaGmHQd1dExVje2O8TZwwCACAmiqKIgu0igMAgFqrdRg0R5s4WoZbbd7651YHu3urHQZpEwcAQJ215wapDAIAgHqqdRhkZhBt7cqefWkRN/n22sQBAFBn7blB69Z3chUAAMB0aUQYNGJmUOO1K3valT57S2UQAABNMN4mbn1HlwEAAEyTWodBvSqDaBkPg/axMqh9+23CIAAAamzxomq7bl3Z2YUAAADTotZhkDZxtLXbxO1vZdC27cnYmANjAADqafHiaq7mXWs7vBAAAGBaNCIM2i4Marx2ZdD+zgwqy2T79qldEwAAzBQTlUGdXQcAADA9ah0GzWnPDBIGNd7W/awMmhwetauLAACgbha1wqC1KoMAAKCWah0G9fZWW23i2N82cV1dxXggNGxuEAAANbV4cbVVGQQAAPVU6zDIzCDatm6t5v3sa5u4yd+jMggAgLpqh0Fr1yVlaVYmAADUjTCIRmjPDBrYx8qgZKKaaKvKIAAAamrRwmq7c2eyaVNHlwIAAEyDZoRBI51dB53XDnL6+4t9/t65/bveBwAA1E1fX5F586rza7WKAwCA2ql3GGRmEC3tFm/axAEAwJ4tXlRt167t7DoAAICpV+8wqFUZNDqa7Nyp73WTbW0FOQfUJk4YBABAjY3PDRIGAQBA7TQiDEpUBzXdeJs4M4MAAGCPxsMgbeIAAKB2ah0GzZkzcX67MKjR2i3e5moTBwAAe9RuE7duna4KAABQN7UOg7q6ivS25gZt397ZtdBZw62qnrkqgwAAYI8WLSqSJHdpEwcAALVT6zAoSea0WsXtEAY12rZ2m7j9qAxqVxNt3eoTkgAA1NdEZVBn1wEAAEy9+odBrVZxKoOabbxN3AFUBmkTBwBAnY3PDFIZBAAAtVP7MKivHQaZGdRo423i9mtmUNUuQ5s4AADqbEkrDNImDgAA6qf2YZDKIMqyHK/q6T+QmUEqgwAAqLFDDqm2a9cmO3dqkQwAAHVS/zCoNTNIGNRcO3YkZetYduBAwiCVQQAA1NiiRUl3VzI2lqxf3+nVAAAAU6n+YZDKoMYbnhTitH8f9kV/q7WcmUEAANRZd3cxPjfojjs7uxYAAGBqNScMMjOosba1wqA5c6oD3H01oDIIAICGOGRptb3jjs6uAwAAmFrNCYNUBjVWO8SZ279/39+uDBIGAQBQd4csqbZ33tXZdQAAAFOr/mGQmUGNt7XV3q1/P8OgudrEAQDQEEsPqbZ33Fl2diEAAMCUqn8YpDKo8dohzty5+/f97e/bKgwCAKDmli6t2irfqU0cAADUSu3DoL5WGLRjh0+2NVW7vVv/foZB/WYGAQDQENrEAQBAPdU+DFIZRLuiZ39nBrW/b+fOZGREqAgAQH0dMt4mrrPrAAAAplb9wyAzgxqvXdFzoG3iEq3iAACot6VLq602cQAAUC/1D4PmVD2vhUHNta3dJm4/K4N6e4t0d+96XwAAUEdLW5VBm7ckW7eqigcAgLpoQBhUbbfv6Ow66Jx2Nc/AflYGJRPVQdtUBgEAUGMDAxNtks0NAgCA+mhOGKQyqLHan2jc38qgZOKAeFhlEAAANVYURQ5ptYq7Q6s4AACojfqHQWYGNV67mqf/ACqD+lUGAQDQEIe2wqDbhUEAAFAb9Q+DVAY1XruaZ2Busd/30W4Tt1VlEAAANbdsWbW97bbOrgMAAJg69Q+DWu29zAxqrvHKoCloE7dVZRAAADV32LLqQ1S33VZ2eCUAAMBUqX8Y1GoTt0NlUGNta1XzzD2ANnHt792mMggAgJo7/LBqe9uazq4DAACYOvUPg7SJa7x2m7i5B1AZ1K4q0iYOAIC6G28TJwwCAIDaEAZRe9rEAQDA3jusVRl0+5pkbEyrOAAAqIPmhEFmBjVWOwyakjZx2xwMAwBQb4cuTbq6kh0jybp1nV4NAAAwFeofBrVmBqkMaq7hKZgZ1N/6Xm3iAACou56eIocsqc5rFQcAAPVQ/zBoUpu4slTV0URT0yauSKJNHAAAzTA+N+i2zq4DAACYGo0Jg5Jkh1ZxjbR1CiqD5qoMAgCgQdpzg1QGAQBAPTQqDDI3qHl27iwzMlKdn3sAlUHtqqJtwiAAABpgIgzSXQEAAOqg9mFQT0+R7ta/coe5QY2zbVJbtwOpDBpoVwZpEwcAQAMctqxqk6xNHAAA1EPtw6Ak6Zs0N4hmabd16+5Kenv3/37alUHaxAEA0ARHHF5tV63q7DoAAICp0YgwaI4wqLHalTz9c5OiKPb7ftpVRdtUBgEA0ABHLq+2q1Yno6NaxQEAwGzXqDBomzCocdqVPAfSIm7y9w+rDAIAoAEOPbSqrB8ZSW6/vdOrAQAADlQzwqC+aqsyqHnalTztNm/7q/39KoMAAGiC7u5ivFXcrSs7uxYAAODANSMM0iausdqVQQNTVBm0TWUQAAANceSR1faWWzu7DgAA4MA1Kwza0dl1cPC1w6ADrQya2/r+rduSsTE90wEAqL92GLRypde/AAAw2zUrDFIZ1DhT3SYu8XsEAEAzHLm8SKIyCAAA6qAZYVBrZtAOb+I3zvAUtYmbHAZt1SoOAIAGOHJ5tV1pZhAAAMx6jQiD+lQGNdZ4ZdABhkFdXcV4ILR124HdFwAAzAZHtdrErVqdjI5qFQcAALNZI8IgM4Oaa9u26qB17gGGQZPvY5swCACABli6NOntTUZGkjW3d3o1AADAgWhEGNTfCoO8id887TZxcw9wZlAy0SpOmzgAAJqgu7sYrw666abOrgUAADgwjQiDJio6tDZomm2t4Ka/vzjg+2rPHRIGAQDQFMcdW22vv6GjywAAAA5Qo8KgYW/iN86W4Wo7OHjg9zUwUG2Hhw/8vgAAYDY47rjqQ1U33OiDdQAAMJs1IgxqV4VoE9c87eBmYApnBgmDAABoiuOOq7YqgwAAYHZrSBhUbbepDGqc8TBo4MDvS2UQAABN024Td9NNyeio6iAAAJitGhEGjc96URnUOO3WgFMRBrVbzW0RBgEA0BDLj0j6epPt25PVqzu9GgAAYH81IgxqVwZtVRnUOMNbqu2UVAaNt4nziUgAAJqhu7vIMcdU52+4saNLAQAADkAzwqDWm/hmBjWPNnEAAHBg7mduEAAAzHqNCIPmtiuDhEGNM5Vt4gYGiiTaxAEA0CzHHVe9Dr7uehXyAAAwWzUjDGpXBmkT1yg7d5bj1WBTMjNIZRAAAA104opqe9VVnV0HAACw/5oRBpkZ1EiT9/egNnEAALBf2mHQylXJxk2qgwAAYDZqRBjUr01cI7VbxPX0JH19xQHfnzAIAIAmmj+/yBFHVOevvrqzawEAAPZPM8KgVpu40dFkZMQn2ZqiHdpMRYu4yfcjDAIAoGna1UFXahUHAACzUiPCoHabuER1UJNs2VJtp6JFXCIMAgCguR5wYlVpf9XVPlwHAACzUSPCoN7eIj091Xlzg5qjva+nujJoizAIAICGecCJ1VZlEAAAzE6NCIOSiblB24RBjTHVbeLaFUZbtyZl6RORAAA0x4oTqu3q1cm69V4LAwDAbNOYMKjdKk6buOZot4mbO3dq7q8dKpWlCjMAoLnKsvTBmAYaGipyv+Oq8z//RWfXAgAA7LvmhEGtQGCbMKgxhluBzeDg1Nxff3/S1bXrfQMANMmtt47lWc8t84Lf2iAQaqBTTqm2P/uZfQ8AALNNY8Kg/lYYpKKjOaa6TVxRFBlo/R4Nb5ma+wQAmE0++m/JunXJL3816kNWDXTqKUWS5Gc/7/BCAACAfdaYMEibuOYZHq4+sTgwRW3ikolgqR00AQA0RVmW+enPOr0KOukhrcqga65NtmxRHQQAALNJY8Kg/lYYtE1lUGNsaQU2U9UmLpkUBvk9AgAaZtWq5PbbO70KOmnp0iLLj0jGxpJfXN7p1QAAAPuiMWFQe2aQyqDmmGgTV0zZfbbDoC3axAEADXPJZZ1eATPBQx5SbS/9qcogAACYTZoTBrXbxKnoaIzxMEibOACAA3bJJd78J3n4w6oPWl18cYcXAgAA7JPGhEH9rUBg2zYHsU0xURk0dfepTRwA0ERlWebSn3Z6FcwEZzws6epKrrs+uf12x1YAADBbNCYMGq8M0iauMcbDoGmYGaRNHADQJHfdlaxblxRT132XWWrBgiIPfEB1/qIfdXYtAADA3mtOGDS3OnLdpqKjMaajTdzgeJs4n4IEAJrjhhur7fIjOroMZohHnlkdW114kdfEAAAwWzQmDOo3M6hxtmgTBwAwJa6/odoee2xHl8EMceYZ1fYnlyTbtwuEAABgNmhMGKRNXPNMz8yg6lOQw9rEAQANcsMN1Rv+xxzd4YUwI6xYkRx6aPVBO63iAABgdmhMGNTfahW2TRjUCGVZjleBDU5HZdDw1N0nAMBMpzKIybq6ijzh8dX5b/+3yiAAAJgNGhMGtSuDhEHNsHVrUraOS6eyMmhwsNpuVhkEADREWZbjM4OOPaajS2EGOfsJVcX8D3+YbNsmEAIAgJmuOWFQqzLIrJdm2LS52vb0TMyLmgrz5lXbLcIgAKAh1qypPmjT05MsP6LTq2GmeOADksOWVW24L7q406sBAADuS2PCoHYgsE0Y1AibW2HQvHlJURRTdr9DrTCoHTYBANRduyroqCOT3t6pe13F7FYURZ74hOr8t7+jMggAAGa6xoRB7TZxW4VBjTA5DJpK7fvbLAwCABrixpuqrXlB3N0TH99qFXdhsnWrQAgAAGayxoRBA61ZL9rENUO7cmdoGsOgsnTACwDU36rV1WueI5d3eCHMOCeemBxxRDWX9YcXdXo1AADAvWlMGDSvFQZt25bs3OlN/LprV+4MDU3t/bbDpZ07k+3bp/a+AQBmotWrq+0Rh2sRx66KosgTH1+d//Z/O8YCAICZrDFh0ODgxPktWzq3Dg6OTZuq7VS3iZs7N+nu2vUxAADqbNWqanv44Z1dBzPT2U+sQsIf/DBZt04gBAAAM1VjwqCeniJz5lTntwx3di1Mv+maGVQUxfh9bjI3CACoubGxMqtvq84fcURn18LMdMLxRR74wKpy/itf7fRqAACAe9KYMCiZaBW3xZv4tbd5c/WpxKkOgybf52a/RwBAzd11VzIyUlVGH7q006thpnr2s6rqoC99uczYmOogAACYiRoVBrVbxW3WJq722lU7Q/Omvre9MAgAaIpVrXlBhy6rKu1hT85+QvUaefVtyfnf7vRqAACAPWlkGGRmUP1NV5u4yfcpDAIA6q4dBh1hXhD3or+/yAtfUIWF7/2XMhs3qQ4CAICZRhhELU1nGDRkZhAA0BCrVlVv6guDuC+//cLk2GOSdeuS975PGAQAADNNo8KgedrENYbKIACAA7f6tmp7xBFaxHHv+vqKvP51Rbq6kq99I/nmfwmEAABgJmlUGDTYehNfZVD9TcwMmvr7ngiDHOACAPW2alW1Pfywzq6D2eGUk4u8/KVVcPgP7y5z8y1eLwMAwEzRqDBoojLIQUndbZ7GMGhoqNjlMQAA6mp8ZtARnV0Hs8fLXpKcdmqydWvyV28rs327Yy8AAJgJGhUGDQxU2y3exK+1sbFyvPprOtvEmRkEANTZ9u1l7ryzOn+4mUHspe7uIn/15iILFyTXXJu8+z1lylIgBAAAndaoMGjevKqiY8twhxfCtNq6NRkbq85PRxg0ZGYQANAAt7XmBc2dmyxc0Nm1MLscckiRt/xlNT/oP7+afPwTnV4RAADQqDBosNUmzsygemtX7PT1JnPmTP2w43nCIACgAVa1wqAjjkiKYupfU1FvjzijyB//z+r35sMfKfONb6oOAgCATmpUGDQ+M8ib+LW2aVO1nY6qoMn3234cAIA6WrWq2h5xWGfXwez13GcX+e0XVuff+b/L/OjHAiEAAOiURoVBKoOaoR32TXcYJFQEAOps9erqjfsjjujwQpjVXv2qIk98QrJzZ/KGN5W56GKBEAAAdEIzwyAzg2ptusOgyTODDMMFAOpq1epqe/jhWsSx/7q6irz5jUUe++hkx0jyxjeXufAir6EBAOBga1QY1G4Tt0VFR60drMqg0bFk69bpeQwAgE4bbxN3eGfXwezX11fkb/66yFmPS0ZagdB/fUsgBAAAB1OjwqDJlUFjYw4+6mpTKwwamqYwqL8/6e6uzmsVBwDUUVmWWX1bdf5wYRBToLe3yNveOtEy7m3/q8zH/q1UaQ8AAAdJo8KgdkVHWaroqLNNm6oDyqGh6bn/oigyf351fsPG6XkMAIBO2rhxYs7m4Yd1di3UR09Pkb96c5Hf+s3q6498rMzfvL3Mtm0CIQAAmG6NCoP6+iYqOtoHt9TPuvXVduHC6XuMBQuq7YYN0/cYAACdsro1L2jJkqS/38wgpk53d5E/+P2u/MXrinR3J/91fvL/varMNdcKhAAAYDo1KgwqimJ8btBmYVBtrV9fbRcunL43LhYu2PWxAADqZGUrDDIviOnyzGcU+ad/LLJkSXLTzcmrfr/MZ84rMzoqFAIAgOnQqDAoSQZbreJUBtVXu1qnXb0zHRaqDAIAamzVqmorDGI6nX5akX/7SJHHPDoZGUn++V/K/N5ryvzqSoEQAABMtcaFQSqD6m+8Mmg6w6CFrcfa4EAVAKif1bdVr3EOFwYxzRYuLPLOvy3yuj+tujhceVVVJfT3/zCW29Z4rQ0AAFOlcWHQYCsM2rK5s+tg+qxvVetM58yg9n235xMBANTJRGWQeUFMv6Io8uxnFfnUuUWe+pSkLJMv/2fywheV+cd/EgoBAMBU6On0Ag62+fOr7YaNnV0H02NsrMzGdhg0jZVBCxYUSUpt4gCAWlrdmhmkMoiDafHiIm95U5FnPaPMRz5W5tLLki98MfnSl8s85tFlnvvsIg89vQqPAPbHxk1lLrwwueTSMtden6xdm+wcSRYtWpcjl4/l5JOLPP4sH4YAoJ4aFwaZ9VJvmzcno2PV+WmdGbSw2rZb0gEA1MXoaJnb1lTnjziis2uhmR5ySpF//qcil15W5t/OLXPJpcn3vp987/tlDj88edITyzzp7CL3O04wBOydm24q85nPlvn6N5Pt23e/fv2GsdxwY/L9H5T51w8kD39YmZe9pMipD/F/DAD10bgwqB0QrF9fJvFHvW7a4czgYNLXN337d+H479G0PQQAQEfccUcyOpr09iaHLOn0amiy008rcvppRa6/ocznv1jm69+oqtbO/WRy7ifLHHtscvYTksc9NoIhYI9uW1Pmff9a5jvfnbjs2GOTxzwqedADiyxblvT2JDtHh3LpZZty4UVVVeKPf5L8+CdlHv+4Mv/ztUUOPdT/LwDMfo0LgxYu1N6rztrzgqazKihRYQYA1NeqVou4ZcuS7m5vftF59zuuyJ/9cZHX/F6ZH16YnP/tMhddnNx4Y/KRj5X5yMeqloaPe0yZxzy6yMknJT09fnehyXbsKPPp85JzPl5m+/akKKoA6AXPL3LqQ3YPjxct6s2KE4q88AVFVq0u88lPlfnKfybf+V5y6U/LvPH1yWMf4/8VAGa35oVB7YoOb+LXUrtSZzrnBSUTbeI2bKjmFHV1eVEIANTDqlXV9gjzgphh5s4tcvYTk7OfWGTTpjLfvyD57vfL/PgnVcXQp89LPn1emQXzk0c9ssxjH1PkjIcn/f1eq0OT/PgnZd79njK33FJ9fepDkj/5oyL3v9/e/V9wxOFF/vzPijz3OWXe8Xdlrro6eeObyzzvOWVe8+oic+b4PwWA2al5YdDCaisMqqf2fm3v5+nSrjwaHavmFM2fP72PBwBwsKxaXSYRBjGzDQ0VefrTkqc/rcjWrWV+9JPk+xdUlUMbNiZf+0bytW+U6etLznh4VTH06EclixZ6Exfq6vbby7z3X8v893eqrxcvSv7g94s85cn710by/vcr8oF/ST744TL/8Znkc59Prr6mzDv+1v8lAMxOjQuDFpj1Umvttm3TXRnU11dkYKDM8HAVQAmDAIC6aLeJO+IIb3QxO8ydW+SsxyZnPbbIzp1lfv6L5IIfVJVDq29LLvhB9XVXV3LSg6uKocc+OjnySL/jUAc7dpQ573PJOf9WZuu2pKsree5zklf+TpF58w7sed7bW+S1ryny0NPLvO1/lfnF5cnvvabMP7wzOeYY/4cAMLs0Lgya3N6rLEtDRmtm/frqk6zTXRnUfozh4SpYPPqo6X88AICDod0m7vDDOrsO2B89PUVOPy05/bQif/gHZa67Pvn+Bcn3f1Dm6quTn/8i+fkvyvzL+5Pjji3z2MdUc0BOXBGtn2GWKcsyF/wged+/llnZ+tt18knJn/5xkROOn9rn8yPPLPKBf03+/A1lVq1Kfu8Pyrzzb5PTTvX/BgCzR/PCoFbFyM6dyZYtybx5nV0PU2t8ZtBBKNleuKB6s0SVGQBQJytXVtsjj+zsOuBAFUWR4++fHH//5HdeVuS2NWV+8IPkexeU+enPkhturE4f/0SZpYckj350mcc9pshpp1bVAMDMVJZlLvtp8rF/q7ZJsmRx8upXFXnqU6Yv2D32mCIfen/yxr8sc/kVyZ/+eZk3v7GaYwYAs0HjwqA5c4rM7a9KhzdsEAbVTXtm0IJpbhOXTASLG8yfAgBqYuOmMhs2VueXH9HZtcBUO2xZkec9N3nec4ts3FTmoouqiqGLLk7uuDP5wheTL3yxzOBgcuYjyjz20UXOfEQOuM0UMDV27ixz4UXJp/6jateWJH29yW++IHnJi4oMDEz/c3XRwiLveXfyN28v893vJX/1N2VuvyN54Qv2by4RABxMjQuDkioo2LotWbc+Wb6806thKo1XBh2MMGhh6zGFQQBATaxqVQUtWZyD8qYadMr8oWqo/FOeXGT79jKX/jT5/gVV5dBda5NvfTv51rfL9PQkp59WzRl6zKOSpUs9L+BgKssy11ybnP+tMl//RrJ2XXV5X2/yjF9Pfvu3ihy27OA+L+fMKfI3f5W891/LfPZzyb+8v8zttyevfU3S3e3/CABmrkaGQQsXJretUdFRR+1g5mDNDErac4q84AMAZr9bW2GQD0zRJHPmFHnkI5JHPqLI2J+U+eWvkgt+UOb7FyQ33Zz86MfJj35c5l3/VM0ZOv205PTTi5z2kGT+fMcBMJVGR8vcdHNy1dXVfK8LL0ruvHPi+kWLkqc9NXnBbxQ55JDOPf+6u4v80WuTZYdWYdB5n0vuuKPMW/6y+j8FAGaiRoZB7RZiKjrqpSzLrF1bnV+8aPofb9GiIkmZu9ZO/2MBABwM7TDoSGEQDdXVVeSkBycnPbjIq1+V3Hxzme//oAqHLr9iYs7Q5z5fpiiS4+/fCodOK/KQU7SUg32xc2eZm1vBz1VXl7nq6uSaa5Nt23a9XX9/8vCHJU97apFHPTLp6ZkZz7OiKPJbv5ksPSR5+9+V+c73krvWlvm7tycLFsyMNQLAZI0Mg9oVHSqD6mX9hmRkJCmK5JBDpv/xDl1abe+4Y/ofCwDgYFi5skySLF/uTSxIkqOPLvKio5MX/VaR9evL/PRnyaWXVW3lbryxeuP6mmuTT59XpqsrWXFCmVNOTk45ucjJJyVLlnguQVIFPzfdnFx11a7Bz/btu992bn9ywgnJA05MHnFGkVMfMrOrbZ50dpHFi5M3vbmaZfT7ry3zD3+fLD9i5q4ZgGZqZhjUrgzS3qtW2qHM4kVJb+/079elrTDodmEQAFATKoPgni1cWOTxZyWPP6s61rjrrjKX/TS59KdlLr20ev5ceVV1+sxnW8HqEWVOPjk5+aQqHDrmaDNFqL8dO8rccGNy9dXJVdeUueaa5Nrr7iH4mZusOCE5cUVy4ooiJ56YHHXk7HuenH5akX99X/K615e5+Zbklb9X5q/fUoVZADBTNDMMWli199Imrl7aoUw7pJluhx5abe+8MxkbK9PV5UUeADC7mRkEe2/JkiJPOruqCkiSNbeX+dnPqzknv7g8uf76ZOWq6vT1b1Th0Nz+5Pjjy5y4IllxQpEVK5Jjj5k5ba9gX23dWua666tWb1e3gp/rb0h27tz9tgMDdwt+ViRHHZXaHEvf77giH/zX5E1vLfOrXyWv+4sy/9/vJC99cX3+jQDMbo0Mg8ZnBq3v6DKYYrffXm0PVhh0yJKqJd3ISNVycNFBmFMEADBdtmwps25ddX75EZ1dC8xGyw4t8pQnJU95UvWm76ZNZa74ZfKLy8v8/BfJr65Mtm5LfnF5dUqqgKinJzn88DJHLU+OPLJq03jIIcmSxdUxxuJFydy53kim8zZtKnPNtcnV1yRXX13m6muTm29OxsZ2v+38+VXwU52KrDih+v2ueyiydGmRf3lP8p73lvnil5P/+9GqgvBNb6j+jwCATmpkGLRQGFRLd9xRHUwdepDCoJ6eIosXl7nrrqoqSRgEAMxm7aqghQuSoSFvWMGBGhoqcuYjkjMfUT2fRkfL3HLrROusq6+u3lTfsiW55ZbqVCl3u6/urjJ9c5I5rVN/a9vdk/R0J92TTnf/un3q66veoB8aKjJ/KBlqnRYuqI5l5s+ffa25mD7r1pVV6HNNNePn6muSVav2fNslS5ITT6jm/KxYUeTEE5Jly5KiaObvU19fkT//syIPelCZf3pPmUsuTV72O2X++I+Spz65uT8XADqvkWHQIYdU2zvv7Ow6mFoTbeIO3gurQ5cmd91VzSs6ccVBe1gAgCl3003V9phjOrsOqKvu7iLHHlO1hXvKk6tjlrGxMrffkdx6a+u0sszKVcnata3TumrOyuhYsnVrdTpwu4dNSdLVlSxYUGbRwiocWriwqkpatKioLltcBUf9/VWwNGdOa9tXhU3Vvycpy+rUPl8UY9m4sUxRVI9x921PjzfHO6ksy9xxx66hz9VXJ3fcw/slhx+WrFiRnHB81ebthBOSQ5bYf3vy608rcsrJyf96e5lf/ir523eU+eZ/JX/8P5Ojj/IzA+Dga2QYtKw96+WuZOfOUn/mmrijFQYdrMqgpGpJ96srJ4IoAIDZ6oabqjeIjxUGwUHT1VXksGXJYcuShz00SXY9Ni3LMlu3JsPDybZtyfYdVTjUPo2OJjtHq237tHPnnr/evr3Mpk0ZP23clGzcWLW83rCxCm/WratOuWGXVRzgv3LdvV7b3Z3MGywzOC+ZN1jNlRmaV1WbLFlS5JAl1Qc629uFC4VH+6Msq64WN95Uhf833FRW2xuyx3nKRZEcdWQV/LTbvK04IZk/389+Xxx1ZJF/fW/yyX9Pzvl4mR/9OHnp75T5rd8s8+LfLjI46OcJwMHTyDBo0aKkt7ea9XLnnclhh3V6RUyFicqgg/eY7eCpalHnRRwAMHvd2Hrz99hjvaaBmaIoigwMVAHJFNzbPV6zc2eZDRuSdesnAqHqfLnLZes3JDtaQdSOHcmOkQNf1ehoFUZt2Lina3cPovp6k6VLyxx6aMZPy5YWu3w9NO/gB0ajo2U2bqx+buvXVyHb+g2t8xvL8Z/Z5O3oaFUd1T51d09se3urf2vP3bc9VRuynp5dLx8rW4Hh9mq7ZbgKf9qnO++qQsU96e5Kjj2u3eqtCn5OOD4ZGPD3YCr09BR52UuSs5+Q/J/3lrno4uTcTyZf/HKZ335h8rznmAsGwMHRyDCoq6vIoUur8vs1twuD6qBd2p4c7MqgIkmpMggAmPVuaLWJUxkEzdPTU7Qqce5+zb2/QT02VmZkpKo+KoqJ1m/JRCu4xYsXZe3adbu0jhsbq24zOlq1vtu8pZqdtHlLsmVzsmlzsnZtmTvvnAgy7ryzClp2jCQrV1WnCbuGRnP7k0MPnRQYLU0WLCgyr1V9NDiYzJtXbXu6k672bKWu6kOj23dUgcqO7cm27cnWbcnGDe2Ap8z69vn1E6eNm6p/28Gxfw/U1ZUccURy3DFVS9Bjjily3DHJccclc+YII6bbkUcW+Ye/S75/QfLBD5e56ebkAx8q8x+fSZ7zP8o8538UWbzYfgBg+jQyDEqqYYYrVyVr1nR6JUyFTZuqT0AlEzOhDoal45VBB+8xAQCm2vbt5fhg8OOO6+xagNmjq6vInDnV/KB70t1d3Gtr9qGh5NA9XrP794yMVAHR7XdUH+y8/fbk9turD+fd3vp6/YYqvLnp5uo04eAkNfPnV7OVFi6cOC2Yn/T3F+nr23XeUk9PMjZahWOjY61tq73fyI5kZGcVTo2MlNX5SZftHKmCsfa2q6u63/7+pH9OMndusnhxFfIdsiRZsjg5/PCqqojOKYoij3ts8uhHJd88P/noOWVWr04+9m/JJz5V5klPLPOMX69mDWmHCMBUa24Y1Hq1ueb2zq6DqdGuzFm48OB+oqldhaQyCACYzW65tXoTct686g1DgJmot7fI4YdXocaEXY//tm/fNRxac3vV1nvjpqr6aNPmqvqoXYk02gpj2tVKyUSgMqc/mdNXfT1/frJgQRX0LFhQjIc8ixZNhD/z52eaZhILBeqmu7vI056aPPns5LvfSz59Xplf/ir52jeSr32jzJFHJr/2lORxj6k+pCEYAmAqNDcMWlZt16wx66UOOtEiLtm1MqgsSy/QAIBZ6cYbq+1xx3rDCZjd5swpctSRyVFHTr70vv9fGxsrMzZWtYvz/yAHS09PkbOfmJz9xCKXX1Hmy/9Z5tvfTm69Nfm/Hy3zfz+aHLk8OfPMMqefWuTUU5P5Q34/Adg/zQ2DDq1mvagMqod2W5Nle+4vMG2WHlKV42/fnty1tiq/BwCYbW64sWqfZF4Q0FRdXcX4vCPohJMeXOSkBxf5o9eW+c53k+98r8xPfpLcujL57OeSz36uTFEkJxxf5pRTkhNXFFlxQnLM0dNVkQZA3TQ3DBqvDOrsOpgaN91cvYFx9NEH93H7+oocfliZlauSm28WBgEAs9P111fb4471ZhIAdNLAQJGnPy15+tOKDA+X+dGPk59cWuayy6o5WFdfU53ac7D6epP73a/MMcckRx1Z5Oijk6OPqqrjDmYbfQBmvuaGQWYG1Up7MOgxRx/8FzrHHJ3xMOj00w76wwMAHLArr6q2J57Y2XUAABMGBoo8/qzk8WdV73XceVeZy36a/PKXZa65tgqFhoerv+PV3/Jyl+9ftqzM0UeldSpyzDHJscdW8wG1QwRonsaGQYe2wqAtW5LNm8vMm+eP4Gx2cysMOtiVQe3H/OFF7eokv0cAwOxy511l7rizan17wvGdXg0AcE8OWVLkyWcnTz67eu9hbKzMqlXJddcnN9+S3HxzWW1vSTZtqrrhrFmT/PgnyeSgaN5gcswxVTXRsccUOfaY5JhjksMPq1omAlBPjQ2D5s4tsmB+mQ0bk9vWJMfP6/SK2F9btlRvYCSdCYOOPaaaP9WuTgIAmE3aVUHHHF19AhkAmB26uooceWRy5JHtS6q/42VZZv2G5JZWMHTzLWVuvjm58aZq5vLmLckVv6xOk0OiuXOT4+9f5oTjkxUnFDn++OS4Y7WbA6iLxoZBSXL44cmGjcmttybH37/Tq2F/3XxLtV28KJk/dPBfoLQDKGEQADAbXXll9SbQA7SIA4BaKIoiixYmixYmp5ycTO5isn17mVtXVsHQjTdWH2y96abqvZWtW5NfXF6d2iFRd3dy7DFljm8FRCtOqCqJddgBmH0aHQbd//7VJyGvu74c77/K7DM+L+iYzjz+Ma0waM2aZOvWMnPn+l0CAGaPdmXQAx7gNQwA1N2cOUXuf7/k/vdLJodEO3dWLeauvTa5+ppqJtE11yYbN1Zt6K67PvnGNyeqiI5cXubEFcmKFVVAdOKKZP58ryUAZrJmh0H3q9p7XXtdp1fCgahm9UyEMgfbggVFFi4ss359csutyYoTOrMOAIB9VZZlrryyOq8yCACaq6enyP2OS+53XPKUJ0+0m7vjjowHQ1dfU+bqq6txC7eurE7f+u+JgOjww8qsWFFVEJ24ogqIFi0SEAHMFI0Og9qt4a67vrPr4MDc3KoMOvrozr3AOOboZP36qkpJGAQAzBa3rkzWb0h6erRNBgB2VRRFDj00OfTQ5NGPStqVRBs2lLn6muSqqycColtXJqtvq07f/d5EQLT0kCogOnFFUW1PSA45pLpvAA6uRodBVUlsNTxveLg0MHeWuvbaantsh9rEJVWLup/9PLnm2jJPPtvvEQAwO/z4J9X2pAcbDg0A7J0FC4o8/GHJwx+WtAOiTZvKVvVQctXVVUB08y3JHXdWpx/8cCIgmjeYHHlkmaOOSo5cnhx1ZJEjj0wOW5YsXJh0dXlNAjAdGh0GLVhQZOkhZe64M7n+huogmNnlzrvKrFyVFEXywAd0bh0nPajIl75c5he/6NwaAAD21Y9/Ur0x8/CHedMFANh/Q0NFTj8tOf20pB0QDQ/vHhDdeFOyeUs1s7A9tzCZCIp6e5NDDilz6NJk6dJk6SHV+3fz5ycL5md8OzgvmdOX9PUlc+Yk3d1eywDcl0aHQUly//tXn1C49jph0Gz081b4cv/7VS88OuUhp1TbX12ZbN9e+mQtADDj7dxZ5tLLqvPVJ3sBAKbOwECRh5zSfs+kep9k+/YyK1cmt6xMbr01ueWWcnz+0F13JSMjyerV1WlCuYd731V3d5k5c5LenqSnt2qB29tTbbsnnZ986m1dt6fb9vZW7zMtWlhVKy1cmPHzAwPa3AGzkzDofslFFyfXXlem/YeJ2ePnv6heELTDmE454ohkyeLkrrVVIHTqQzq7HgCA+3LlVcmWLcnQUDXgGQBgus2ZU+R+90vud7/2JRPvxe3cWebOu5Lbb0/uuCO5/Y6qI8zGDcmGjcmmTcmGDcnGjcmW4So4ahsdTYaHp3q1ew6h+nqTRYvLLF6cHLI4Wdw6LVlSZMniZMmS1mWLkr4+7zUCM0fjw6ATTyySlPnpzzq9EvbHz39ebU85ubN/XIuiyCmnlPnv71Szg4RBAMBMd/GPqjc4Hnqa1ioAQOf19BQ5bFk1O2jCPb9GGR0ts2NHsmNHsn17sn1HFRDt3Lnn00hrOzrp/O6nMjt3JjtGko0bkvUbkvXrk3XrqvPbtlXXrVlTnXa1e3g0f34VGi1pnargqMghSyZCpEWLkvlDXo8B06/xYdDDHpp0dyU33pisXl3m8MP9xztbbNlS5trrqvOdrgxKqkDqv79TtqqV/B4BADNXWZb5xn9V5x/zGK9bAIDZp7u7yNy5ydy5U3mv9/66aOvWMuvXJ2vXJWvXVq3t7lpb5q61ra9bl61dW4VLGzdWpxtvnHwvu4dGRZEsmF+Ot6RbuGCiPd3ChUUWLkgWLEiOPnpnurvKLFiQ9PZ6DQfsm8aHQfOHipx0Upmf/Ty58OLkuc/u9IrYWz+5JBkbq1q0HXJI5/8AntoKpH7282pI4sBA59cEALAnP/9FsmpV9ebJWY/t9GoAAGaHuXOrAOrwwydfuvv7P2VZZuPGTIREd7WCorXl+Ndr1yZ33pVs3pyUZasKaUOSm3a7t0nnN4yfmzc4KTxaODlAKnYPlBYkc+aYdQRN1/gwKEnOfESRn/28zEUXlXnus/2nOFt87RvVH8MnnNXhhbQcf3yy/Ihk5arku99PnvbUTq8IAGDPvvr16nXUE59QvakBAMDUKYoiC1rVPPc7bpdrdrvtzp3leDu6Da3t+vXJ+g1l1q2f/HWycUOR9RvKjI0lm7dUp1tX3v0e9zzrqCiSvr4yc+Ykc/qqcGjOnKSv9XXfpMvmzEn65yRz+qvr+vuL6utdLpt02/6MXz93ruAJZiphUJJHPTL54IeTn1yabNtWpr/ff1Yz3dq1ZS68sDr/tF+bGfurKIo87deS//vRMl/7epmnPXVmrAsAYLK1a8t8+9vV+afPkNdRAABN1dNTzRA6ZMndr9n9ddqiRYty551rs2nTxDyjyWHR+vWTAqQNE+HSzp1V9dH27dVp3+05YLon3V3J3LllBgaqcGhgoHVqnZ+4rNjDZdXtJn8tXIKpIQxKldAftiy5bU3yw4uSJz6+0yvivnzz/GR0LHnQA5Njj5k5fwx+7anJRz6WXHpZsmp1mSPMoAIAZpiPfbzM1m3JA05MTjm506sBAGBfdHcX4+3fcszdr91zy7rh4WTbtmT7jolAaPv2ZMeOict2tC8fv02Zbe3LtmXi/PbWfW1vXda6323bqvtLqvfs2pVL927vQqauVrjU25t0d+966une/bK7n7qKak1lWY2caG/bp/HL7+EnWRTtyqpWNVW7kqrv7pcVu1zX2ztx2/apq6u6r8n3O36+9cDFpFNvb9Lbk/S0t61TX18VJMK+EAalSpafdHaZT3wq+a/zyzzx8Z5IM9nGTWU++e/Vf89Pf9qB76vR0dF85zvfybXXXptly5blkY98ZLq7u/frvg5bVuRhDy3z458kH/q/Zf76LX6XAICZ44Yby3zpS9X517y68AlLAKbF6OhoLrzwwqxZs+aAj7OBA1MURQYHk8HBff7OfX6ssbEy27Ylw8OTTltbp9bXW7dWs7Z3v2zidlsnbav7TbbcZ7A0E+xbBdWB6usrM28wGZyXzJuX6vxgMtT6emioyNBQMn8oGWqf5lXbefOqYLEOyrLM6GgVtHV1Tf+/qSyr/Twbj6WEQS1PeVKRT3yqzEUXV2HD/KHZtzOb4gMfKrNuXXLsMcnTf+3A7uvLX/5y3vSmt2TlypvHL1u+/Oi84x3/K8985jP36z5/73eLXHJpmfO/lTztqWUecYbfJQCg8zZuKvOmt5QZHavaJJ9+mtcoAEy96TjOBmaHrq5ivLXbvdu716Hj4VIrHNq5Mxkd3Y/TWNW6ruhqBQZF63wxUanT1XW3ZZV3X0tV+dSuptrRrqjaUU5UWW1PRkaq60dGJm6/Y9L5sdHqrsuyeoxWrrDbZWWq2+4cre5rdGd1P5Pt2JGs3ZGsXXdPP8F7D6cGB8vdgqKBwVb1UXdVldTdOt/TU6Snp7W29prba21d0P56Tv9wtm4dG/+33NM+2dPlO3dOnEZGqn//zpFkpH35yMTPZPJtJ/+buybt567uav+29393+7ruajv592Lyde3z7e9p5z5lmVx/Q7JkSfKpj8++QEgY1HK/+xW5//3KXHd98u3/Tp79rE6viD25+poyX/pydf51f1qkr2//n3Bf/vKX8/KXvzzzhp6SY+//ofT3PyDbtl2Zu+74p7z85S/POeecs18vVB9wYpHnPafMeZ9L/s97y5z7MWWbAEBn3XZbmbe8rcwttySHHpr8xeu8NgFg6k3XcTbQTLuES7vNVJopDt7r6nYFzM6dVRC0pdWKr73dvDnZsjnZtDnZvLnMps3Jxo3Jpk3VZZs2Vad2xdWW1veuvm2vHn0fVrp1P/51U6fd/i9JMnKvN91v7daLc+dOz/1PF2HQJL/21CL/8v4yH/1YmSc+IaqDZqAPf6T6j+dJZyenPmT/98/o6Gje9Ka3ZN7QU3LkMeemKLqSJAODD8vcgXNz600vyV/+5Vvz9Kc/fb9K2V/5iiLf/K/qDZevfzN5xtP3e6kAAPttbKzMl75SVVZv3ly1jvj7dxRZssTrXACm1nQfZwM0XVEU4zOD+vuT+fPv9db3eM3OnVVQtGljsvFuQdHwcFWhMzJSVlU3rcqc9vkirSqZ1kyjrknn29f19/dn+45t41/39Eya39RV3PN8p65WVdLd5iP19k6c7+mZmKE0fr51/+NzoMqqqmqsPRvqbudHJ82LGh2dmBk1Ojrp8rGknHS+/dPcsaPMX/+vKdqhHSAMmuS5z06+8p/JTTcnr3t9mWOPmd4+j319m7Njx9h935AkVSnkhRdVT+5X/s6BvYFx4YUXZuXKm3Ps/T80/gK1rSi6smTpH+fG656WCy+8MI95zGP2+f4HB4u8+EXJv7y/zIc+XObnPz+4PUNnM8+LmcF+mDnsi5nBfpg57Iu9t3NnctlPkzvurL5+4AOTt721yBGHC4IAmHrTfZwNwNTo6SmyaGGyaOG93Wr/jxkWLRrMunU79vv7Z7Kqqmr2vs8rDJpkzpwib3h98po/LPPLXyW//NV0P+L26X6AWnrG05MjjzywNzHWrFmTJOnvf8Aer5/Turx9u/3x3GcnnzmvegPmq1/f77tpIM+LmcF+mDnsi5nBfpg57It9NW9e8v/9TpHn/A+tawGYPgfjOBsAOqm/Pzn5pInzs40w6G5OPqnIu/53cs210/9Yc+fOzdatne2hONv0z0me9msHfj/Lli1LkmzbdmUGBh+22/Xbt125y+32x5w5Rd79j1U1Uzl7A+ODzvNiZrAfZg77YmawH2YO+2LfHHNM8vCHVq9LpkL74Kenpyf9/aNTcp8A1MPBOM4GgE4qiiL/+t6J87ONMGgPznh4kTMePv2Ps2jR3Kxbt236H4jdPPKRj8zy5Ufnrjv+KXMHzt2lhL0sx3LXHf8nRx55TB75yEce0OMcd2yR4449wMU2jOfFzGA/zBz2xcxgP8wc9kVntQ9+Fi2an/Xr13d6OQDMIAfrOBsAOmk2hkBtXfd9E6if7u7uvOMd/yubN30zt970kgxv+XFGRzdleMuPc+tNL8nmTd/M29/+N4ZaAgDcTVEUs/oACIDp4TgbAGY2YRCN9cxnPjPnnHNO5g/9Kjde97RcdcVxufG6p2XB/Ctzzjnn5JnPfGanlwgAAACzhuNsAJi5ZmWbuNk+qImZ45nPfGae/vSn5/LLL8+1116bZcuW5ZGPfKRPKgEAMGP19yenndqTnTt3Oh4CZpz2cfaFF16YNWvWOM4GgBliVoZBs31QEzNLd3d3Hv/4x+chD3lIp5cCAAD3qSiKnHvO/Kxbt87xEDAjdXd35zGPeUynlwEATDIrw6BECAQAADSX2U0AAMC+MDMIAAAAAACgxoRBAAAAAAAANSYMAgAAAAAAqDFhEAAAAAAAQI0JgwAAAAAAAGpMGAQAAAAAAFBjwiAAAAAAAIAaEwYBAAAAAADUmDAIAAAAAACgxoRBAAAAAAAANSYMAgAAAAAAqDFhEAAAAAAAQI0JgwAAAAAAAGpMGAQAAAAAAFBjwiAAAAAAAIAaEwYBAAAAAADUmDAIAAAAAACgxoRBAAAAAAAANVaUZVl2ehEAAAAAAABMD5VBAAAAAAAANSYMAgAAAAAAqDFhEAAAAAAAQI0JgwAAAAAAAGpMGAQAAAAAAFBjPZ1eQJ38/Oc/z3vf+95cdtll2blzZ1asWJGXv/zlefrTn75X3////t//yxvf+MZ7vP7jH/94HvGIR0zVcmvri1/8Yi655JJcfvnlufrqqzMyMpJ3vvOdee5zn7tP9zM2NpZPfvKT+cxnPpObbropAwMDedSjHpU/+ZM/yVFHHTVNq6+PqdgPF198cV760pfe4/X7s1+bZs2aNfna176W733ve7n++utz5513ZsGCBTn99NPzyle+Mg95yEP2+r48J/bfVO0Hz4kDt3379rz73e/O5ZdfnptuuikbNmzI/Pnzc9RRR+X5z39+nvWsZ6W3t3ev7stz4sBM1b7wvJgeH/rQh/Kud70rSfLpT386p5566l59n+fF/jvQY4kk2bFjRz70oQ/lS1/6UlavXp0FCxbkCU94Qv74j/84S5YsmcbV0ySOe5npvCfAbOA9E2YD7ynVjzBoilx00UV55Stfmb6+vvz6r/96BgcH881vfjN/8id/kttuuy2veMUr9vq+zj777DzwgQ/c7fLly5dP5ZJr6z3veU9WrlyZRYsW5dBDD83KlSv3637e+ta35rzzzssJJ5yQl7zkJbn99tvzta99LT/4wQ/y6U9/Oscee+zULrxmpmo/JMkZZ5yRM844Y7fL9/Q8YVfnnntuPvzhD+foo4/Oox/96CxevDg33XRTzj///Jx//vl517vetdcH7p4T+28q90PiOXEgtmzZkn//93/PKaecksc//vFZvHhxNmzYkO9///t505velK9+9av58Ic/nK6u+y6e9pw4MFO5LxLPi6l09dVX573vfW8GBgYyPDy8T9/rebF/puJYYmxsLL//+7+fCy64IKeeemqe8pSn5Kabbsp5552XCy+8MJ/5zGeyePHig/Cvoc4c9zIbeE+A2cB7JswG3lOqoZIDNjIyUj7pSU8qTzrppPKXv/zl+OUbN24sn/KUp5QPfvCDy1tvvfU+7+dzn/tcuWLFivJzn/vcdC639n7wgx+M/7w/+MEP7tfP9MILLyxXrFhRvuhFLyq3b98+fvl3vvOdcsWKFeUrXvGKKV1zHU3FfrjooovKFStWlP/8z/88HUtshG984xvlxRdfvNvlP/7xj8sHP/jB5cMf/vBdfsfviefEgZmq/eA5ceBGR0f3+LMeGRkpX/ziF5crVqwo//u///s+78dz4sBN1b7wvJhaO3bsKJ/znOeUz3/+88vXve515YoVK8rLLrtsr77X82L/TNWxxGc/+9lyxYoV5Z/+6Z+WY2Nj45d/6lOfKlesWFG+5S1vmZb10xyOe5ktvCfAbOA9E2YD7ynVj5lBU+Ciiy7KzTffnGc84xm7pO5DQ0N59atfnZGRkXz+85/v4Aqb5VGPetQBf5rsvPPOS5L80R/9Ufr6+sYvP+uss3LGGWfkggsuyKpVqw7oMepuKvYDB+4pT3nKHj8h9LCHPSyPeMQjsmHDhlx11VX3eT+eEwdmqvYDB66rq2uX3+G2np6ePPnJT06S3HTTTfd5P54TB26q9gVT6wMf+ECuueaavOMd70h3d/c+fa/nxf6ZqmOJ9s//T//0T1MUxfjlL3zhC3PUUUfly1/+crZt2zb1/wAaw3Evs4X3BJgNvGfCbOA9pfoRBk2BH/3oR0mSxzzmMbtd177sxz/+8V7f3y9/+ct89KMfzYc+9KF89atfzbp166Zmoey1iy++OAMDAzn99NN3u+6xj31skon9zvS78cYbc8455+SDH/xgvvCFL2TNmjWdXlIt9PT07LK9N54T02df9kOb58TUGxsby/e///0kyYoVK+7z9p4T02df90Wb58WBu+KKK/KBD3wgr33ta3P88cfv8/d7XuyfqTiW2L59e372s5/luOOO2+2NpaIo8qhHPSrDw8O5/PLLp2jVNJHjXprE3zRmE6+D6QTvKc1OZgZNgRtvvDFJcswxx+x23dKlSzMwMLBPn2w999xzd/m6v78/f/AHf5BXvepVB7RO9s7w8HDuuOOOrFixYo+fiG3vZ59WPni+8pWv5Ctf+cr41z09PXnxi1+c17/+9fv8qWUqq1atyg9/+MMsXbr0Pt9s9ZyYPvuyHybznDhwO3bsyAc/+MGUZZn169fnwgsvzPXXX5/nPve5eeQjH3mv3+s5MbUOZF9M5nlxYHbs2JG/+Iu/yAMe8IC88pWv3Ofv97zYf1NxLHHzzTdnbGzsHnutty+/8cYb87CHPexAlkuDOe6lKfxNY7bxOpiDzXtKs5cwaAps3rw5SVUevyfz5s3Lpk2b7vN+jjzyyLzlLW/JYx7zmBx22GHZsGFDLrzwwrz73e/Ou971rsydOzcveclLpnTt7K69r+bNm7fH69uX780+5cAsXrw4f/Znf5YnPOEJWb58ebZu3ZrLLrss73rXu3LOOeekKIq84Q1v6PQyZ52RkZG8/vWvz44dO/K6173uPl8cek5Mj33dD4nnxFQaGRnJ+973vvGvi6LIK17xivzZn/3ZfX6v58TUOpB9kXheTJX3vOc9ufHGG/P//t//2683DTwv9t9UHEvs7c+//ViwPxz30hT+pjFbeB1MJ3hPaXYTBs0gZ5xxxi59GPv7+/PsZz87D37wg/O85z0v73vf+/Jbv/Vb+9RKCGazE044ISeccML41wMDA3nSk56UhzzkIXnWs56Vc889N7/7u7+bJUuWdHCVs8vY2Fje8IY35Mc//nFe8IIX5NnPfnanl9RI+7sfPCemzuDgYK666qqMjY3l9ttvz7e//e380z/9U37605/mwx/+8D2+UGXqHei+8Lw4cJdddlk++tGP5rWvfe0+VSkC7A/HvQBTw+tgDjbvKc1+ZgZNgftKMDdv3nyPn57aGyeccEIe+tCHZv369bnuuuv2+37YO+19dU+fnLyvT8Qx/ZYuXZqzzz47O3fuzM9+9rNOL2fWGBsby5ve9KZ85StfybOe9ay87W1v26vv85yYWvu7H+6N58T+6+rqymGHHZbf/u3fzt/8zd/k0ksvzfvf//57/R7PiemxP/vi3nhe7J2dO3fmDW94Q0488cQDas3kebH/puJYYm9//oJuDoTjXprC3zRmO6+DmQ7eU6oHYdAUaPfg3lNvwzvuuCPDw8N77Ku8LxYtWpQk2bp16wHdD/dtYGAgS5cuza233prR0dHdrm/v5wPdpxwYz4l9MzY2lje+8Y35/Oc/n2c84xn5u7/7u3R17d2fAM+JqXMg++G+eE4cuPbw6/saXOk5Mf32dl/cF8+L+zY8PJwbb7wxv/rVr3LSSSflxBNPHD99/vOfT5L85m/+Zk488cScf/7593g/nhf7byqOJY466qh0dXWNz3S5u/bl9zRTCPaG416awt806sD/p0wl7ynVhzBoCjz84Q9PklxwwQW7Xde+rH2b/TE6OprLL788SXLEEUfs9/2w984444wMDw/n0ksv3e2673//+0kObJ9y4Nqfblm+fHmHVzLztf9of+ELX8jTn/70/O///b/3eR6E58SBm4r9cG88Jw7c7bffniR71ZbGc2J67cu+uDeeF/etr68vv/Ebv7HHU/uN3yc+8Yn5jd/4jfv8OXpe7J+pOJbo7+/PKaeckhtuuCErV67c5bqyLPPDH/4wAwMDOemkk6Zo1TSR416axN80Zjuvg5kq3lOqF2HQFHjkIx+Zo446Kl/5ylfyq1/9avzyTZs25QMf+EB6e3t36aF4++2357rrrtutvL79wney0dHR/OM//mNuuummPOIRj8ihhx46bf+OJlq7dm2uu+66rF27dpfLX/CCFySphinv2LFj/PLvfve7+dGPfpTHPOYx/qBOoXvaD3t6TiTJv/3bv+Xiiy/Osccem5NPPvlgLHHWapfxfuELX8iv/dqv5R/+4R/u9Y+258T0mKr94Dlx4K699to9fjpu69ateec735kkOeuss8Yv95yYPlO1LzwvDkx/f3/e/va37/F02mmnJUl+7/d+L29/+9vzwAc+MInnxVSbqmOJ9s//3e9+d8qyHL/8P/7jP3LLLbfkmc98Zvr7+6f3H0OtOe6ljvxNYzbwOphO8Z5S/RTl5CMF9ttFF12UV77ylenr68uv//qvZ3BwMN/85jezcuXK/MVf/EVe8YpXjN/2DW94Qz7/+c/nne98Z5773OeOXz65LceyZcuyYcOG/OhHP8qNN96Yww47LJ/4xCdy1FFHdeKfN6ucd955ueSSS5IkV199da644oqcfvrp4+WGD33oQ/P85z8/SfLe974373vf+/La1742f/iHf7jL/bz5zW/OeeedlxNOOCFnnXVW7rjjjnz1q1/N4OBg/uM//iPHHXfcwf2HzTJTsR+e+MQnpqenJyeddFKWLVuWrVu35mc/+1l++ctfZv78+fnIRz6SU0455eD/42aR9s92YGAgL33pS/f4KfsnPelJ42/weU5Mj6naD54TB+69731vPvaxj+WhD31oli9fnnnz5mXNmjX53ve+l/Xr1+dhD3tYPvKRj4y/Yeo5MX2mal94Xkyf9mvWT3/60zn11FPHL/e8mHpTcSwxNjaW3/3d380FF1yQU089NQ9/+MNz880355vf/GaWL1+e8847L4sXL+7EP48acdzLbOA9AWYD75kwG3hPqX4OrPcG484888x86lOfyj//8z/nq1/9anbu3JkVK1bkda97XZ7+9Kfv1X284hWvyE9/+tP88Ic/zIYNG9Lb25ujjz46v//7v5/f+Z3fyYIFC6b5X1EPl1xyyXiP+7ZLL710l1LE9h/Ue/M3f/M3WbFiRT7zmc/k4x//eAYGBvLkJz85f/Inf5Kjjz56ytddN1OxH174whfmggsuyI9//OOsX78+XV1dOeKII/Kyl70sr3jFK3LYYYdNy9rrpN0qZnh4OB/4wAf2eJvly5eP/+G+N54T+2+q9oPnxIF7/OMfn9tvvz2XXXZZfvrTn2Z4eDjz5s3LiSeemF//9V/P8573vL1uTeY5cWCmal94Xswsnhf7ZyqOJbq6uvL+978/H/rQh/LFL34x55xzThYuXJjf+I3fyB//8R8LgpgSjnuZDbwnwGzgPRNmA+8p1Y/KIAAAAAAAgBozMwgAAAAAAKDGhEEAAAAAAAA1JgwCAAAAAACoMWEQAAAAAABAjQmDAAAAAAAAakwYBAAAAAAAUGPCIAAAAAAAgBoTBgEAAAAAANSYMAgAAAAAAKDGhEEAAAAAAAA1JgwCAAAAAACoMWEQAAAAAABAjQmDAAAAAAAAauz/B0Z1RhZuVSnuAAAAAElFTkSuQmCC\n", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "# We can try plotting the density and view the high density intervals to understand the variable better\n", + "az.plot_density(\n", + " trace,\n", + " group=\"sample_stats\",\n", + " var_names=\"accept\",\n", + " point_estimate=\"mean\",\n", + ");" + ] + }, + { + "cell_type": "code", + "execution_count": 16, + "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ - "Last updated: Sun Feb 07 2021\n", + "Last updated: Tue Apr 06 2021\n", "\n", "Python implementation: CPython\n", - "Python version : 3.8.6\n", - "IPython version : 7.20.0\n", + "Python version : 3.9.2\n", + "IPython version : 7.21.0\n", "\n", - "numpy : 1.20.0\n", "seaborn : 0.11.1\n", - "pymc3 : 3.11.0\n", - "pandas : 1.2.1\n", - "matplotlib: None\n", + "pandas : 1.2.3\n", + "arviz : 0.11.2\n", + "matplotlib: 3.3.4\n", + "numpy : 1.20.1\n", + "pymc3 : 3.11.2\n", "\n", - "Watermark: 2.1.0\n", + "Watermark: 2.2.0\n", "\n" ] } @@ -536,13 +1688,20 @@ "%load_ext watermark\n", "%watermark -n -u -v -iv -w" ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Updated by Meenal Jhajharia" + ] } ], "metadata": { "kernelspec": { - "display_name": "Python PyMC3 (Dev)", + "display_name": "Python 3", "language": "python", - "name": "pymc3-dev-py38" + "name": "python3" }, "language_info": { "codemirror_mode": { @@ -554,7 +1713,7 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.8.6" + "version": "3.9.2" } }, "nbformat": 4,