job_arrival_characterization.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# The code in this cell comes from Caspar Greeven at SURFsara \n",
    "\n",
    "# Collected job data from Lisa using the following command:\n",
    "# sacct -a --starttime 2020-01-01 --format=jobid,gid,uid,partition,submit,start,end,elapsedraw,cputimeraw,ncpus,nnodes,nodelist,exitcode,state,timelimit > $HOME/jobdata.csv\n",
    "\n",
    "# Required packages for processing\n",
    "import os\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.patches as mpatches\n",
    "import matplotlib.pyplot as plt\n",
    "import time, datetime, pytz\n",
    "from matplotlib.ticker import MultipleLocator, FixedLocator, LogLocator, NullFormatter, ScalarFormatter\n",
    "from datetime import date, datetime, time\n",
    "import calendar\n",
    "from multiprocessing import cpu_count\n",
    "from pandarallel import pandarallel\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# The code in this cell comes from Caspar Greeven at SURFsara \n",
    "\n",
    "# Required preprocessing/parsing of the job data\n",
    "def preprocess_jobdata_to_df(name):\n",
    "    with open(os.path.join(location_job_data_csv, name),'r') as file:\n",
    "        filedata = file.read()\n",
    "        filedata = filedata.replace('None assigned','NoneAssigned')\n",
    "    with open(os.path.join(location_job_data_csv, str('processed_'+name)),'w') as file:\n",
    "        file.write(filedata)\n",
    "    jobdata = pd.read_fwf(os.path.join(location_job_data_csv, str('processed_'+name)), delimiter=r\"\\s+\", header=None)#, low_memory=False)\n",
    "    jobdata = jobdata.apply(lambda x: x.str.strip() if x.dtype == \"object\" else x)\n",
    "    jobdata = jobdata.rename(columns=jobdata.iloc[0]).drop(jobdata.index[0])\n",
    "    jobdata = jobdata.iloc[1:]\n",
    "    jobdata = jobdata.astype({\"ElapsedRaw\": int, \"CPUTimeRAW\": int, \"NCPUS\": int})\n",
    "    return(jobdata)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "location_job_data_csv = \"path/to/surfsara-jobdata/\"\n",
    "jobdata = preprocess_jobdata_to_df(\"jobdata.csv\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# The code in this cell comes from Caspar Greeven at SURFsara \n",
    "\n",
    "# Some extra preprocessing required for the job data, e.g. filtering out irrelevant login nodes\n",
    "filtered_jobs = jobdata[(jobdata[\"Start\"] >= '2019-12-29 23:00:00') & (jobdata[\"Start\"] <= '2020-08-07 21:59:45')]\n",
    "filtered_jobs = filtered_jobs[(~filtered_jobs[\"NodeList\"].str.contains(\"None\")) & (~filtered_jobs[\"NodeList\"].str.contains(\"software\")) & (~filtered_jobs[\"NodeList\"].str.contains(\"login\"))]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "filtered_jobs[\"dt\"] = pd.to_datetime(filtered_jobs['Submit'], utc=False, format=\"%Y-%m-%dT%H:%M:%S\")\n",
    "filtered_jobs[\"submit_hour_of_day\"] = filtered_jobs[\"dt\"].dt.hour\n",
    "filtered_jobs[\"submit_day_of_week\"] = filtered_jobs[\"dt\"].apply(lambda x : x.weekday())\n",
    "filtered_jobs[\"submit_date\"] = filtered_jobs[\"dt\"].dt.date\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO: Pandarallel will run on 8 workers.\n",
      "INFO: Pandarallel will use Memory file system to transfer data between the main process and workers.\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b7c88918d4ed494e8ab802647e4d67f6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "VBox(children=(HBox(children=(IntProgress(value=0, description='0.00%', max=281058), Label(value='0 / 281058')…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Creates a column that indicates if a job was executed on ML nodes \n",
    "gpu_nodes = {\n",
    "    \"r28n1\", \"r28n2\", \"r28n3\", \"r28n4\", \"r28n5\",\n",
    "    \"r29n1\", \"r29n2\", \"r29n3\", \"r29n4\", \"r29n5\",\n",
    "    \"r30n1\", \"r30n2\", \"r30n3\", \"r30n4\", \"r30n5\", \"r30n6\", \"r30n7\",\n",
    "    \"r31n1\", \"r31n2\", \"r31n3\", \"r31n4\", \"r31n5\", \"r31n6\"\n",
    "    \"r32n1\", \"r32n2\", \"r32n3\", \"r32n4\", \"r32n5\", \"r32n6\", \"r32n7\",\n",
    "    \"r33n2\", \"r33n3\", \"r33n5\", \"r33n6\",\n",
    "    \"r34n1\", \"r34n2\", \"r34n3\", \"r34n4\", \"r34n5\", \"r34n6\", \"r34n7\",\n",
    "    \"r35n1\", \"r35n2\", \"r35n3\", \"r35n4\", \"r35n5\",\n",
    "    \"r36n1\", \"r36n2\", \"r36n3\", \"r36n4\", \"r36n5\",\n",
    "    \"r38n1\", \"r38n2\", \"r38n3\", \"r38n4\", \"r38n5\",\n",
    "}\n",
    "\n",
    "def split_nodes(s): # parses node strings like r12n[1-30,32] to r12n1, r12n2 ... r12n30, r12n32\n",
    "    if s is None or len(s) == 0:\n",
    "        return set()\n",
    "    \n",
    "    s = s.replace(\"\\r\\n\", \"\").replace(\"\\n\", \"\").replace(\"\\t\", \"\")\n",
    "\n",
    "    start = 0\n",
    "    index = 0\n",
    "    rack_chunks = []\n",
    "    in_bracket = False\n",
    "    while index < len(s):  # Separate them in parts like r12n[1-30,32] or r13n1\n",
    "        if s[index] == \"[\":\n",
    "            in_bracket = True\n",
    "        elif s[index] == \"]\":\n",
    "            in_bracket = False\n",
    "        elif s[index] == \",\" and not in_bracket:\n",
    "            rack_chunks.append(s[start: index])\n",
    "            start = index + 1\n",
    "        index += 1\n",
    "    rack_chunks.append(s[start: index])  # Add the last line\n",
    "    \n",
    "    node_names = set()\n",
    "\n",
    "    for rack_chunk in rack_chunks:\n",
    "        if \"[\" in rack_chunk:\n",
    "            prefix, postfix = rack_chunk.split(\"[\")\n",
    "            postfix = postfix[:-1]  # Remove the last bracket\n",
    "            nodes = postfix.split(\",\")\n",
    "            for node in nodes:\n",
    "                if \"-\" in node:\n",
    "                    start, end = node.split(\"-\")\n",
    "                    if not start.isnumeric() or not end.isnumeric():\n",
    "                        continue\n",
    "                    for i in range(int(start), int(end) + 1):\n",
    "                        node_names.add(\"{}{}\".format(prefix, i))\n",
    "                else:\n",
    "                    node_names.add(\"{}{}\".format(prefix, node))\n",
    "        else:\n",
    "            node_names.add(rack_chunk)\n",
    "\n",
    "    return node_names\n",
    "\n",
    "def calculate_perjob(row):\n",
    "    nodes = row[\"NodeList\"]\n",
    "    splitnodes = split_nodes(nodes)\n",
    "    return any([n in gpu_nodes for n in splitnodes])\n",
    "\n",
    "pandarallel.initialize(nb_workers=min(cpu_count(), 8), progress_bar=True)\n",
    "filtered_jobs[\"is_gpu\"] =  filtered_jobs.parallel_apply(calculate_perjob, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "std_per_hour = filtered_jobs.value_counts([\"submit_date\", \"submit_hour_of_day\"], sort=False).groupby(\"submit_hour_of_day\").std()\n",
    "x_vals = np.arange(filtered_jobs[\"submit_hour_of_day\"].nunique())\n",
    "y_vals = filtered_jobs.value_counts([\"submit_date\", \"submit_hour_of_day\"], sort=False).groupby(\"submit_hour_of_day\").mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 792x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(11,5))\n",
    "negative_direction_values = np.zeros(len(std_per_hour))  # We create a 2d array to make sure matplotlib does not create downwards errorbars\n",
    "ax.bar(x_vals, y_vals, yerr=[negative_direction_values, std_per_hour], color=\"lightcoral\", edgecolor='black', capsize=3)\n",
    "\n",
    "ax.set_ylim(bottom=0, top=None)\n",
    "ax.set_xlabel(\"Hour of Day\", fontsize=20)\n",
    "ax.set_ylabel(\"Jobs Submitted\", fontsize=20)\n",
    "ax.tick_params(axis='both', which='major', labelsize=18)\n",
    "ax.tick_params(axis='both', which='minor', labelsize=16)\n",
    "fig.tight_layout()\n",
    "\n",
    "date_time = datetime.now().strftime(\"%Y-%m-%d-%H-%M-%S\")\n",
    "\n",
    "fig.savefig(f\"num_jobs_diurnal_hourly_{date_time}.pdf\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 792x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Cropped version of the plot above\n",
    "fig, ax = plt.subplots(figsize=(11,4))\n",
    "negative_direction_values = np.zeros(len(std_per_hour))  # We create a 2d array to make sure matplotlib does not create downwards errorbars\n",
    "ax.bar(x_vals, y_vals, yerr=[negative_direction_values, std_per_hour], color=\"lightcoral\", edgecolor='black', capsize=3)\n",
    "\n",
    "threshold = 1500\n",
    "ax.set_ylim(bottom=0, top=threshold)\n",
    "ax.set_xticks(range(0,24))\n",
    "every_nth = 4\n",
    "for n, label in enumerate(ax.xaxis.get_ticklabels()):\n",
    "    if n % every_nth != 0:\n",
    "        label.set_visible(False)\n",
    "ax.set_xlabel(\"Hour of Day\", fontsize=20)\n",
    "ax.set_ylabel(\"Jobs Submitted\", fontsize=20)\n",
    "ax.tick_params(axis='both', which='major', labelsize=18)\n",
    "ax.tick_params(axis='both', which='minor', labelsize=16)\n",
    "\n",
    "scatter_x = []\n",
    "scatter_y = []\n",
    "for x_val, y_val, std_val in zip(x_vals, y_vals, std_per_hour):\n",
    "    if y_val+std_val > threshold:\n",
    "        scatter_x.append(x_val)\n",
    "        scatter_y.append(y_val + std_val)\n",
    "\n",
    "index = 0\n",
    "for scat_x, scat_y in zip(scatter_x, scatter_y):\n",
    "    ax.text(x=scat_x, y=threshold * 1.02, s=int(round(scat_y, 0)), fontsize=14, ha='center', rotation=30)\n",
    "    index += 1\n",
    "\n",
    "# Annotate the mean for ML jobs\n",
    "x_vals_ml = np.arange(filtered_jobs[filtered_jobs['is_gpu'] == True][\"submit_hour_of_day\"].nunique())\n",
    "y_vals_ml = filtered_jobs[filtered_jobs['is_gpu'] == True].value_counts([\"submit_date\", \"submit_hour_of_day\"], sort=False).groupby(\"submit_hour_of_day\").mean()\n",
    "ax.scatter(x_vals_ml, y_vals_ml, marker='o', c=\"steelblue\", alpha=1, zorder=2)\n",
    "\n",
    "date_time = datetime.now().strftime(\"%Y-%m-%d-%H-%M-%S\")\n",
    "fig.tight_layout()\n",
    "fig.savefig(f\"num_jobs_diurnal_hourly_cropped_ml_mean_{date_time}.pdf\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 792x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "std_per_day_of_week = filtered_jobs.value_counts([\"submit_date\", \"submit_day_of_week\"], sort=False).groupby(\"submit_day_of_week\").std()\n",
    "x_vals = np.arange(filtered_jobs[\"submit_day_of_week\"].nunique())\n",
    "y_vals = filtered_jobs.value_counts([\"submit_date\", \"submit_day_of_week\"], sort=False).groupby(\"submit_day_of_week\").mean()\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(11,5))\n",
    "negative_direction_values = np.zeros(len(std_per_day_of_week))  # We create a 2d array to make sure matplotlib does not create downwards errorbars\n",
    "ax.bar(x_vals, y_vals, yerr=[negative_direction_values, std_per_day_of_week], color=\"lightcoral\", edgecolor='black', capsize=3)\n",
    "\n",
    "# ax.set_xlim(left=0)\n",
    "ax.set_ylim(bottom=0, top=None)\n",
    "ax.set_xlabel(\"Day of Week\", fontsize=20)\n",
    "ax.set_ylabel(\"Jobs Submitted\", fontsize=20)\n",
    "ax.tick_params(axis='both', which='major', labelsize=18)\n",
    "ax.tick_params(axis='both', which='minor', labelsize=16)\n",
    "\n",
    "def get_converted_xticks(ax):\n",
    "    days_of_week =dict(enumerate(calendar.day_abbr))\n",
    "    return [days_of_week[tick] for tick in ax.get_xticks()]  # Get the hour number\n",
    "\n",
    "ax.set_xticks(x_vals)\n",
    "ax.set_xticklabels(get_converted_xticks(ax))\n",
    "fig.tight_layout()\n",
    "\n",
    "date_time = datetime.now().strftime(\"%Y-%m-%d-%H-%M-%S\")\n",
    "fig.savefig(f\"num_jobs_day_of_week_{date_time}.pdf\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "submit_day_of_week\n",
      "0    11610.187500\n",
      "1     7191.406250\n",
      "2    13880.093750\n",
      "3    10471.500000\n",
      "4    14753.906250\n",
      "5     7429.193548\n",
      "6     5160.156250\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "y_vals = filtered_jobs.value_counts([\"submit_date\", \"submit_day_of_week\"], sort=False).groupby(\"submit_day_of_week\").mean()\n",
    "y_vals.max() / y_vals.min()\n",
    "print(y_vals)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 792x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# This cell prints the (sorted) CoV in number of CPUs (cores) requested per user\n",
    "data = filtered_jobs[[\"UID\", \"NCPUS\"]].groupby([\"UID\"]).std() / filtered_jobs[[\"UID\", \"NCPUS\"]].groupby([\"UID\"]).mean()\n",
    "data.sort_values(\"NCPUS\", ascending=False, inplace=True)\n",
    "\n",
    "x_vals = np.arange(len(data))\n",
    "y_vals = data[\"NCPUS\"]\n",
    "fig, ax = plt.subplots(figsize=(11,3))\n",
    "ax.bar(x_vals, y_vals, color=\"lightcoral\")\n",
    "\n",
    "ax.set_xlabel(\"Users (sorted by CoV)\", fontsize=20)\n",
    "ax.set_ylabel(\"CoV\", fontsize=20)\n",
    "ax.tick_params(axis='both', which='major', labelsize=18)\n",
    "ax.tick_params(axis='both', which='minor', labelsize=16)\n",
    "\n",
    "fig.tight_layout()\n",
    "\n",
    "date_time = datetime.now().strftime(\"%Y-%m-%d-%H-%M-%S\")\n",
    "fig.savefig(f\"cov_cpus_requested_per_user_{date_time}.pdf\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO: Pandarallel will run on 16 workers.\n",
      "INFO: Pandarallel will use Memory file system to transfer data between the main process and workers.\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e2134163b9bc40c5b7771f401c078eec",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "VBox(children=(HBox(children=(IntProgress(value=0, description='0.00%', max=140529), Label(value='0 / 140529')…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 792x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "data = filtered_jobs[filtered_jobs['is_gpu'] == True][[\"UID\", \"NCPUS\"]].groupby([\"UID\"]).std() / filtered_jobs[filtered_jobs['is_gpu'] == True][[\"UID\", \"NCPUS\"]].groupby([\"UID\"]).mean()\n",
    "data.sort_values(\"NCPUS\", ascending=False, inplace=True)\n",
    "\n",
    "x_vals = np.arange(len(data))\n",
    "y_vals = data[\"NCPUS\"]\n",
    "fig, ax = plt.subplots(figsize=(11,3))\n",
    "ax.bar(x_vals, y_vals, color=\"lightcoral\")\n",
    "\n",
    "ax.set_xlabel(\"Users (sorted by CoV)\", fontsize=20)\n",
    "ax.set_ylabel(\"CoV\", fontsize=20)\n",
    "ax.tick_params(axis='both', which='major', labelsize=18)\n",
    "ax.tick_params(axis='both', which='minor', labelsize=16)\n",
    "\n",
    "fig.tight_layout()\n",
    "\n",
    "date_time = datetime.now().strftime(\"%Y-%m-%d-%H-%M-%S\")\n",
    "fig.savefig(f\"cov_cpus_requested_per_user_ml_jobs{date_time}.pdf\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAxAAAADQCAYAAACX+YfUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeXxU5dn/8c81M9kXyMYOorKIiEsFFXfApZWq1datKq5VW7VFfPr8WutuW2sf91q12laWWrcqdQE3qohWLCBVUFRAZJE9bAGyZ+7fH2cSk8lMMpNMMkn4vl+veQ25zz33uTJzJpzrnHsx5xwiIiIiIiKx8CU7ABERERER6TyUQIiIiIiISMyUQIiIiIiISMyUQIiIiIiISMyUQIiIiIiISMwCyQ4gEQoLC93AgQOTHYaIiIiISJfx4YcfFjvnisLLu0QCMXDgQBYsWJDsMEREREREugwzWxWpXF2YREREREQkZkogREREREQkZl2iC1OylNx9N2737kbllpVF7v/8TxIiEhERERFpW7oD0QqRkoemykVEREREOjslECIiIiIiEjMlECIiIiIiEjONgRARERGRJu3YsYPi4mIqKyuTHYq0kt/vJycnh/z8fNLS0lrUhhIIEREREYmqvLycjRs30q9fPzIyMjCzZIckLeSco6qqipKSElavXs2AAQNalESoC1MrWFZWXOUiIiIinc3mzZspKioiMzNTyUMnZ2akpqZSWFhIXl4eW7dubVE7ugPRCrVTtVZ/9RW7p04la8IEAnvvneSoRERERBKnvLycXr16JTsMSbDc3FxWrlxJ7969435t0u9AmNkQM7vdzD4ws81mttPMPjKzX5lZp7iUbzk5AAR37UpyJCIiIiKJVV1dTSCga85dTUpKCjU1NS16bdITCOBS4DrgS+B24OfAF8CvgffNLCOJscXEMjMBcKWlSY5EREREJPHUdanrac1n2hHSyX8AdzrndtQre9TMlgG/Ai4DHkpKZDGyjAww0wJyIiIiItLlJf0OhHNuQVjyUOuZ0PMB7RlPS5gZlpGhOxAiIiIi0uUlPYFoQr/Q88akRhEjy8oiqARCREREpNO4/PLLMTMmTZoUcfutt97aqKuPmXHrrbc22e7kyZMxM5YvXx53TGbGjTfeGPfr2lOHTCDMzA/cDFQDf49S5wozW2BmCzZv3tyu8UWMJzNTdyBEREREOomysjKee+45AJ588kmqq6uTHFHn0RHGQERyP3AEcINz7otIFZxzjwGPAYwcOdK1Y2wRWWYmwQ6QyIiIiIh0ZCV33x1x3KhlZdVNkd8epk+fTklJCaeccgozZ87ktdde47vf/W677b8z63B3IMzsDuAa4DHn3J3JjidWPt2BEBEREWlWtEln2nsymilTppCXl8fkyZPJyMhg6tSpbbo/5xz33XcfQ4cOJTU1ld69e3PNNddQUlISse5vfvObutW/jz32WD766KMGdV5//XWOPPJIunXrRnZ2NkOHDuX2229v09+hVoe6A2FmtwI3Ak8AVyU3mvhYVhaurAwXDGK+DpeXiYiIiCRU2WuvUbNhQ0Lb3DV5clz1/b16kfHtb8e9n3Xr1jFr1iyuuOIKioqK+N73vscLL7zAtm3byMvLi7u9WPzqV7/izjvv5Oqrr+bUU09lyZIl3HTTTXz88ce88847+OqdP06dOpUBAwbw0EMPUVFRwc0338y4ceNYtmwZ+fn5rFixgtNOO40f/OAH3HzzzaSmprJs2TJWrFjRJrGH6zAJhJndAtwCTAUud84lvVtSPCwnB5zD7dqF5eYmOxwRERERiWLatGkEg0EmTJgAwEUXXcRTTz3FM888w1VXJf4a9tatW7n33nu56KKLeOghb3WCk08+maKiIi688EJeeeUVTjvttLr6ZWVlvPHGG2RleWsqH3744QwePJj77ruPO+64g4ULF1JZWckjjzxCbui8c+zYsQmPO5oOkUCY2c3ArcA04BLnXDC5EcXPV7sa9c6d+JRAiIiISBfXkiv/ADtuuy3qtuyLL25hNPGZOnUqgwcPZvTo0QCccMIJ9OnTh6lTp7ZJAvHBBx9QUVHBBRdc0KD83HPP5ZJLLuGdd95pkECccsopdckDwMCBAzniiCOYO3cuAAcffDApKSmce+65XHrppRx77LH06NEj4XFHk/S+NmZ2NXAbsBqYBfzQzC6o9zgxuRHGpjZpcDt3JjkSEREREYlm/vz5LFmyhDPPPJPt27ezfft2du7cyZlnnsncuXNZunRpwve5detWAHr37t2gPBAIUFBQULe9Vs+ePRu10bNnT9auXQvAoEGDeP311wkGg1x44YX06tWLww8/nHfeeSfhsUeS9AQCGBV6HgBMwbsLUf/xqyTFFRervQMRYSCMiIiIiHis3pX1WMoTbcqUKQDcdddd5OXl1T1quxa1xWDq/Px8ADaEjRmprq5my5YtFBQUNCjfuLHxMmgbN26kb9++dT+PGTOG1157je3btzNr1ixSUlIYP348xcXFCY8/XNK7MDnnLgYuTnIYrWZZWeDz6Q6EiIiISBPac6rWcJWVlTz99NMcfvjh/O53v2u0/brrrmPatGnccccdjRaQa40jjjiCtLQ0nn76acaNG1dX/swzz1BdXc1xxx3XoP7MmTPZvXt3XTemlStX8sEHH/CLX/yiUdtpaWmMHTuWXbt2cfrpp/PVV19RWFiYsNgjSXoC0VWYGZadTVAJhIiIiEiH9Morr7Blyxbuuecejj/++Ebbr7zySn784x8ze/ZsxowZ0+r91SYh+fn5TJo0iTvvvJOsrCxOOeUUPvvsM2688UaOPvpoxo8f3+B1GRkZnHTSSfz85z+noqKCW265hdzcXK677joAHn30UebMmcMpp5xC//79KS4u5s4776RPnz4ccMABrY67OUogEsjfowc1q1fjnEto1ioiIiIirTdlyhRycnI466yzIm4/77zzmDRpElOmTGlVAlFWVgZ4dwdq/eY3v6GoqIhHH32Uhx9+mIKCAiZMmMCdd97ZYApXgAkTJpCVlcU111xDcXExo0aN4umnn67rCnXQQQfx6quv8stf/pJNmzaRn5/P0UcfzZNPPklGRkaL446VdbLZUiMaOXKkW7BgQbLDoHLhQspefpm0448nPexWlIiIiEhn9NlnnzFs2LBkh9GpTJo0iYcffphdu3YRCHTc6/XNfbZm9qFzbmR4eUcYRN1lpAwbBqmpVMyeTU3YaHoRERER6do+/fRT/vKXv/DXv/6VM844o0MnD63RNX+rdvTW4rU88fYXbN5RRlG3DC466lRGvf08wfXr8YduM4mIiIhI13fNNdewdOlSfvCDH3DvvfcmO5w2owSiFd5avJb7ZyymoqoGgE07ynjw32u4wlfACevXkzJ8eJIjFBEREZH28vbbbyc7hHahLkyt8MTbX9QlD7UqqoM85etPzfr1SYpKRERERKTtKIFohc07yiKWbwn6qf76a1ww2M4RiYiIiIi0LSUQrVDULfI0WYUZAais1F0IEREREelylEC0wiVjhpKW4m9Qlpbi5+LjBgFQ/cUXyQhLRERERKTNKIFohbEj+jJx/AhS/N7b2KNbBhPHj+CEUYMI7L8/FXPnEty1K8lRioiIiIgkjhKIVho7oi+H7FPIoF65TPvpWMaO6AtA2uGHQ3U1NevWJTlCEREREZHEUQKRAJmpAUorqxuU+QoLAQhqQTkRERER6UKUQCRAZlqAsoqG07laRgaWnk5wy5YkRSUiIiIi4SZPnoyZYWYsXbq00fbZs2fXbZ81a1Zd+a233oqZUV1d3eg1Tbn44ovp169f3HHWxlE/ho5CCUQCZKT6G92BMDN8BQW6AyEiIiLSAeXk5DBt2rRG5VOnTiUnJycJEXUeSiASIDM1QEVVDTVB16DcV1hI9ddfU71mTZIiExEREel43lq8lgsffItv3zGDCx98i7cWr233GM4880z+9re/4dw3529lZWU8//zzfP/732/3eDoTJRAJkJEWAKAs7C5E2rHH4svOZveUKUoiRERERPCSh/tnLGbTjjIcsGlHGffPWNzuScSFF17IqlWreO+99+rKpk+fTk1NTZsnECUlJVxzzTX06dOHtLQ0hg4dyn333dcgmam1Y8cOLr74YvLy8sjNzeX8889nS1gX+QceeIBhw4aRkZFBXl4eI0eOZPr06W0Wf6DNWt6DZIYSiNKKarLTU+rK/fn5ZF12Gbsee4yyGTPIvvJKzCxZYYqIiIgkzCOvf8qKjSVxv+6zr7dTVRNsUFZRVcO9Ly/i1f+ujqutfXrm8uOTh8cdA8Bee+3Fsccey7Rp0zjmmGMAr/vSGWecQXZ2dovajEUwGGT8+PEsXLiQ22+/nREjRjBjxgwmTZrE5s2b+e1vf9ug/sSJEznhhBN46qmnWLZsGTfccAPr1q3j7bffBuDJJ5/k+uuv5+abb+aYY46hrKyMRYsWsbUNu9ErgUiAjNTIdyAAfJmZpB9zDGWvvEJw0yb8PXu2d3giIiIiHUZ48tBceVuaMGEC119/PQ8++CDbtm1j1qxZvPrqq226z5kzZ/Lee+/xxBNPcPHFFwNw0kknsXv3bu655x4mTZpEYWg2T4Dhw4fzxBNPAPDtb3+b/Px8LrjgAv71r38xbtw45s6dy4EHHsjNN99c95pTTjmlTX8HJRAJkNlEAgEQGDIEgOply5RAiIiISJfQ0iv/Fz74Fpt2lDUq79Etg/+bMLq1YcXlrLPO4pprruHll19m1apV9OrVi3HjxjFnzpw22+ecOXPw+Xycd955DcovuOAC/vKXvzB37lxOPfXUuvKzzz67UcwTJkxg7ty5jBs3jlGjRvHwww9z7bXXcvrpp3PkkUeSmZnZZvGDxkAkRG0Xpt0VkRMIX04Ovl69qFq+vD3DEhEREelwLhkzlLQUf4OytBQ/l4wZ2u6x5OTk8L3vfY9p06YxdepUzj//fHy+tj093rp1K/n5+aSlpTUo79WrV932+nqGXXxOTU0lLy+PtWu9MSMTJkzgkUce4T//+Q8nn3wy+fn5nHnmmaxcubLNfgclEAmQk+GNe9hZWhW1TsqgQdSsXo0rL2+vsEREREQ6nLEj+jJx/Ah6dMvA8O48TBw/grEj+iYlngkTJjBjxgwWL17MhAkT2nx/+fn5bN26lcrKygblGzZsAKCgoKBB+caNGxv8XFlZybZt2+jb13u/zIwrr7ySefPmUVxczJQpU5g3bx7nnHNOm/0OSiASoFtmKgA7Siui1gkMHgzOUf3VV+0VloiIiEiHNHZEX6b9dCyv3TSeaT8dm7TkAeDEE0/k7LPP5qqrrmL48JZ1y4rHcccdRzAY5LnnnmtQ/uSTT5KamsoRRxzRoPzZZ59t8PNzzz1HMBhk9OjG3b3y8vI455xzOPvss/nkk08SH3yIxkAkQE5GKgZsL62MWqd27ENw27Z2ikpEREREmuP3+3nqqadirv/CCy806ubUu3dvjjrqqKivqT8L53e+8x2OPvporrrqKjZv3szw4cOZOXMmf/7zn/nlL3/ZYAA1wKeffsoll1zCueeey9KlS/nVr37Fcccdx7hx4wC44ooryMnJYfTo0fTo0YOlS5cybdo0TjrppJh/p3gpgUgAv8/IzUxlRxMJBKmp4PPhyhoPGhIRERGRziFS16Dx48fzyiuvRKxfVlbWYLyDz+djxowZ3HDDDdx1111s2bKFgQMHcu+99zJx4sRGr3/ggQd46aWXOOecc6ipqeHUU0/lwQcfrNt+1FFH8cQTTzBt2jR27NhBnz59uOCCC7jtttsS8NtGZpEWrOhsRo4c6RYsWJDUGH70yDsMKMzmprMOjVqn5J57CAwZQma9kfUiIiIiHdlnn33GsGHDkh1Gp/Wtb32L/Px8Zs2alexQGmnuszWzD51zI8PLdQciQbo1dwcCsMxMXGlpO0UkIiIiIsny7rvvMmvWLP773//y6KOPJjuchFICkSDdMlNZXbyryTqWkaEEQkRERGQPMGbMGPbee2/uuusufvSjHyU7nIRSApEg3bJS2bG6+TsQweLidopIRERERJKlujry+mBdgaZxTZDe3TPZUVrJlxtKotbxqQuTiIiIiHRySiAS5NuHDCA7PcAjr39KTTDywPTaLkxdYeC6iIiI7Dl07tL1tOYzbTaBMLN7zWy/Fu9hD5GTkcKVJ+3P4tVbee+z9RHrWGYmOEfVwoXtHJ2IiIhIy6SkpFCmaei7nPDpZeMRyx2IicCnZjbHzC4ws/QW7WkPcPzwPgCs2bI74vbAoEFgRvl777VnWCIiIiIt1qNHD9auXUupelF0es45qqqq2Lp1K19//TUFBQUtaieWQdQ3AZcCRwNHAQ+a2TTgcedc262R3QmlBvzkZ6exeUfkLN1fVETaccdRMXs2rrISS01t5whFRERE4pObmwvAunXrqKqqSnI00lqBQID09HQGDBhAenrL7gs0m0A4534D/MbMxgE/Ak4HrgWuMbMPgMeBZ5xzurcFFOVmsKkk+lvh79EDgJriYgJ9+rRXWCIiIiItlpubW5dIiMQ8iNo59y/n3LlAX+B64DNgNPAXYJ2ZPWRmB7dNmJ1HUW46m6LcgQDwFRUBENy4sb1CEhERERFJmLhnYXLObXXO3eecOwCvS9MUvDsZPwE+NLN5ZnZZguPsNHp0y2DzjrKofQR9+flYVhZlr75K6fTpVH/9dTtHKCIiIiLScq1aSM45NxeYa2Y/BX6Il0SMBL6Fd2dij1PULYOK6iA7SivpntV4ZLv5fGRfdhnl775L1WefUbVoUcR2LCuL3P/5n7YOV0REREQkLolaB2IQcCCwV+hnS1C7nc6AwmwAVm3eFbWOLy+PzNNOI3fixKh13O7IMzmJiIiIiCRTixMIM8sxsyvNbD7wId7dh1RgKnBMguLrdPbukQPAV5uir0hdy1o4966IiIiISLLE3YXJzEbjzcZ0FpCJd7dhEd5sTH9zzu1IaISdTH52GrkZKazY2HwCISIiIiLS2cSUQJhZPjABuBwYhpc07AYmA4855/7TVgF2NmbG4D7def2jrzlyaC+OGNIz2SGJiIiIiCRMs12YzOwpYC1wD7A/8BFed6U+zrnLlDw09rNTDqAwN50n5yxLdigiIiIiIgkVyxiIc4Aq4M/AKOfcoc65R51zO9s2tM6rZ/dMzh69D0vX72i2K5NlZcVVLiIiIiKSTLF0YboS+LtzTtMCxeG44X149I0lzFmynn16Rl+5MXyq1t3PPkvNqlXkTJrU1iGKiIiIiMSt2TsQzrnHoyUPZpZrZv3NTGubh+melcaBAwt4/oMVLFq1JebXpR58MK60lN1TplC5cCEuGGzDKEVERERE4hP3NK5m5jezX5jZcmAbsBLYZmbLQ+WtWpyuK/nxScMJ+Hw8/e8vY35NYPBg0sePx+3cSdnLL1M5f34bRigiIiIiEp+4EggzSwXeBH4DDATWAPNCzwND5bNC9WJt85dm9pyZrTAzZ2Yr44mpIxvYI4exI/qwZM1WamK8k2BmpI0cSfa11+Lv14/KefNwNTVtHKmIiIiISGzivQMxCTgemAEMc84NdM6Nds4NBIYCL+MtIhdPB/7fAmOBL/HuaHQpI/YqoKyyhmXr41sXwsxIO+YYglu3UvbyywS3b2+jCEVEREREYhdvAvFD4BPge865BnOUOue+BM4EPgXOj6PNfZ1zBc65E4F1ccbT4R08sAAD5i/fFPdrU4YMIe2YY6j6+GN2PvAA5bNmEdytsewiIiIikjzxjlcYBPzBORexP45zLmhmrwLXxtqgc25FnDF0Kt2z0ti/fx5vLvqaMw7fm+z0lLhenz52LIH99qPivfeo+Pe/qfzwQ/z9++Pr3h1fQQEpw4fjy85uo+hFRERERBqK9w5EJdDc2WoW3roREvKdQwawcXsZVz46J64ZmWoF+vQh6+yzyb7qKgL77kuwpITKjz6i/LXX2PXoo9Rs3twGUYuIiIiINGbOudgrm83BG+twgHOu0VmrmRXidXFa6pw7Nu5gzD4BskNjKmI2cuRIt2DBgnh3166+WLedu6Z/xMbtpTz+4+Pok9+6heKccwQ3bmTX5MmkDB1K5hlnJChSEREREREwsw+dcyPDy+PtwvQQ8DQwz8x+DbwNrAd64Q2uvhEoAn7aqmhjYGZXAFcADBgwoK1312pD+3Tn9xcewYQ/vMVLC1Zx1Un7t6o9M8Pfqxepw4dTuWgRJcuX40pLG9fLymq0WJ2IiIiISEvFlUA45541s4OBXwCPRahiwO+dc88mIrhmYnmsNoaRI0fGfhsliQpz0xlzQB9enPcVqX4fF40Zgt8X91IcDaSOHk3lp59GTB4AnAZdi4iIiEgCxX326py7ATgS+CvwX2BF6PmvwFHOuV8kNMIu5tpTRnDovkU88/6XvP/5xla35y8sJOuCCxIQmYiIiIhI81p0+ds594Fz7kfOuZHOucGh5x855+YmOsCuJj3Fz23njKRbZiqvfbSGYBxjUKIJ9OvX5PaaLVuIZ6yLiIiIiEg0zSYQZpZmZvPM7F9mFnUOUjNLDdX5oKl6An6fj3EH9mXBl5t5fNZnbb6/XQ89xM7776dsxgyqV6zQytYiIiIi0mKxjIE4HzgUONU5F3V6VudcpZn9HzAz9JrJsQRgZhcCe4V+LAJSzezG0M+rnHPTYmmns7l83H5s3F7GjAWrOOfIfemeldZm+0ofP57qFSuo/PhjKhcswLKzyTzjDAL77NNm+xQRERGRrqnZaVzN7BVgqHNucEwNmn0BLHfOjY+x/mzguCib33HOHd9cG51hGtdIVm3eyU8ee5fC3HSqg44tJeUUdcvgkjFDGTuib1xtldx9d8QB0/VnYXJVVVQvX075W28R3LmTjBNPxLp1I9C/P5aWFlMbIiIiIrJnaM00rofg3VWI1RzglFgrx5IgdFV7FeVw2qiBvPCfr+rKNu0o4/4ZiwHiSiJiOcG3lBRShg3D36sXu/78Z8peeSW0wfD36RN1xibN5CQiIiIitWJJIAqBeKYL2ggUtCycPc97n29oVFZRVcMTb38R912IWPny8siZOBFXWkpwyxaqV62i6vPP22RfIiIiItK1xDILUxmQHUeb2UB5y8LZ82zeURZXeaJYSgq+bt0I7LMP6WPGkH3llU3W1yxOIiIiIgKx3YFYA4yKo82RwOqWhbPnKeqWwaYIyYLfZ8xcuBq/z0jx+xjapzu98zPxmbVJHNbMgnblr75K6mGH4S8sbJP9i4iIiEjnEEsCMRv4iZmNdM41OVLZzA7FW2TuDwmIbY9wyZih3D9jMRVV30ytmuL34ffBA6GxELX27pHDreeMpFf3zPYOk8r586mcPx9fUREp++2Hv29ffEVF+PLysDZKakRERESk44klgXgI+DHwnJmd4pyLuHCBme0HPAfUAA8nLsSurXacwxNvf8HmHWV1szAdtV8vSsoqqQk6yiqq+WjlFv42ZykT//o+t587kiF9uic8FsvKijoLU/aVV1K1eDFVy5dT8e67ddt8+fkE9t2XlBEj8Pfpg/n9CY9LRERERDqOZqdxBTCzm4FbgUrgH8BbwNeAA/oB44DvA2nAzc65X7dRvBF11mlc47V6805ufGo+20srOfngfvQryObUkXvVdWt6a/HaRolIWwzEDpaVEdyyhZr166n+/HOq16yBqiosN5fUgw6icuFCTQcrIiIi0slFm8Y1pgQi1MANwC1ACl7i0GAzUAXc6py7s5Wxxm1PSSAAtu4q584X/svy9SWUVlYzYkA++/XtTnFJOe9+vp7qmm8+mrQUPxPHj2iz2ZxqBcvKqF6yhIr58wlubHrCrm633NKmsYiIiIhIYrQ6gQg1shdwKXAU0BsvcVgHvAc84ZxblZhw47MnJRC1nHNMn7eSVxasYtOOMqpqghHr9eiWwbSfjm23uKpXrmT3lClRtyuBEBEREekcEpJAdFR7YgJRn3OO7/x6ZqPbQrX6FWTxrX0KGVCYTa/umWSlp9CzWwYFOeltEs+O226Lui31iCNIHzcOC8Qy/EZEREREkqU1K1FLB2dmUaeDzUwLkJeVxqyP11JaWd1g26H7FHLaqIGMGJBPVnoK0PbjKCo/+ICqzz8n0K8fgX33xTIysMxMfNnZWPfudTM6ldx9t8ZRiIiIiHRASiC6iEjTwaal+Ln2OwcwdkRfnHNsLimneGc5u8urWLZ+B/+ct5JbnllAasDH8P75ZKcH+M+yTVRWe92hNu0o4/7QVLKJSiIyzzuPynnzqF6+nKpPPmmwzd+7N2njxuHLy4uYPABRy0VERESkfSiB6CKiTQdbW25m9OiWQY9uGQCMGtSDMw7fm8/Xbuffn2/gk9Vb+e9XxY3araiq4Ym3v4grgWhqOtiUIUNIGTIEV11NsKQEV1aGC83qVP7225T+7W/Nth/ctg3LzdWUsSIiIiJJoDEQUufkO2ZELDfgtZvGt/n+g6WlBDduJLhjB2UvvthsfX+/fqQecgiBffbBunXTgnYiIiIiCaQxENKsHlHGUQT8Pj5ZvZVBvbuRntJ2V/19mZn49t4boMkEIuP00wlu20bl/PmUvfxyKMgA/l698PfuTWDQIAKDBmE+n8ZSiIiIiCSYEgipE2kcRcBv1ASDXD9lLgGfsX//PPoVZHPywf3Yr29eUuJMPfhgANKOOYbgli1Ur17tLWy3YQOVCxdSOX8+lpmJv3dvjaUQERERSTAlEFIn2jiKkfsW8emabXy6ZisfrdzC25+sZebC1RTkpHHzWSPZr2/3hMfS1DiKun8HAvh79sTfs2ddmauooGrZMqq//JKatWsTFo/uZIiIiIh4NAZC4lZaUc2bH6/hr299Qe+8TG4/dxRZ6QEyUwMdbhxCU2tSWHa290hN9dalSEnB1707vm7dsOxsfDk5dc8lv/td1Ha0OJ6IiIh0RRoDIQmTmRbg9MP2pndeFrc+u4ALH3wLAL/POHTfIsaN6Esw6Np0PYlECAwahNu9G1dVhausxO3cSfWKFVBVFVc7VV988c1aFtnZWEpKxHq6iyEiIiJdgRIIabHDBvfg/kuO5It12ymvqqG4pJwZH65m3rJNpPh9VNW03VV7M+EAACAASURBVHoSiZB5+umNypxzUF5OcNcu3M6d3vOuXZS/+WbUdkqffrphgc+Hr0cP0kaPxrKyvDscaWkajyEiIiJdghIIaZUhfbozpM83YyAmHD+EH/zfG3XJQ62WrCeRCLGMpWhQbgYZGfgzMqCoqK68qQQi6/LLcaWluF27CO7aBRUVVH7yCWXTp8ccp3Ouw3X/EhEREYlECYQkVFZaCtGG1WzaUcbPp86lKDeDo4f14sihvdo8nvboGhTo2zgpSjv+eILFxV7XqIoKqKyk9B//iNrG7smT8ffu7Y2/yM3Fl5vr/Tsnp1Fioa5QIiIikkxKICThiqKsJ5Ge4qeqJsj85Zv41+K1TPzuCA4ZWEjP7hkd/up73HcyQutSNNBEAuFKS6lcuLDR+At/nz6kHn64ukKJiIhIh6EEQhIu0noSaSl+fjZ+BGNH9KWiqoZr//Ie97/ijYvISPVz0MBCTjiwL726Z1KQk0b3rDRmf7KuwwzEbusr+zlXX/3N+IuSEu9RXEzF++/H1RVKREREpK0pgZCEi7aeRG15WoqfBy87muXrd/DVphK+2rST9z7bwAdLN9a1kZHipyoYpLrG6w/VUQdix6O5uxj1x1/4e/aEwYNJPewwgps313WFchUVlD3/fNR9VC5ciH/AAG9GqPT0NvtdREREZM+ldSCkQ6isrmHFxp1s3VlO8c5y/vTmkrrkob6C7DT+dNVxZKUH8HXwbk9tpam1Leqz3Fz8PXrUrXXh79+fQN++WPfu7LznHo2jEBERkSZpHQjp0FID/gYrWj/82qcR623ZVcEP7n4Dn0FORiopAR8A+/TI4afjR1CUm9Eu8XZUWZdeSnDbNq8L1KZN1GzejNu0CVdeDvPmeZVSUqKudaFxFCIiItIcJRDSIUUbiJ2bkcJ5xwympLSSkrJKqmuC1AQdsz9Zx30vL2LSqQdRkJPWYFD2W4vXdpixFInQVFeoQP/+0L9/o20uGKTm668Jbt5MzcaNVM6f3x6hioiISBekLkzSIb21eG3EgdgTQwOxw700fyV/DN21OHCvfPbpmUtWWgrrt+1mzmfrG3SHaqqdPUVT3aD8ffrgKywkZf/98RUW4svLw3y+doxOREREOgJ1YZJOpbmB2OFOGzWQYf3y+GDpRv61eC3LN5RQWlEdsW6yFrXrLCwzk+qlS6latMgrCATw5eWRMmwYgUGD8PfrV3eHR2tSiIiI7HmUQEiHNXZE37hO8gf37sbg3t248LghANQEHeN/M5NI99g27SjT6s9RZJ1/Pq662uvytH07NRs3UrNhAxVz5lAxZw6WkYF164ave3etSSEiIrIHUgIhXZbfZ1HHUgBMmjyXkw7uR15WGnnZaeRlpdE9K5XUgL+dI21/zU4pGwgQGDiwwbZgaSnVy5ZRs2aNN0h7y5Ym91H6z396yUa9hy87G3///pi/67/HIiIiXZXGQEiXFmksRWrAx3H79+bDFcVs3VXR6DVZaQEvmchOoyA7jZMP6c+h+xR1ucHYidDUWArr1g1XWtp4xiczfEVFBPbdl8A++1D2z3+qG5SIiEgHpDEQskdqaixFdU2Q4p3lbN9dwbZdlWwvrWDbrgq2765k2+4Ktu+uYPHqrbyzZD0HDsjni3XbqagOAl1jYbu2ljtxIgCuuhpXXo4rLSW4ZQvVX39Nzfr1VM6bR+XcuVFfr25QIiIiHZMSCOnyoo2lCPh99OqeSa/umVFfW1ldw9/eWcYz73/ZaFtFVQ1/fetzJRDNsEAAy86G7Gz8PXqQMmwYAK6qiupVqyh98smor939978TGDyYQP/++IqK1PVJRESkA1ACIdKE1ICfS8ftFzGBANhcUk5ldc0eMW4ikubGUjT52pQUUgYNarJOzZo1VC9b5v2QkoK/d2/8PXoQGDLEW2U7KwsLeH/GNCOUiIhI+1ACIRKDHk0Mxr7s4XcYPaQnl5+w3x6XSLT1iXnu//t/1GzYQE1xMTWrV3uL4H38MZX1xjz5Cgrw9+2rGaFERETaiRIIkRhcMmZoxIXtxh86gBUbS3hx/kpenL+SgM84aO9CDtm7AJ8Z2ekpjDmgT11ioYHY8fP36oW/Vy844ADAG1NRvWoVbscOgjt3UrNuHdUrVzbZRk1xsbcgnrpAiYiItJoSCJEYNLew3QdLN7J8Qwm7yqt47b+r+fDLzXWvfXnBKob3z2NXWSWzl3yzKrYGYnvi7QZlgQAp++7bqLypGaF2/fGP3uxP3bvjy8/3HgUF+PLzCfTrh2VktPwXEBER2cNoGleRBKuqCVJVHcTheP/zjTz7/pdsLimjrLImYv2i3HT+9rNx7Rxl19NUApFx+ukEt24luHUrNVu3emtYVFZ6G1NS8PfqRWDAACoWLoSyxl3VNI5CRET2RJrGVaSdpPh9pPh9AJx4UD9OPKgfQec45deRV8XeXFLOVX+ag99n5GSkcubhe9M7L5PC3HQyUvUVTYTUgw9u8LNzDrd7N8HNm6lctIjgli1UzJ0LwWDE12schYiIyDd0diLSDnwWfVXsjFQ/PbtlEARWbCjhpqfn123LzUhhYI8cbjtnFJlp+ro2JZ6uUGaGZWfjy84msPfeALiyMkp+//uo7Qd37sSXk5O4gEVERDopdWESaSeRVsVOS/EzcfyIujEQ5ZXVLF69lZLSSop3lvPlhhLeWbKewwb34LSRe3HgXgWkpfg1GLuNNNUNCp+PlGHDCAwdir+oCF9hYd0UsuE0payIiHQF6sIkkmTNDcQGSE8NMGpQjwav26toGc++/yXzlm0CYEBhFhu2l1GpVbHbVephh1H53/9S9emnXkEggL9vXwIDBpB6xBH4Mr9ZkFBTyoqISFemBEKkHUVbFbsp5x87mLOO3IePvtrCp2u28sz7XxJ+47CiqoYn3v5CCUQbyjj5ZNLHjSO4ZYu3LsXXX1OzZg0V771H5aJFBPr1g5SUqHclREREugr9TyfSCaQG/Bw2uAeHDe7B0/+Osip2lIXuJHbNjaOwQAB/z574e/aE4cMBqF69mvI33qBm/XpcdTVUVze5j4q5c7GcHPwFBfh7945aT92gRESko1ICIdLJRFsV2wxmLlzNQQML6NU9A7/Pl4ToOreWnJgHBgwg+/LLG5Q1NZai/I036v7t69ULf8+e3iJ3mZn4Cwvx5edjubnqBiUiIh2WEgiRTibSqtgpfh8FOWk8EBoLkeL38e1D+nPwwAIOH9KzblpZSb6c//1f3M6dVH/1FVWffkr18uWNkwJ1gxIRkQ5M/0uJdDLRBmOPOaAPX6zbzuriXXy8cgszPlzFywtWMWpQERPHH0hhbnqSI99zNNUVypeRARkZ+Hv0IO3wwwFwNTXeuhTFxd5Cd8XFVP7nP1Hbr1m/Hl+vXphZm/0OIiIi0SR9Glcz8wE/A64EBgKbgWeBm51zMd2r1zSuIo1VVNXw1HvLeeq95QAcNLCAG848hO5ZaUmOTGLR5JSyQGCffUgbN45Anz7tFJGIiOxpok3j2hH6NdwH3AssAa4FngN+CrwcSi5EpAXSUvxcdPwQ7r14NBceN4SPV27htf+uSXZYkgDpJ51E9bp17H78ccrefJNkXwgSEZE9S1K7MJnZcLyk4QXn3PfrlX8FPAicC/w9SeGJdHpmxvD++Qzvn8+C5Zt4Z8l6zj16EEDCFqNLRDtaGK+xprpBpY0eTeohh1D+5ptUvv8+VZ98gi8/H1+3bvi6dydl2DBvpigSM5tTomaEUiwdvx3FolgUS8eJpSNL9hiI8wAD7g8rfxz4HXABSiBEEuK44X149I0lrC7exfL1OxoMxG7pYnThq2u3pJ1EtNEVNfefjKWnk/7d7+Lfay+qly0juGMH1V99hSspoeLdd8m68EICAwcmZDanRM0IpVg6fjuKRbEolo4TS0eW7ARiFBAE5tUvdM6Vm9lHoe0ikgDH7t+bP72xhPtfWcSXG0oazOIE3piJB2Ys5uNVW2Juc/Yn61rdTiLaqK9PXibnHDUo7td1RmZG6oEHknrggXVlwV272D1lCqXPPENg2LAmX1/60kutjiERbSSqHcXStu0olrZrI1HtKJa2ayNR7cTTRvoxx+DLy2v1PttCUgdRm9lioIdzrmeEbc8CZwFpzrnKCNuvAK4AGDBgwKGrVq1q63BFOr37Xl7E/C83sWVnRdQ6BTmxD7JORDuJiqXW4N7due2cRuO99ig1W7dS+o9/4Hbtwu3cGbWe5eTE1F4i2khUO4qlbdtRLIpFsXScWLLOPRd/kifKiDaIOtl3IDKBaGcP5fXqNEognHOPAY+BNwtTm0Qn0sVcd6p3pfrCB9+KuBhdj24ZTPvp2JjbS0Q7iYpFvuHPzyfniiuApmdzyp00Kab2EtGGYukc7SgWxaJYOk4sHVmyZzkqBaJdYkyvV0dEEuiSMUNJS/E3KEtL8XPJmKHt3k6iYhEREZH2kewEYh1QaGaRkoi+QHGk7ksi0jpjR/Rl4vgR9OiWgeFd7Z84fkTcg5YT0U6iYpHILCsrrvK2akOxdI52FItiUSwdJ5aOLNljIH4N/Ao41jn3br3ydGALMMc5953m2tFCciIiIiIiidVRF5J7BnDAxLDyH+GNfXiy3SMSEREREZGokjqI2jm32Mz+CFxjZi8AM4FheCtRv4PWgBARERER6VCS2oUJwMz8eHcgrgAGAsV4dyZuds7tirGNzUCy53EtxItdJBY6XiQeOl4kXjpmJB46XiSavZxzReGFSU8gugozWxCpj5hIJDpeJB46XiReOmYkHjpeJF7JHgMhIiIiIiKdiBIIERERERGJmRKIxHks2QFIp6LjReKh40XipWNG4qHjReKiMRAiIiIiIhIz3YEQEREREZGYKYEQEREREZGYKYEQEREREZGYKYFoITPzmdl1Zva5mZWb2Rozu8fMspIdm7Q9MxtiZreb2QdmttnMdprZR2b2q0jHgJkNNbN/mtk2M9ttZu+a2dgobevY2gOYWaaZfWVmzsweirBdx4xgZvlmdreZLQ99tpvN7G0zOyasno6XPZyZZZvZDWa2OPR/UrGZvW9mF5uZhdXV8SKtogSi5e4D7gWWANcCzwE/BV42M72vXd+lwHXAl8DtwM+BL4BfA++bWUZtRTPbF3gfGA38PlQ3G3jdzE6I0LaOrT3D7XirvzaiY0YAzGwv4EPgIuAfwE+A3wIrgb716ul42cOFPrdXgTuA+cD1eP8f+YEngN/Vq6vjRVrPOadHnA9gOBAEng8rvxZwwA+THaMebX4MjAS6RSj/degYuKZe2bNADXBwvbJsYBVe0mH1ynVs7QEP4FtANTAp9Lk+FLZdx4weAO8Ca4DezdTT8bKHP/CSAQfcF1aeCqwAtut40SORD2WOLXMeYMD9YeWPA6XABe0ekbQr59wC59yOCJueCT0fABC6xXsaMNs591G91+8C/gwMAUbVe72OrS7OzPx4n+drwAsRtuuYEczsWOBo4PfOufVmlmJmmRHq6XgRgNzQ87r6hc65SqAY2A06XiRxlEC0zCi8jHxe/ULnXDnwEQ2/fLJn6Rd63hh6PhBIA+ZGqPtB6Ln+8aJjq+u7DtgPuCbKdh0zAnBK6Hm1mb0MlAG7zWypmdU/adPxIuB9ntuB/zWzs8xsQGicw53AocCtoXo6XiQhlEC0TB+g2DlXEWHbWqDQzFLbOSZJstCV5Zvxuqb8PVTcJ/S8NsJLasv61ivTsdWFmdnewG3A7c65lVGq6ZgRgKGh58eBfLxxEJcBlcA0M7sktF3Hi+Cc24Z3Z2ErXhelVcDnwNXA951zj4eq6niRhAgkO4BOKhOI9GUCKK9Xp7J9wpEO4n7gCOAG59wXobLaLgeRjpfysDq1/9ax1XU9AnyFNyAxGh0zApATet4JjAl1RcHMpuP1af+tmU1Bx4t8YxfwCfAS3iDpfLwE4u9mdrpz7k10vEiC6A5Ey5Ti3QKMJL1eHdlDmNkdeF1SHnPO3VlvU+1xEOl4iXSs6NjqokLdTk4CrnLOVTVRVceMgNdlCeCp2uQB6q40vwT0wrtLoeNFMLMReEnDm865nzvnpjvn/oI3jmYD8HjoLrmOF0kIJRAtsw7vtl2kL1VfvNt9ysb3EGZ2K3Aj3lR5V4Vtrh3Q1pfGasvq30rWsdUFhT7Pe4GZwAYzG2Rmg4C9QlW6hcq6o2NGPF+HnjdE2LY+9JyHjhfxXId3Qv9c/ULnXCkwA+9vzUB0vEiCKIFomfl4791h9QvNLB04GFiQjKCk/ZnZLcAtwFTgcue8+e3qWYx3+3d0hJcfEXquf7zo2OqaMoAiYDywrN5jdmj7BaGfL0fHjHhqB632i7CttmwTOl7EU3vy74+wLVDvWceLJIQSiJZ5Bm/+44lh5T/C6wv4ZLtHJO3OzG7Gm9liGnCJcy4YXic0Nd7LwPFmdlC912bjnSwuo+HsFjq2uqbdwFkRHj8JbX8t9PNLOmYk5J944x8uCH32AJhZb+B7wDLn3HIdLxKyJPR8cf3C0F3N04FtwJc6XiRRrPEFU4mFmf0Br8/7dLxuCcPwVmb8NzA20smkdB1mdjXwELAauAlvmrv6NoYGrBHqqjIPqMJb0bME74/vCGC8c+71sLZ1bO0hzGwg3qDqPzrnrqlXrmNGMLMrgD8BnwJ/xVsU7MdAb+C7zrk3QvV0vOzhQquWL8Tr1vYk3meZj3ccDASuds49HKqr40VaL9kr2XXWB95twuvxVm2swOszeC+QnezY9GiXz38y3lWZaI/ZYfWHAS/izdNdCrwHnBClbR1be8gD7z/2RitR65jRo95neybe/Py78e5IvAEcpeNFjwif677AFLzxM1V4icEc4EwdL3ok+qE7ECIiIiIiEjONgRARERERkZgpgRARERERkZgpgRARERERkZgpgRARERERkZgpgRARERERkZgpgRARERERkZgpgRARERERkZgpgRCRPYKZDTGzSjP7ebJjSQYzO97MnJndGmP9i0P1L27byETanpn9wcy2mVlhsmMR6QqUQIjsAeqdPM5uos7AUJ2V7RdZu7oX2AL8MdmB1DKz2Wam1TybYGYnmtmTZvaVmZWaWZmZLTezaWb2nVa0+/fQ8f7jGOq+Gar7vZbuT5LuN0AacGuS4xDpEpRAiEiXZ2ZHAuOBPzjnSpMdjzTPzHLMbDrwBnAmsAR4BHgA+BA4BZhpZne3cBePhZ5/1EwcA4FxwHrglRbuS5LMObcBmAxcaWYDkhyOSKcXSHYAIiLt4GogCExLdiDSPDPzAc8BJwNvAxc459aF1UkDrgKGtGQfzrnZZrYUOMTMvuWcWxil6mWAAU8456pbsi/pMKYAPwauAG5MciwinZruQIhIk0JXgm8ys0/MrMTMdprZl2b2jJkdGqH+4Wb2DzPbEBpzsMbM/mRmfSLUnR3qGpJqZjeb2RdmVmFmk0PbU83sp2a2MNR/udTMVprZi2Z2Qozx5wI/AN53zq2JsH0fM3ss1C2mzMy2mtliM3vUzArC6qaZ2S/MbFEolhIze9fMzo7Qbm2XsMmh8RfPmNkmMwvWji8AjgvVdfUes8Pa6WdmD5nZitB7s8XMXjKzUVF+355m9hcz2xj6fT4ys4tiea+iMbPxZva+me0OfQ7/MLPBYXWeDsV/bJQ2fhDa/ocYdnkeXvKwHDg1PHkAcM5VOOceACaF7Sfmzwh4PPQc8S6EmfmBSwAH/DmGuKOqd6ynmdmvQ12yKkLfpVvMLDXK6/YLHUNrQvU3hrpfDY1Qd3JoH/uY2bWh96As/JhqIsZ+ZvagmS0zs/LQd2Gemd0UVi9qd8h6MQysV9bUd+F4M/s89Lci4viE0OfpzOzqCPHG/N1wzv0HWAlcamYWy3siIpHpDoSIRBX6T/Y14EhgLt5JVDXQHzgeeBevO0lt/UvwTsoqgJeANcBg4HLgVDM7wjm3OsKungdGAa8C/wQ2hcon451MfgJMBcqAPsDRwLeBWTH8GscCqcB7EX6/3sB8IBeYGYojHdgbuBB4CG/cBKETvNfxTvo/xxtLkYmXnDxjZgc7526IsP99gf8AS4EngQxgEXAbcDGwV+jftVbWi+9beF148kP7fgEoBL4HvGdmZzjnZtarXwC8D+wT+n3fA3oDj4baaYkzge8A04HZwMHA94ExZnakc+6LUL2HgXOAK4E5Edq5IvT8WIRt0ere7Zzb3VRF51xF7b9b8BlNwesb/0Mzuz5C97bvAH2BN51zX8UQdyyexTvW/wFUAafj9csfaWanOefqxsSY2bfxPvMU4GW8hKof3mcy3szGRLlz8gBwDDAD77iuaS4oMxuJ997l431+L+C9d/uH4rsj/l+1kUjfhRK8z+G3eN/1SAnmBKASeLpevHF9N+r5N3A+MBzv74qItIRzTg899OjiD7yTfQfMbqLOwFCdlfXKRoTKpkeo7wPy6v08BO8/+eVA37C6Y/FOYqaHlc8Otb8IKAzb1g2v29ECwB9h/wUx/u6/C+3j+xG2XRva9rMI27KAjHo//zJUdyYQqFfeA++k3wFHRng/HfDbKLHN9v4MR9wWCL2X5cBxYdv6AGvx+uWn1St/LLS/+8Lqj8Q7WXXArTG+bxfXi/+7Ydt+Fir/V1j5J6F4wz/LvUOf5b9j2G8ALwF1wKA4j/O4PqPQtmdC5RdHaO/F0LYfxPudi/ZZ45081//epOMl5w64sF55HrANKAb2D2trOLALWBhWPjnUzlpg7zhiSwW+Cr32hxG29w/7OerfknoxDIz1u4CXpNUACyJsGxV63fOt+W5EOHZ/0trPVA899uSHujCJSCzKwgucc0Hn3LZ6RT/Gu1L6M+fc2rC6b+HdkTjVzHIitH+Tc644fBd4fc8r8E4+w/e/JcbYawdMrm+iTqTfb7dzrn75paGYJrl6feGdc5v45urs5RHa3kjDOwyxGo93xfYPzrl3wmJbB/we6IU3wBczS8G7srqTsJlmnHML8K74tsRbzrnwwcMPAV8CY81sr3rlj+DNdBPeZeoKvM/yTzHsLx/vhBbg6zhjbcln9Fik8tDdqVPwPr8X44yjKXfU/94458rxEh/w4q81AegO3OKcW1K/Aefcp3h3+g4xs/0j7OP3Lr47JqfineS/5Jz7e/hGF6HrXwtF/C6E/l78CzjUzIaHba49lqbUK4vruxFmQ+hZA6lFWkFdmESkKUuAj4DzQieKL+J1i1ngnKsMqzs69HxclD7IPQA/3p2KD8O2zQuv7JwrMbOX8U5uPjKz5/G6TP3HxTeTUu04hm0Rtr2E13Xij2Z2Ml5XiH8DS5xz9buS5ACDgLXOuc8jtPNW6PmQCNs+dvW62cSh9v3cyyKv3VA7BmEY3hX3/fC6nLzrnNsRof5sGp/Yx+Kd8ALnXI2ZvYd3EncIsCq0aSreHZ8rgHugLrG5GO/9fzaG/bWob3orPqO38JKho8xsmHPus1D5JXj/R052zlW1JKYoGr2feMd1dVhstZ//QVE+/9rB48Pwvqf1Nfo+NeOI0POrcb4uXk19FyYDJ+Ido/8LdV3SzgU24x3jteL9btS3NfSs9SBEWkEJhMieofYKflN3HWu31V3tD50ojgVuxutHfldo004zmwL80jm3K1RWe6Le3EJt2RHKNkQoA69P/f8Dfsg3Vy7LzewfwP845zY2sy/45u5CevgG59wqMzsM74r9t/H6lgOsMbO7nXMPhn7uFnqOdhejtrx7hG3Rfrfm1L6fZzVTr/b9rI0x2nvS0jiaa692vzjndprZ34CrQv3z38br498LuD90tb05W/C6wqXidW35MsY4W/QZOeecmf0ZuBPvLsT1obE/tXczWjV4OoJG72foe7YFL8muVfv5NznNLPF9n6KpfU/WNlmr9ZqKazreeIgLzOyXzrka4Lt478P9ruEMWPF+N+rLCD03uusoIrFTFyaRPUPtFemCJurUXpHbXr/QObfNOXedc64/3wyI/hy4Bq/LSvg+ujnnrIlHpCvaERdTc86VOedudc4NwetycAHeHZAL8AahxqJ2QHbE390595lz7pzQ9pHAL/D+Nj5gZpeF/W69ouyjd1i9BruIMc5wtW2d3sz7eVtY/Z5R2osWe3Oaay/8d649Jq4MPcczeJrQieIHoR8jdUGJpjWf0RN4Y0QmhK56j8W7u/K2c255HDHEotH7GZrtqQDvBLpWbZwHNfP5Twlvj/iPudrvfN8Y6zuiX4CMlETXf13kDV53wWfxPqcTQ8WRui9B/N+N+mr/DmyKsE1EYqQEQmTP8AXeWIIhFjY1aT213QI+jtaIc265c+4veLPc7MK7ulyr9qTvmFbGGm3fa5xzT+JN77kMOLqJ36W+RaHn/Zppv9o596Fz7i682WDAm9EF59xOvCvhfS1s+tKQMaHnaGsJRFMDdSeQ4eJ9Pz8HSoGDzaxbhO3HxxlbrePCC0LxHh368b/1tznnFuF1AzvDzA4HTgDm1OsaFIvaZON/zCyzqYrmrQfRqs8odCfrJb6Zxad2PERMSU+cGr2feJ9xgIbvZZt+n8LU7ivWlb234c3E1kDouDi4FXFMDj1fFJrS9TvAIufcR2H1WvPe1P4dCG9TROKgBEJkDxDqOvI03knK/4XPgW5m/fim69HkeuV7RxjUCN4MMWk07AbwEN5V3PvMrNHiXuat6RDzf/hmVhQ6AQ2XBeTg9RkPH4cRyezQ8xHhG8zsMDOLdIW9tqz+WIu/4vXP/7/6J/yhE52b6tWJR+1A8EgDOl/EOyG+2sxOifRiMxtde4Id6qf/JN57c2tYvZF4A6xbYqyZfTes7Bq+uUK/KsJrHsHrgvQ83nv2aJz7fApvPMpg4MXQgOYGQsfT1YTGWoS05jOqXRPieuAMvNmPpkcLMLQuQUvuLt1kZnn12knH6z4F3p0Q6v17O3BLqJtd+P59ZnZ8C/Yfyct4s1SdZmbnhW80s/A7E/OAAWZ2Ulj5jXjTEreIc+7fpL7m3QAABEtJREFUeBcHTuebSRkmR6ga13cjzBF4iXukqYZFJEYaAyGy57geb0rES4DRZvYmXpeJvfD+w84B7grrYnQQMN3MPsSbonMdUBSqn8I3YyJwzn1uZpfinaB9amav4U1ZmYJ3gnwM3mDIJu8E1NMX+MDMPsO7arwGb72G7+J1U3kwdNW5Sc65T8zsC2CcmflDfatr/RDvJOQdvGkht+GdGJ+Kd8fm/np178a7Ino68LGZzcQbtHwWXt/13zvnGq010Yx/hV7/Qqi9MmCVc26ac67KzM7EO5GeYWbv4101LcW7+jsKb72H3nyT6NyA1+1nYihpqF0H4hy8waSnxRkfeCeX081sOt57dBDe7ERbgZ9Eec1zwH14n2Ex3hz9MXPOBc3sLLyVw08HVpjZv4DP8E7+9sL7PYvwPpdarfmM3sCbyrT2ZP2hCBMF8P/bu3vQKIIwjOPPi4KxEUEQCwmCYAoNFoKCTURBjEUMaGOnjWIj2JoihXYKNjapTCFBQQuxEC3kGj8KC4WIgpBoJ0gsomJOlLF4xnxcNrm9u8TT5P+DacK7l9m9W9iZnXdeabo+ilSivkKBN/L9MbsOxHa5ZsN0pfSU0kREnJAHMc/z+b+Wc5Q65TeGm1SQ29OolNKPfL0fSRqJiLPyLH+HnIh8SHOfF67KbwLvRcRt+bewX96ut6Lm33ZJTsS/JA/4fkoq2hWqmXtD+c3cXnn74aKlbADKSv/AXrI0Gu3vNDmp8KJcPG1SfoD5KD8kHi2I3yrvUvQkx1XlrTUfSOpd4H90y7OGH3L8Z3nwMSTpYE1sRQvXQdgoJ28/lpM7q3IibEVeYhQNnPefvd97a/6+T54tf5X7+V1+SL4haVfB53Tk6zeaY7/ID+knC2K35f85vEi/1uTrO6aZOg2VmpjN8s5Go/LD0Fd5lvaOnAuytiZ+izyI+5T7+FLeBemAmqsDcUoetD2T9E2eFb8raUed46/l46+0+Js9LD9EjufzmcrXa0TSkVa+o4JjBzRTr6BrkbjdOeZmA+dRycesk3Q5n081n8ugCmoWzPodXc/f+ZR8376VBxv9NbHDqqnB0OC17pQLAo7Lb/cm5MJvAwWxfXKNlqkcd0se2M3rQ5l7oaYPv3L8/Tqxjd4bZ/Ln9tfrB41GW7xFSs3m9wHA/yEiNshLHp6mlI7Vi0frIqIiVwHvSim9a3N3llREnJffTnUn12Qoc0xFLnrW1Da1aF1EvJAnUXamuW8iATSIHAgAK15KaVKe5e3LS3uwjPKa/R5JD1fa4CHrkYuulRo8oP0iol/SHnn7ZwYPQIvIgQCwWgzJy6IW2pYULYqIc3Lew2l5rf5ge3u0PFJKx9vdBzRsvaQLaX5VdQBNYAkTAGBJRMR7OW9mTM61mJcAu1qxhAnASsIAAgAAAEBp5EAAAAAAKI0BBAAAAIDSGEAAAAAAKI0BBAAAAIDSGEAAAAAAKO03y6WR+rUHBzUAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 792x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# This cell prints the (sorted) CoV in number of CPUs (cores) requested per user in general and for ML jobs in a curve plot\n",
    "\n",
    "data = filtered_jobs[[\"UID\", \"NCPUS\"]].groupby([\"UID\"]).std() / filtered_jobs[[\"UID\", \"NCPUS\"]].groupby([\"UID\"]).mean()\n",
    "data.sort_values(\"NCPUS\", ascending=False, inplace=True)\n",
    "\n",
    "x_vals = np.arange(len(data))\n",
    "y_vals = data[\"NCPUS\"]\n",
    "fig, ax = plt.subplots(figsize=(11,3))\n",
    "ax.plot(x_vals, y_vals, color=\"lightcoral\", label=\"All Jobs\", marker=\"s\", markevery=25, markersize=6)\n",
    "\n",
    "data = filtered_jobs[filtered_jobs['is_gpu'] == True][[\"UID\", \"NCPUS\"]].groupby([\"UID\"]).std() / filtered_jobs[filtered_jobs['is_gpu'] == True][[\"UID\", \"NCPUS\"]].groupby([\"UID\"]).mean()\n",
    "data.sort_values(\"NCPUS\", ascending=False, inplace=True)\n",
    "\n",
    "x_vals = np.arange(len(data))\n",
    "y_vals = data[\"NCPUS\"]\n",
    "\n",
    "ax.plot(x_vals, y_vals, color=\"steelblue\", label=\"ML Jobs\", marker=\"o\", markevery=25, markersize=6)\n",
    "\n",
    "ax.set_xlabel(\"Users (sorted by CoV, per curve)\", fontsize=20)\n",
    "ax.set_ylabel(\"CoV\", fontsize=20)\n",
    "ax.tick_params(axis='both', which='major', labelsize=18)\n",
    "ax.tick_params(axis='both', which='minor', labelsize=16)\n",
    "\n",
    "ax.legend(prop={\"size\": 16})\n",
    "fig.tight_layout()\n",
    "\n",
    "date_time = datetime.now().strftime(\"%Y-%m-%d-%H-%M-%S\")\n",
    "fig.savefig(f\"cov_cpus_requested_per_user_line_{date_time}.pdf\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\begin{table}[] % Made by running job_arrival_characterization.ipynb\n",
      "\\caption{Fraction of jobs per job state in Lisa}\n",
      "\\label{surfing:tbl:job-state-fraction}\n",
      "\\adjustbox{width=\\linewidth}{\n",
      "\\begin{tabular}{@{}llllllllll@{}}\n",
      "COMPLETED & FAILED & CANCELLED & TIMEOUT & OUT OF MEMORY & REQUEUED & NODE FAILURE\n",
      "0.916742 & 0.060288 & 0.011632 & 0.010783 & 0.000544 & 0.000007 & 0.000003\n",
      "\\end{tabular}\n",
      "}\n",
      "\\end{table}\n"
     ]
    }
   ],
   "source": [
    "# This cell outputs the total fraction of jobs per state\n",
    "\n",
    "state_counts_fraction = filtered_jobs[\"State\"].value_counts() / len(filtered_jobs)\n",
    "state_counts_fraction.rename({\"OUT_OF_ME\": \"OUT OF MEMORY\", \"NODE_FAIL\": \"NODE FAILURE\"}, inplace=True)\n",
    "\n",
    "print(r\"\"\"\\begin{table}[] % Made by running job_arrival_characterization.ipynb\n",
    "\\caption{Fraction of jobs per job state in Lisa}\n",
    "\\label{surfing:tbl:job-state-fraction}\n",
    "\\adjustbox{width=\\linewidth}{\n",
    "\\begin{tabular}{@{}llllllllll@{}}\"\"\")\n",
    "print(\" & \".join(state_counts_fraction.index))\n",
    "print(\" & \".join([\"{:.6f}\".format(i) for i in state_counts_fraction.values]))\n",
    "print(r\"\"\"\\end{tabular}\n",
    "}\n",
    "\\end{table}\"\"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "               completed  cancelled     failed  out_of_memory  node_fail  \\\n",
      "type                                                                       \n",
      "$<$5 minutes    20906469     941884    2279845            0.0        0.0   \n",
      "$<$1 hour      188182872   10718515   10815368            0.0     4970.0   \n",
      "$<$6 hours     552998001   40811256   40014691            0.0        0.0   \n",
      "$<$1 day       771351584  135774303   66325910            0.0        0.0   \n",
      "$\\geq$1 day   1390810962  200480045  165246396            0.0  1050672.0   \n",
      "\n",
      "                 timeout  \n",
      "type                      \n",
      "$<$5 minutes       93846  \n",
      "$<$1 hour        5603816  \n",
      "$<$6 hours      60027569  \n",
      "$<$1 day       172581325  \n",
      "$\\geq$1 day   1078554531  \n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 792x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Kristian's stacked barplot - largely based on his code\n",
    "def jobType(x):\n",
    "    if(x < 300):\n",
    "        return '$<$5 minutes'\n",
    "    elif(x < 3600):\n",
    "        return '$<$1 hour'\n",
    "    elif(x < 21600):\n",
    "        return '$<$6 hours'\n",
    "    elif (x < 86100):\n",
    "        return '$<$1 day'\n",
    "    else:\n",
    "        return '$\\geq$1 day'\n",
    "\n",
    "def parsedfsum(df, arg):\n",
    "    retdf = df[df['State'] == arg]\n",
    "    retdf = retdf[['ElapsedRaw']]\n",
    "    retdf['type'] = retdf['ElapsedRaw'].apply(lambda x: jobType(x))\n",
    "    retdf = retdf.groupby('type').sum()\n",
    "    retdf.columns = [arg.lower()]\n",
    "    return retdf\n",
    "\n",
    "completed = parsedfsum(filtered_jobs, 'COMPLETED')\n",
    "cancelled = parsedfsum(filtered_jobs, 'CANCELLED')\n",
    "failed = parsedfsum(filtered_jobs, 'FAILED')\n",
    "oom = parsedfsum(filtered_jobs, 'OUT_OF_MEMORY')\n",
    "nodefail = parsedfsum(filtered_jobs, 'NODE_FAIL')\n",
    "timeout = parsedfsum(filtered_jobs, 'TIMEOUT')\n",
    "\n",
    "\n",
    "merged = pd.concat([completed,cancelled,failed,oom,nodefail,timeout], axis=1).sort_values('completed', ascending=True)\n",
    "merged.fillna(0, inplace=True)\n",
    "print(merged)\n",
    "\n",
    "colors = [\"lightcoral\", \"steelblue\", \"yellowgreen\", \"orchid\", \"gold\", \"gray\"]\n",
    "hatches = ['', '/', '\\\\', '+', '*', '.']\n",
    "\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(11,5))\n",
    "ax = merged.plot.bar(stacked=True,\n",
    "                     fontsize=14,\n",
    "                     ax=ax,\n",
    "                     color=colors,\n",
    "                     edgecolor=\"black\",\n",
    "                    )\n",
    "\n",
    "\n",
    "# Add the patterns, based on https://stackoverflow.com/a/22836353\n",
    "bars = ax.patches\n",
    "for num, bar in enumerate(bars):\n",
    "    bar.set_hatch(hatches[int(num / merged.index.nunique())])\n",
    "\n",
    "ax.get_yaxis().set_major_formatter(ScalarFormatter(useMathText=True)) \n",
    "ax.set_xticklabels(labels=merged.index, rotation=0)\n",
    "ax.yaxis.get_offset_text().set_visible(False)\n",
    "\n",
    "fig.tight_layout()  # This forces matplotlib to compute the offset label, which otherwise will be empty. Has no impact on the figure.\n",
    "\n",
    "ax.set_xlabel('Job Length', fontsize=20)\n",
    "ax.set_ylabel(\"Total Runtime ({}) [s]\".format(ax.get_yaxis().get_major_formatter().get_offset()), fontsize=20)  # Based on https://stackoverflow.com/a/45766598\n",
    "ax.tick_params(axis='both', which='major', labelsize=18)\n",
    "ax.tick_params(axis='both', which='minor', labelsize=16)\n",
    "handles, labels = ax.get_legend_handles_labels()\n",
    "ax.legend(handles[::-1], labels[::-1], prop={\"size\": 15})\n",
    "\n",
    "date_time = datetime.now().strftime(\"%Y-%m-%d-%H-%M-%S\")\n",
    "fig.savefig(f\"runtime_per_state_grouped_job_runtime_{date_time}.pdf\", bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Run this once for the next cell\n",
    "def jobType(row):\n",
    "    x = row['elapsedraw']\n",
    "    if(x < 300):\n",
    "        return '$<$5 minutes'\n",
    "    elif(x < 3600):\n",
    "        return '$<$1 hour'\n",
    "    elif(x < 21600):\n",
    "        return '$<$6 hours'\n",
    "    elif (x < 86100):\n",
    "        return '$<$1 day'\n",
    "    else:\n",
    "        return '$\\geq$1 day'\n",
    "\n",
    "\n",
    "plot_df = filtered_jobs.copy()\n",
    "plot_df = plot_df[plot_df['State'].isin(['COMPLETED', 'CANCELLED', 'FAILED', 'OUT_OF_ME', 'NODE_FAIL', 'TIMEOUT'])]\n",
    "plot_df.columns = [col.lower() for col in plot_df.columns]\n",
    "plot_df['group_runtime'] = plot_df.apply(jobType, axis=1)\n",
    "plot_df.fillna(0, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "group_runtime\n",
      "$<$5 minutes    0.137867\n",
      "$<$1 hour       0.128716\n",
      "$<$6 hours      0.205573\n",
      "$<$1 day        0.331014\n",
      "$\\geq$1 day     0.512334\n",
      "dtype: float64\n",
      "group_runtime\n",
      "$<$5 minutes    0.519284\n",
      "$<$1 hour       0.396604\n",
      "$<$6 hours      0.323560\n",
      "$<$1 day        0.469558\n",
      "$\\geq$1 day     0.558288\n",
      "dtype: float64\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 792x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Kristian's stacked barplot - largely based on his code\n",
    "colors = [\"lightcoral\", \"steelblue\", \"yellowgreen\", \"orchid\", \"gold\", \"gray\"]\n",
    "hatches = ['', '/', '\\\\', '+', '*', '.']\n",
    "\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(11,4))\n",
    "\n",
    "# Plot the general bars:\n",
    "bar_width = 0.4\n",
    "offset = [-0.21, .21]\n",
    "\n",
    "sub_df = plot_df[['group_runtime', 'state', 'elapsedraw']].groupby([\"group_runtime\", \"state\"]).sum().unstack(level=1).fillna(0)\n",
    "sub_df.columns = [col.lower() for col in sub_df.columns.droplevel(0)]\n",
    "sub_df.rename(columns={\"out_of_me\": \"out_of_memory\"}, inplace=True)\n",
    "sub_df = sub_df[['completed', 'cancelled', 'failed', 'out_of_memory', 'node_fail', 'timeout']]\n",
    "sub_df = sub_df.sort_values('completed', ascending=True)\n",
    "\n",
    "print(1 - (sub_df['completed'] / sub_df.sum(axis=1)))\n",
    "\n",
    "for x_pos, group in enumerate(sub_df.index):\n",
    "    bottom = 0\n",
    "    for index, col in enumerate(sub_df.columns):\n",
    "        ax.bar(x_pos + offset[0], sub_df.loc[group].iloc[index],  bottom=bottom, label=col, color=colors[index], hatch=hatches[index], width=bar_width)\n",
    "        bottom += sub_df.loc[group].iloc[index]\n",
    "        \n",
    "ax.annotate(\"All jobs\",\n",
    "            xy=(1 + offset[0], sub_df.iloc[1].sum()), xycoords='data',\n",
    "            xytext=(0, 50), textcoords='offset points',\n",
    "            fontsize=18, color=\"black\", ha='center',\n",
    "            arrowprops=dict(arrowstyle=\"->\",\n",
    "                            connectionstyle=\"arc\",\n",
    "                            color=\"black\"\n",
    "                           )\n",
    "            )\n",
    "\n",
    "sub_df = plot_df[plot_df['is_gpu'] == True][['group_runtime', 'state', 'elapsedraw']].groupby([\"group_runtime\", \"state\"]).sum().unstack(level=1).fillna(0)\n",
    "sub_df.columns = [col.lower() for col in sub_df.columns.droplevel(0)]\n",
    "sub_df.rename(columns={\"out_of_me\": \"out_of_memory\"}, inplace=True)\n",
    "sub_df = sub_df[['completed', 'cancelled', 'failed', 'out_of_memory', 'node_fail', 'timeout']]\n",
    "sub_df = sub_df.sort_values('completed', ascending=True)\n",
    "\n",
    "ax.set_xticks(range(len(sub_df.index)))\n",
    "ax.set_xticklabels(labels=sub_df.index, rotation=0)\n",
    "\n",
    "print(1 - (sub_df['completed'] / sub_df.sum(axis=1)))\n",
    "\n",
    "for x_pos, group in enumerate(sub_df.index):\n",
    "    bottom = 0\n",
    "    for index, col in enumerate(sub_df.columns):\n",
    "        ax.bar(x_pos + offset[1], sub_df.loc[group].iloc[index],  bottom=bottom, label=col, color=colors[index], hatch=hatches[index], width=bar_width)\n",
    "        bottom += sub_df.loc[group].iloc[index]\n",
    "\n",
    "ax.annotate(\"ML jobs\",\n",
    "            xy=(1 + offset[1], sub_df.iloc[1].sum()), xycoords='data',\n",
    "            xytext=(0, 40), textcoords='offset points',\n",
    "            fontsize=18, color=\"black\", ha='center',\n",
    "            arrowprops=dict(arrowstyle=\"->\",\n",
    "                            connectionstyle=\"arc\",\n",
    "                            color=\"black\"\n",
    "                           )\n",
    "            )\n",
    "\n",
    "ax.get_yaxis().set_major_formatter(ScalarFormatter(useMathText=True))\n",
    "ax.yaxis.get_offset_text().set_visible(False)\n",
    "\n",
    "fig.tight_layout()  # This forces matplotlib to compute the offset label, which otherwise will be empty. Has no impact on the figure.\n",
    "\n",
    "ax.set_xlabel('Job Length', fontsize=20)\n",
    "ax.set_ylabel(\"Total Runtime ({}) [s]\".format(ax.get_yaxis().get_major_formatter().get_offset()), fontsize=20)  # Based on https://stackoverflow.com/a/45766598\n",
    "ax.tick_params(axis='both', which='major', labelsize=18)\n",
    "ax.tick_params(axis='both', which='minor', labelsize=16)\n",
    "ax.legend(prop={\"size\": 16})\n",
    "\n",
    "# Filter duplicate labels see https://stackoverflow.com/questions/13588920/stop-matplotlib-repeating-labels-in-legend\n",
    "handles, labels = plt.gca().get_legend_handles_labels()\n",
    "by_label = dict(zip(labels, handles))\n",
    "ax.legend(list(by_label.values())[::-1], list(by_label.keys())[::-1], prop={\"size\": 16})\n",
    "\n",
    "date_time = datetime.now().strftime(\"%Y-%m-%d-%H-%M-%S\")\n",
    "fig.savefig(f\"runtime_per_state_grouped_job_runtime_all_vs_ml_{date_time}.pdf\", bbox_inches='tight')"
   ]
  }
 ],
 "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}