notebooks/ns3-replay.ipynb

{
    "cells": [
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "# Ns-3 Replay.\n",
                "\n",
                "You use this notebook to explore the ns-3 replay results and/or recreate Figure 8 (a--c) in the paper.\n",
                "\n",
                "## Prerequisites\n",
                "\n",
                "You must download and evaluate the puffer data, which you can do by running the following CLI commands:\n",
                "\n",
                "```bash\n",
                "./run.py ns3-replay\n",
                "```\n",
                "\n",
                "On 20 CPU Cores, each individual experiment takes between 1 to 2 hours to run.\n",
                "On a GPU, you can typically expect less than an hour per experiment.\n",
                "\n",
                "This will use the simulation data in [ns3-simulation/results](../ns3-simulation/results).\n",
                "If you want to regenerate the simulation data or simulate different scenarios, follow the instructions in [ns3-simulation/README.md](../ns3-simulation/README.md)."
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "## Imports and Settings"
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 1,
            "metadata": {},
            "outputs": [],
            "source": [
                "import sys\n",
                "sys.path.append('..')  # To allow imports from parent directory\n",
                "%load_ext autoreload\n",
                "%autoreload 2"
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 2,
            "metadata": {},
            "outputs": [],
            "source": [
                "import experiment_helpers as eh\n",
                "import pandas as pd\n",
                "import seaborn as sns\n",
                "from experiment_helpers.data import Path\n",
                "from matplotlib import pyplot as plt  # ticker\n",
                "\n",
                "# Load the config, including the local config.\n",
                "from config import Config as LocalConfig  # Load from file: ../config.py\n",
                "from experiments.ns3.config import Ns3ExperimentConfig\n",
                "from experiments.ns3.implementation.load_data import load_frame\n",
                "\n",
                "config = Ns3ExperimentConfig.with_updates(LocalConfig)\n",
                "resultdir = Path(config.output_directory) / 'ns3-replay'"
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 3,
            "metadata": {},
            "outputs": [],
            "source": [
                "# Plot import and color settings\n",
                "%matplotlib inline\n",
                "\n",
                "# Set theme and parameters\n",
                "eh.plot_utils.setup(latex=True)\n",
                "\n",
                "def plot_markers(*markers, offset=3):\n",
                "    \"\"\"Helper to annotate plot phases.\"\"\"\n",
                "    ax = plt.gca()\n",
                "    ylim = ax.get_ylim()\n",
                "    for x, text in markers:\n",
                "        ax.plot([x, x], ylim, color='black', zorder=1, linewidth=1)\n",
                "        ax.annotate(text, xy=(x, ylim[1]),\n",
                "                    va='bottom', ha='left',\n",
                "                    xytext=(offset, offset),\n",
                "                    textcoords=\"offset points\")\n",
                "    ax.set_ylim(ylim)  # Counteract auto-adjustment.\n",
                "\n",
                "base_palette = sns.color_palette(\"deep\")\n",
                "palette = {\n",
                "    \"Memento\": base_palette[0],\n",
                "    \"Random\": base_palette[1],\n",
                "    \"Memento(T=0)\": base_palette[4],\n",
                "    \"FIFO\": base_palette[2],\n",
                "    \"LARS\": base_palette[3],\n",
                "}\n",
                "\n",
                "\n",
                "plotdir = Path(\"../plots/ns3\")\n",
                "plotdir.mkdir(exist_ok=True, parents=True)"
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "# Data Overview\n",
                "\n",
                "Show the average number of samples per ns3 simulation run."
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 4,
            "metadata": {},
            "outputs": [
                {
                    "data": {
                        "text/html": [
                            "<div>\n",
                            "<style scoped>\n",
                            "    .dataframe tbody tr th:only-of-type {\n",
                            "        vertical-align: middle;\n",
                            "    }\n",
                            "\n",
                            "    .dataframe tbody tr th {\n",
                            "        vertical-align: top;\n",
                            "    }\n",
                            "\n",
                            "    .dataframe thead th {\n",
                            "        text-align: right;\n",
                            "    }\n",
                            "</style>\n",
                            "<table border=\"1\" class=\"dataframe\">\n",
                            "  <thead>\n",
                            "    <tr style=\"text-align: right;\">\n",
                            "      <th></th>\n",
                            "      <th>run</th>\n",
                            "      <th>horizon</th>\n",
                            "      <th>wl_samples</th>\n",
                            "      <th>tt_samples</th>\n",
                            "    </tr>\n",
                            "    <tr>\n",
                            "      <th>experiment</th>\n",
                            "      <th></th>\n",
                            "      <th></th>\n",
                            "      <th></th>\n",
                            "      <th></th>\n",
                            "    </tr>\n",
                            "  </thead>\n",
                            "  <tbody>\n",
                            "    <tr>\n",
                            "      <th>skewed1</th>\n",
                            "      <td>0</td>\n",
                            "      <td>128</td>\n",
                            "      <td>52991</td>\n",
                            "      <td>52991</td>\n",
                            "    </tr>\n",
                            "    <tr>\n",
                            "      <th>skewed3</th>\n",
                            "      <td>0</td>\n",
                            "      <td>128</td>\n",
                            "      <td>51469</td>\n",
                            "      <td>51469</td>\n",
                            "    </tr>\n",
                            "    <tr>\n",
                            "      <th>w1</th>\n",
                            "      <td>0</td>\n",
                            "      <td>128</td>\n",
                            "      <td>10082</td>\n",
                            "      <td>10082</td>\n",
                            "    </tr>\n",
                            "    <tr>\n",
                            "      <th>w1_c100</th>\n",
                            "      <td>0</td>\n",
                            "      <td>128</td>\n",
                            "      <td>10082</td>\n",
                            "      <td>10082</td>\n",
                            "    </tr>\n",
                            "    <tr>\n",
                            "      <th>w1_c133</th>\n",
                            "      <td>0</td>\n",
                            "      <td>128</td>\n",
                            "      <td>10120</td>\n",
                            "      <td>10120</td>\n",
                            "    </tr>\n",
                            "    <tr>\n",
                            "      <th>w1w2</th>\n",
                            "      <td>0</td>\n",
                            "      <td>128</td>\n",
                            "      <td>85302</td>\n",
                            "      <td>85302</td>\n",
                            "    </tr>\n",
                            "    <tr>\n",
                            "      <th>w1w3</th>\n",
                            "      <td>0</td>\n",
                            "      <td>128</td>\n",
                            "      <td>6042</td>\n",
                            "      <td>6042</td>\n",
                            "    </tr>\n",
                            "    <tr>\n",
                            "      <th>w2</th>\n",
                            "      <td>0</td>\n",
                            "      <td>128</td>\n",
                            "      <td>165718</td>\n",
                            "      <td>165718</td>\n",
                            "    </tr>\n",
                            "    <tr>\n",
                            "      <th>w2_c100</th>\n",
                            "      <td>0</td>\n",
                            "      <td>128</td>\n",
                            "      <td>99818</td>\n",
                            "      <td>99818</td>\n",
                            "    </tr>\n",
                            "    <tr>\n",
                            "      <th>w2_c133</th>\n",
                            "      <td>0</td>\n",
                            "      <td>128</td>\n",
                            "      <td>39184</td>\n",
                            "      <td>39184</td>\n",
                            "    </tr>\n",
                            "    <tr>\n",
                            "      <th>w3</th>\n",
                            "      <td>0</td>\n",
                            "      <td>128</td>\n",
                            "      <td>6873</td>\n",
                            "      <td>6873</td>\n",
                            "    </tr>\n",
                            "    <tr>\n",
                            "      <th>w3_c100</th>\n",
                            "      <td>0</td>\n",
                            "      <td>128</td>\n",
                            "      <td>6873</td>\n",
                            "      <td>6873</td>\n",
                            "    </tr>\n",
                            "    <tr>\n",
                            "      <th>w3_c133</th>\n",
                            "      <td>0</td>\n",
                            "      <td>128</td>\n",
                            "      <td>6960</td>\n",
                            "      <td>6960</td>\n",
                            "    </tr>\n",
                            "  </tbody>\n",
                            "</table>\n",
                            "</div>"
                        ],
                        "text/plain": [
                            "            run  horizon  wl_samples  tt_samples\n",
                            "experiment                                      \n",
                            "skewed1       0      128       52991       52991\n",
                            "skewed3       0      128       51469       51469\n",
                            "w1            0      128       10082       10082\n",
                            "w1_c100       0      128       10082       10082\n",
                            "w1_c133       0      128       10120       10120\n",
                            "w1w2          0      128       85302       85302\n",
                            "w1w3          0      128        6042        6042\n",
                            "w2            0      128      165718      165718\n",
                            "w2_c100       0      128       99818       99818\n",
                            "w2_c133       0      128       39184       39184\n",
                            "w3            0      128        6873        6873\n",
                            "w3_c100       0      128        6873        6873\n",
                            "w3_c133       0      128        6960        6960"
                        ]
                    },
                    "execution_count": 4,
                    "metadata": {},
                    "output_type": "execute_result"
                }
            ],
            "source": [
                "statframe = eh.data.read_csv(resultdir / 'overview/overview.csv.gz')\n",
                "statframe.groupby(\"experiment\").min().astype(int)"
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "The data is skewed heavily towards w2. w1 and w3 make up a small fraction of total packets.\n",
                "Let's compute the exact fractions for the periodic experiment, where w2 is present all the time and w1 and w3 only appear sporadically."
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 5,
            "metadata": {},
            "outputs": [
                {
                    "data": {
                        "text/plain": [
                            "(0.012987468649322886, 0.004674763206681801)"
                        ]
                    },
                    "execution_count": 5,
                    "metadata": {},
                    "output_type": "execute_result"
                }
            ],
            "source": [
                "total = (\n",
                "    statframe\n",
                "    .query(\"experiment == 'w2'\")\n",
                "    [lambda df: df['run'].isin(range(2,23))]\n",
                "    ['wl_samples'].sum()\n",
                ")\n",
                "\n",
                "w1 = (\n",
                "    statframe\n",
                "    .query(\"experiment == 'w1'\")\n",
                "    [lambda df: df['run'].isin(range(2,23, 5))]\n",
                "    ['wl_samples'].sum()\n",
                ")\n",
                "w3 = (\n",
                "    statframe\n",
                "    .query(\"experiment == 'w3'\")\n",
                "    [lambda df: df['run'].isin(range(2,23, 10))]\n",
                "    ['wl_samples'].sum()\n",
                ")\n",
                "\n",
                "\n",
                "w1/total, w3/total"
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "## Packet size distributions \n",
                "\n",
                "Looking at the packet size distribution, we can see that w1 and w3 have a much larger fraction of max-MTU-sized packets, while w2 contains smaller packets, which also explains why w2 has so many more packets overall.\n",
                "\n",
                "Going by packet sizes, this makes w2 easily distinguishable from w1 and w3, which are very similar to each other and more difficult to keep apart."
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 6,
            "metadata": {},
            "outputs": [
                {
                    "data": {
                        "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjkAAAHPCAYAAACxyMv0AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8pXeV/AAAACXBIWXMAAA9hAAAPYQGoP6dpAABMRElEQVR4nO3de1hc953n+U/duEkU6C5ZOli27gIpsmUlVil3EQs5090TZWw807M9VmLZuztPIN0NO/tsLLqD88zuGtI7aCez00JupXu6d0yyg3umeyyw5U4nMaXY8kUxoJtl2eboYt2hCnGry9k/MCVhblVQRVHF+/U8fnw4dU6dL/qp4KPf+Z3fz2ZZliUAAIA0Y092AQAAAIlAyAEAAGmJkAMAANISIQcAAKQlQg4AAEhLzmRevKGhQS0tLTpw4EDU55imqZqaGuXn50uScnNzVVlZmaAKAQBAqrIl4xFyr9erpqYmNTQ0qLCwUI2NjVGdZ5qmiouLVVdXp5KSEklSWVmZJMUUlAAAQPpLSsgZsm3bNhmGEXXI2bt3r0zT1NGjRyP7hoLP4cOH5fF4ElUqAABIMSkzJsfn88nr9Y4IMoZhyO12q76+PkmVAQCAmShlQo7X65UkFRYWjnjNMIzI6xMJh8Pq6elROByOa30AAGBmSerA41iYpjnma3l5eZIGe3vcbve479PX16dTp07FtTYAwNQ5Pv5EBauXSJLeu3BLuUvmT/q9tm7dGq+ykMJSJuR0dnZKuhNo7pabmytJ6urqmjDkDFm7dq1ycnLiVh9mplAopNbWVm3atEkOhyPZ5WAa0Oapq91sjmyvXr1ay1YVTHgO7Y3xpEzIKSgY+y+73++XNHoAGovD4eADMYvQ3rMPbZ7abDZ7TO1He2M0KTMmp6ioSNJgb81nDe2LthcHAACkv5QJOYZhSJI6OjpGvGaaJo+PAwCAYVIm5Ljdbnk8HjU3Nw/bb5qmfD6fSktLk1QZAACYiWZsyDFNU9u2bRs2/011dbVM0xz2uHhNTY08Hk9kBmQAAAApSQOP29vb1dDQIJ/Pp/b2dtXX18vj8Yw6B87d42wMw9DRo0dVU1OjlpYW+f1+GYbB2lUAAGCEpIScwsJCVVdXq7q6esxjDMPQ8ePHR93POlUAAGAiM/Z2FQAAwFQQcgAAQFoi5AAAgLREyAEAAGmJkAMAANISIQcAAKQlQg4AAEhLhBwAAJCWCDkAACAtEXIAAEBaIuQAAIC0RMgBAABpiZADAADSEiEHAACkJUIOAABIS4QcAACQlgg5AAAgLRFyAABAWiLkAACAtETIAQAAaYmQAwAA0hIhBwAApCVCDgAASEuEHAAAkJYIOQAAIC0RcgAAQFoi5AAAgLREyAEAAGmJkAMAANISIQcAAKQlQg4AAEhLhBwAAJCWCDkAACAtEXIAAEBaIuQAAIC0RMgBAABpiZADAADSEiEHAACkJUIOAABIS4QcAACQlgg5AAAgLRFyAABAWiLkAACAtETIAQAAaYmQAwAA0hIhBwAApCVCDgAASEuEHAAAkJYIOQAAIC0RcgAAQFoi5AAAgLREyAEAAGmJkAMAANISIQcAAKQlQg4AAEhLhBwAAJCWCDkAACAtEXIAAEBaIuQAAIC0RMgBAABpiZADAADSEiEHAACkJUIOAABIS4QcAACQlgg5AAAgLTmTeXHTNFVTU6P8/HxJUm5uriorK6M61+v16sUXX5RhGPL7/ers7NSPfvQjud3uBFYMAABSRdJCjmmaKi4uVl1dnUpKSiRJZWVlKisr04EDB8Y91+v1qqqqSkePHo3sq6+v15NPPqnGxsaE1g0AAFJD0m5XVVVVyTCMSMCRpMrKSjU3N8vr9Y57bn19vTZu3Dhsn8fjUXt7u0zTTEi9AAAgtSQl5Ph8Pnm9Xnk8nmH7DcOQ2+1WfX39uOd3dXXp2LFjw/YNhZu8vLz4FgsAAFJSUkLOUE9NYWHhiNcMw5iwJ+fpp5+Wz+fTnj175PP55PP5dPDgQVVXVzMmBwAASErSmJzxbikN9cT4fL4xA0tJSYmqq6tVVVWlnTt3yjAM1dXVyTCMqGsIhUIKhUKxFY6UM9TGtPXsQZunB8sKR9WGY7W3w+FISF1ILUkJOZ2dnZJGv7WUm5srafCW1Hi9MqWlpWppaVFzc7Pa29vV1NSkffv2RV3D2bNnYysaKa21tTXZJWCa0eapxxG2Itvnzp3TFf/NqM/9bHtv3bo1bnUhdSUl5BQUFIz5mt/vlzTx2JqamhoZhqGjR49q7969qq2tVWdnZ9SPoK9du1Y5OTnRF42UFAqF1Nraqk2bNvEvu1mCNk9d7WZzZHv16tVatmrs3xVDaG+MJykhp6ioSNJgb81nDe0brxenpqZGzc3NkUfIGxsb9eSTT+rQoUPasWPHiAHNo3E4HHwgZhHae/ahzVObzWaPqf1ob4wmKQOPh8bOdHR0jHjNNM0JQ8rPfvazYce43W7V1dVJklpaWuJYKQAASFVJCTlut1sej0fNzc3D9pumKZ/Pp9LS0nHPNwwjMq7n7n1utzsyezIAAJjdkjYZYHV1tUzTHPa4eE1NjTwez7AJAk3T1LZt24bNnfP000+rubl52FNa7e3tkjRhQAIAALND0pZ1GBo0XFNTo5aWFvn9fhmGMebA4bvH6JSUlKiuri4ya3Jubq5M01RjYyPz5AAAAElJXqDTMIwJ16kyDEPHjx8fsb+kpGRYjw8AAMDdkna7CgAAIJEIOQAAIC0RcgAAQFoi5AAAgLREyAEAAGmJkAMAANISIQcAAKQlQg4AAEhLhBwAAJCWCDkAACAtEXIAAEBaIuQAAIC0RMgBAABpiZADAADSEiEHAACkJUIOAABIS4QcAACQlgg5AAAgLRFyAABAWiLkAACAtETIAQAAaYmQAwAA0hIhBwAApCVCDgAASEuEHAAAkJYIOQAAIC0RcgAAQFoi5AAAgLREyAEAAGmJkAMAANISIQcAAKQlQg4AAEhLhBwAAJCWCDkAACAtEXIAAEBaIuQAAIC0RMgBAABpiZADAADSEiEHAACkJUIOAABIS4QcAACQlgg5AAAgLRFyAABAWiLkAACAtETIAQAAaYmQAwAA0hIhBwAApCVCDgAASEuEHAAAkJYIOQAAIC0RcgAAQFoi5AAAgLREyAEAAGmJkAMAANISIQcAAKQlQg4AAEhLhBwAAJCWCDkAACAtEXIAAEBaIuQAAIC0RMgBAABpiZADAADSEiEHAACkJUIOAABIS4QcAACQlpzJLiAeTNNUU1OTJMnj8aiwsDDJFQEAgGRLasgxTVM1NTXKz8+XJOXm5qqysjLq830+n2pra+X1elVXV0e4AQAAEUm7XWWapoqLi/Xoo4+qurpa1dXVMk1TZWVlUZ+/Z88emaapo0ePEnAAAMAwUffk+P1+7d+/X8eOHZPP5xv1GJvNppMnT0b1flVVVTIMQyUlJZF9lZWVKi4ultfrlcfjGff8vXv3qqurS42NjdF+CwAAYBaJOuQ8++yzam5ulmEYevjhh5WXlzfpi/p8Pnm9XpWWlg7bbxiG3G636uvrxw059fX1Mk1T1dXVcrvdk64DAACkr6hDzrFjx1RaWqof/vCHU76o1+uVpFFvMRmGEXl9LAcPHoxs7927V6ZpauPGjXrmmWeivm0VCoUUCoViqBqpaKiNaevZgzZPD5YVjqoNx2pvh8ORkLqQWqIOOXl5edqxY0dcLmqa5rjXkQZ7e0brpTFNM/KaYRg6fPiw2tvbVV5erj179qixsTGqoHP27NnJfwNIOa2trckuAdOMNk89jrAV2T537pyu+G9Gfe5n23vr1q1xqwupK+qQ88gjj6ilpUWPPPLIlC/a2dkpSaPe8srNzZUkdXV1jRlyJOnxxx+P3NIqLCxUXV2d9uzZo9raWh0+fHjCGtauXaucnJzJfgtIEaFQSK2trdq0aRP/spslaPPU1W42R7ZXr16tZasKJjyH9sZ4og45TzzxhMrLy/WHf/iH2r17t3JzcyOPft9tw4YNE75XQcHYf3H9fr+k0QPQ3T577cLCQhmGoba2tgmvLw12ZfKBmD1o79mHNk9tNps9pvajvTGaqEPOt7/9bfl8Pp08eTIy8d7dLMuSzWbTqVOnJnyvoqIiSYO9NZ81tG+sAcWGYUi60xv02dfGuxUGAABmj6hDznPPPRe3iw4FlY6OjhGvmaY57pNVQ+eOFWaGXgcAALNb1CFn165dcbuo2+2Wx+NRc3PzsBmOhwYVf/bR8s8qLS1VQ0PDiP0TBSQAADB7xDzjcXd3t1544QV9//vf1/e//339xV/8hS5cuBDzhYdmOL77cfGamhp5PJ5hEwSapqlt27apvr4+sq+iokJut1s1NTWRfV6vV11dXaqoqIi5FgAAkH5iWrvqlVde0f79+yNPPvl8PjU1NammpkbV1dV67LHHon4vwzB09OhR1dTUqKWlRX6/X4ZhjLl21d1jdNxut1577TU9++yzKisrk2EY8vv9eu2115gcEAAASIoh5Jw8eVJlZWV6/PHHVVlZGXnUe2iRzaqqKm3atEnr16+P+uKGYejAgQMTHnP8+PER+91u94TnAgCA2Svq21W1tbUqKSlRdXV1JOBId4LKww8/POz2EQAAQDJFHXLa2tr06KOPjvn67t27o56jBgAAINGiDjl5eXmjzmszxOfzTWnRTgAAEH9NTU1at26d2tvb43JcIlRVVWndunVxf9+oQ8727dtVX1+v7u7uEa91d3fr4MGD2r59e1yLAwAAmKyoBx5XV1eruLhYO3fu1DPPPKMVK1ZIGlwU7dChQ3K73WM+GQUAADDdYnqE/KWXXtIPfvADPf/887LZbLKswRVjPR6PfvjDH2ru3LkJKRIAACBWMYWc3NxcHThwQH6/PzLIuKioaNjTVgAAADNBzDMeS4NhZ/v27dq+fTsBBwCACYw1sLa4uFjFxcUj9m/btk1lZWXDzi8uLta2bdu0d+/eEYODvV6vtm3bJp/PJ6/Xqz179mjv3r3j1lRTUxP11C8TXd/n86m+vl7FxcVat26dtm3bNuZ7t7e3a8+ePVq3bp327Nkz6jJN8TJqT84rr7wSCTJDolldXJI2bNgQn8oAAEgTJSUlamhokNfrjayxaJqmurq65PP5ZJrmsAWofT6fduzYIZ/Ppz179kiS9u3bp7y8PL388svas2eP6urqIssg+Xw++Xw+NTQ0qLa2VoWFhcOWSPqsvXv3yuv16vDhw+PWHe31vV6vjhw5otLSUhUWFso0TVVVVenkyZPDrjEUcNxutyoqKmQYhl5++WU1NzdP8k92fKOGnLKyMhUUFOiVV16J7PvWt74lm8025htZliWbzRZ1GAIAYLYYCjZNTU2Rba/Xq+3bt+vChQvyer2RxambmpokDc4/9+yzz6qrq2vYskUlJSWqqqrS/v375fF4hi1nVFtbOyx8jKasrExer1d1dXUTLmod7fVLSkpGXNPn86m2tnZYgKutrZUkNTY2RvaVlJREQle8jRpy/viP/3jEnDd1dXVxvzgAALOFx+MZ9ot8KPDk5+erpaUlEnK8Xq8KCwslSc3NzZFFqe9WUVGhhoYGNTQ0aN++fZH9u3btGjfgVFVVqbm5ecIgJA2GlFivf7e7e6aGtofC3NDXnz023kYNOaMVvGvXroQUAADAbDDUAzL0S9/r9Wrfvn1yu92qqqqKHOf1evXUU09FHvAZLQC43W653W61trYO2z/eygS1tbWRkDVRD46kmK8/dLvM6/Wqra1NPp9v2DlD43iGAtx0mNTAYwAAEJu7b1MN/cL3eDzavXu3pMEQMLR/vLAy5LMhQhq/R2QoPLndbpWXl8dc/3jXb2pq0s6dO3XkyBF5PB7V1dWpurp6yteYqqhDzo9//GNduHBhzNe/+93vDhvDAwAA7jAMQ4ZhqL29XW1tbZHQ43a7Iz07Xq9XbrdbhYWFKioqkqQRvTXS4C0gSdq0aVPU13/qqadUWVmp5557Tl6vd8KnmmK5/v79+7V9+3Y1NjZq37598ng8I4a9DPXgTOeyEVGHnEOHDsnv94/5+vbt2xP6GBgAAKluaFxOS0vLsFtGHo9Hra2tkcHI0mD48Xg8OnTo0Ihem6HHsycaV3O3J554InKOx+NRVVXVqL1BQ2K5vs/nG9GLNFo48ng8amhoGPF+iRh0LMUQcoaenhqL2+1mFXIAAMZRUlIi0zR17NixYSGnpKREx44dU1tbm3bs2BHZX11dLbfbrZ07d6q+vl4NDQ3au3evmpubVV1dPekBu3V1dVHdtor2+oWFhTp06JDq6+vl9XpVU1OjQ4cOjXi/iooKSdLOnTvV0NCgpqYm7dmzZ9wFwKdi3BmPjx07FnmUTRpMbkNrVg3p6uqS3+9XS0sLq5ADADCOu4PN3QNwPR5PpHdjaIyONHiL67XXXlNtba0aGhrU1dWloqIiNTY2TmkAr9vt1nPPPafy8nI1NDREnuz6rGivX1dXp/LyctXW1sowjMi4nIMHDw7LBoWFhTp8+LBqa2tVVVWlwsJCVVRUqL29XUeOHJn09zMWmzW0ANUompubI8+0m6Ypt9s9ZpDZuHGjnnjiiRm/EnlPT49OnTqlDRs2KCcnJ9nlIMFCoZBOnDihLVu2yOFwJLscTAPaPHW997cva8X9iyVJvdmLtHzNvROeQ3tjPOP25OzatSvy6Pj69ev1l3/5l8xoDAAAUkLUY3Ief/xx1qkCAAApI+qQc+HCBR07diyRtQAAAMRN1CGns7OTR8QBAEDKiDrkVFRUqK2tTa+++moi6wEAAIiLcQce323z5s2qqKhQWVmZSkpKtHv37lGfz2dgMgAAmAmiDjlf//rX5ff7ZVmWjhw5Mmz+HOnOZIGnTp2Ke5EAAACxijrk1NXVjTv9MwAAwEwSdciZ6ZP8AQAA3C3qkDOku7tbDQ0NkYW3Nm/erEceeWTEcg8AAADJFFPIeeWVV7R//351dXXJ7XbL5/OpqalJNTU1qq6u1mOPPZaoOgEAAGIS9SPkJ0+eVFlZmXbt2qXjx4/rzTff1OnTp/Xqq6/qG9/4hqqqqnT69OlE1goAABC1qENObW2tSkpKVF1dPWx5B8MwdODAAT388MOqqalJSJEAAACxijrktLW16dFHHx3z9d27d6utrS0uRQEAAExV1GNy8vLy1NXVNebrPp9PeXl5cSkKAACkPp/PJ9M09fLLL8s0TR04cGBarx91T8727dtVX1+v7u7uEa91d3fr4MGDPGYOAAAkSV6vV9u2bdOTTz6pQ4cOye/3T3sNUffkVFdXq7i4WDt37tQzzzwTeWS8tbVVhw4dktvtVmVlZcIKBQAAqcPj8ej48eNyu93as2dPUmqI6RHyl156ST/4wQ/0/PPPy2azybIsSYPfyA9/+EPNnTs3IUUCANLDQE+ffJ/c1K2Pr+rG+cu6cqpD/d29kqTNJax9mG7cbndSrx9TyMnNzdWBAwfk9/sjg4yLioqGPW0FAJjdLMuS79INme+8r5sfXdGNDy8r0NMf03s4s1wJqm7mGOi8qZ4rl2SFQ8kuZVQ2u0M5S+5RRv78ZJcyaTHPeCwNhh3G3wAAJKnnVrc+Otau86+3qedmfMZdOJyOuLzPTNZ77ROF+/uSXcaYLAXUe+3K7As5AIDZKRQI6urZCzrz6tu6etqc1HvMWeBW/opFmrdyieYVLNa8gsXKys2R96X/psVxrncmy160dMb35GQvWpLsMqYk5pDz85//XE1NTTJNU7m5uSooKFBpaakefvjhRNQHAEgSy7LUaV5Tx/Ez+ug3p9Tv74np/Jx5ubpvR6FWPLBauUvnye5I/96ZWGTkz0/pXpJUEHXI8fv9+va3vy3TNGVZljZu3CjLsnTkyBE1NTXJ4/HohRdeSGStAIAECgWC+vjNMzr/eqtufvhJTOdm58/VigfXaNWXNil36TzZbLYEVQlEL+qQU15ero6ODj311FOqqKgY9lpVVZV+9rOf6c/+7M/0R3/0R3EvEgAQX6FgSDc+uKSP3zyjC++cVaB3IKbzl226TwXb1mnFg2tmxfgZpKaoQ47X61VJScmIgCMNzqHT2dmphoYGQg4AzDD93b26eOIDXXrvvC63fhiZ/iNajgynNu/5klZsWaXsfKYKQeqIOuQYhqEdO3aM+fo3v/lNvfHGG3EpCgAwOeFQSNfOXlTHW2f00bGTssKxBRpJumfz/Vpf8pDmr1wquz3qifGBEXw+X+T/eXl5ka+na/6cqENOaWmp2tra9Nhjj436umma2rVrV9wKAwCMzwpbunK6Qx1vndGlEx9oIMa5aCRpWdFKGQ+t1T2b7lfGnKwEVInZyjRNFRcXD/t627ZtkhSZCTnRog45Ho9H3/72t1VUVBRZ0mHIhQsXVF9fr+eee07Hjh0b9hrz6QDA1FmWpa5LN3Tpt+d1uvm4gv2BmN9j9de2yNi6RvMKlsiZwQwiSCzDMHTmzJmk1hD13/Jnn31WlmVp//79I0bND93fLS8vH7bPZrPp1KlTcSoVAGaPPn+POo6fUcfxMzE/6SRJC+5fpjVf26Jlm+6TKysjARUCM1/UIaeysjJyLw0AED/hcFjXzlzQ+ZY2mW+djfn83CXzVPS727VkfQG3nIC7RB1yuO0EAFNnWZa6Ll7Xh96TuvTeed2+3hXT+fMKFmv5A6u1/HOr5F42n/logHHEfFO2u7tbDQ0Nam1tlSRt3rxZjzzyyIhxOgAAKdDbrw+97ep46+ykbjut+foDWvnwBuUbiwg0QIxiCjmvvPKK9u/fr66uLrndbvl8PjU1NammpkbV1dVjPnkFALNBKBDU1TOmLr13Xubb72vgdmyLL85fuURrvvaAlm1aqYwcbjsBUxV1yDl58qTKysr0+OOPq7KyUrm5uZIGHwmrqalRVVWVNm3apPXr1yesWACYKYIDAV17/6I+OnZSV05+HPPj29l5c7S+ZJuWb1mt7Pw59NIACRB1yKmtrVVJSYmqq6uH7TcMQwcOHNDevXtVU1PD+lUA0k44FNK19y/p/K/fk/n2+5N6j5UPb9Carz/AbSdgGkUdctra2vSjH/1ozNd3796tH//4x3EpCgCSZWhg8Llf/lYX3/1A/d29Mb/HwtXLdZ9no+7ZfL8y52YnoEoA0Yg65OTl5amra+ynAIambAaAVDLQ06f3f/FbXXj7rLou3YjpXJvNpqVFK2VsXavln7tfruzMBFUJYDJieoS8vr5eu3fv1ty5wxdo6+7u1sGDB7V79+64FwgA8RIOhXXr4ys63fyWLr13PuaFKpdtuo9AA6SQqENOdXW1iouLtXPnTj3zzDORR8ZbW1t16NAhud1uVVZWJqxQAIjF0G2nj988rYsnPpD/yq2Yzs9y52jdIw/p3s+vV5Y7J0FVAkikmB4hf+mll/SDH/xAzz//vGw2W+RfQR6PRz/84Q9H9PAAwHQJBYL6sKVd51va1Gle0wf6dUzn37P5fhX+zsPKX8HAYCBdxBRycnNzdeDAAfn9frW1tUmSioqKIo+TA8B0sCxLneY1nX+9Tedfb5MVDsd0/qI1y7X8gdW6z1PIuk5AGos65Hz3u99VSUmJHnvsMeXm5rLMA4BpE+wPyHz7rM4efSfmwcH5KxZq5faNWvHgGuXM4x9kwHRrampSbW2tTNNUYWGhSktLVVpaOi3Xjjrk3Lp1Sy+++CKzGgNIKMuy5P/kls4cfVsfHTsVcy/Nko0Fcq6Zpy9840tyOmNeuQZAHDU0NKi2tlYVFRUyDENNTU2qqqpSe3v7iHn3EiHqnwAVFRX6zne+o1dffVXf+MY3ElkTgFkk0Degj46dVMebp3UjxrWdcpfM0+qvbZHx4GpluedIkkKhkE6cOMG4GmAGqK2t1U9/+lMVFhZKGhzDm5ubq0OHDmnfvn0yDCOh14865GzevFkVFRUqKytTSUmJdu/ePWpxGzZsiGuBANJHoG9A5ttndbn1Q1088UHM5xtb12jtzgc1/76lhBhghvP5fNq+fXsk4Ax54okndOjQIXm93oTftoo65Hz961+X3++XZVk6cuSImpqahr1uWZZsNptOnToV9yIBpJ5wOKxO85o+fuO0PvjVewoHQzGdn28skrF1re7zFPIIN9KS+fZZtf3dMQX7AskuZVTOLJeKfscjY+uaSZ3vdrt14MCBEfuHJg72+XxTqi8aUYecurq6aSkIQOoZGkfzobddF949p9vXx54dfTQ2m00rtq7Rum9sVb6xSHa7PUGVAjPHmVfelv+T2OZvmm5nXn1r0iFnLF6vV9LgratEi2nGYwCQpN7Obp1vadf5X72n3q7bMZ9vPLRW939xkxbct1TOTFcCKgRmvnWPPKS2v/PO6J6cdd94KK7v6fP5tH//fpWWlo64jZUISX30wDRN1dTUKD8/X9LgPDyTmTXZ5/Np586damxsTPggJmC2Gbjdp0utH+qDX72nG+cvx3x+3j0LtPqrn9PyLasig4MBDI4xi3cvyUxXXl6u7du3T8uTVdIEIeeFF15QQ0ODTNOUYRh64okn9J3vfCcuFzZNU8XFxaqrq1NJSYkkqaysTGVlZaPewxtPeXk5t9KAOLDClq6du6jzv25Vx/EzMZ+fv2Kh7vncKt23faPmLGTBXgB3lJWVRSYVni5jhpzvfve7amlp0caNG/XII4/o5MmTev7559XS0qIXXnhhyheuqqqSYRiRgCNJlZWVKi4ultfrjfpeXX19PTMuA5NgWZa6r3bq4zdO6/1fvKtA70BM5zsynFr1pU26/0ublLtkHk87ARjT3r17ZRjGtPXgDBk15DQ3N6ulpUV1dXXatWtXZH9DQ4P+5E/+RD//+c+nNCmgz+cb9dExwzDkdrtVX18fVcgxTVOmaerRRx9Vc3PzpOsBZoNQIKgb5y+r/e9/o2vvX4zpXJvNpvu/VKQVD67VojX3yO5wJKhKAOlm79692rhxY1IW8R415DQ0NKikpGRYwJGk0tJSeb1eNTU1TSnkDI2sHm3QkWEYkdcnUlNTowMHDox4nD0aoVBIoVBsj7Qi9Qy18Wxra8uy5Lt8Ux1vnlbHm2fUF+Pg4Hn3Ltb9X9qk5VtWyZWdOfy9NbP/PGdrm6cD667tUNiKqg3Ham8HQXxG2Lt3r7q6urRjx44Rv9sNw0jOZIBtbW360Y9+NOoJjz76qKqqqqZ0UdM0x3zt7ufn3W73mMfV1NToiSeemHQNZ8+enfS5SD2tra3JLiGhAv4+3f7wuvzvX9HAzZ6YznXkZCh/03LNvX+RnHPvBJpODajzTOrOe5XubZ6OgoE7TxmdOXNaFz6Jvsfxs+29devWuNWFyRsKNnv37h3x2lNPPZXw3p1RQ47P5xszXRmGMeVBvp2dnZLuBJq7DY2v6erqGjPktLe3S5raM/Zr165VTg4TjKW7UCik1tZWbdq0KW3+ZWdZlm51XNUHv/jtpAYHLy1aqeVbVqlg2zrZnenxZ3K3dGzz2eI3HXdCzbp167Vw6aIJz6G9Z7YzZ2L/GRVPSXmEvKCgYMzX/H6/pNED0JDa2lodPnx4SjU4HA4+ELNIKre3ZVm6euaC2v/+N7p+LraxNPkrFuqezat0/5eKZt0K3Knc5rPV3UPXHXZbTO1He2M0Y4acRD4pUVRUJGmwt+azhvaN1YtTVVWlkpKSSG+OdKebsr29XT6fb1omGAISxbIsXT1tDoaaDy7FdO7Gb35By7esUv6KRTztBGDWGzPk/OAHP4hM0ne3oVtN3/3ud0c9L5rHy4duhXV0dIx4zTTNcW9DtbW1qaGhYdTXysvLJUnHjx8fdzwPMNN0X+9S+9/9Rh+/Ef0YGPey+Vr9tS269/Pr5crKSGB1AJCaxgw5d/eUjKalpWXEvmj/5eh2u+XxeNTc3Dxs0JFpmvL5fOOuStrY2DhiX1NTk8rLy3X06FFmPEZK6Lnl15lX3lbHW2fU7++N6py5i/K05fGvamnhvaztBABRGDXkHD9+POEXrq6uHjHxX01NjTwez7AJAk3T1J49e/T0009r3759Ca8LSIT+231qfel1nX+9Lepzcubl6oHSr2rZpvtkdxBqACBWo4ac6ZhB2DAMHT16VDU1NWppaZHf75dhGGM+TsbtJ6SSQN+Azr/epta/bVE4GP18LRsf/YLWFj+gjJysBFYHALNDUhfoNAxjwjUsDMOYsGeppKQk6Y+pAZ0XrulY/cvyX7kV1fEOl1PrS7Zp1Zc2KcvNdAYAEG9JDTlAKgsOBPX+P7yj1r+NboZuSVrztS0q/CcPK2MOPTUAkGiEHCBKwYGgzrz6ttr/7ljU56x4YLU2/dMdyl0yL4GVAQBGQ8gBxtF16Ybe/pvXop6vxu506At7S7TiwdXMUwMASUbIAe4SHAjo5H9/Q2dfezfqAcMrH96gwt/drjnzGRwPADMJIQezXtfF63rrb17TjfOXozp++ZZV2rznS8pdnJ/YwgAAU0LIwawTDoV09rUTeq/x11Edb7PZ9OC/+Lru21HIJHwAkEIIOUh7VjisjjdPq+2/HVPvre6ozlnp2ajN3/qisnJ5tBsAUhUhB2nHsixdbv1Q7730unyXb0qSzk9wjt3p0OeffETG1rUMGAaAOGpqatLBgwfV3t4uwzC0a9euMSf+jTdCDtJCoG9ArS+16Nwvfxv1Ocu3rNIDj39VOfMTP8M3AMxGDQ0Nqq2tVUVFhTwej0zTVHl5uUzTnHAy4Hgg5CAlWWFLH/3mpN558RcKDQSjOmdewWJt3vNFLV5n0FsDANOgqalJP/3pT1VYWChpcBWDp59+WrW1tdNyfUIOUkbPTb/e/fkvdfHdc1Edn7d8oYp+z6PLAzf1wAMPyOFwJLhCAMDdDh8+POZrPp8v4etSEnIwo117/4LeONysnpv+qI6/z1Oozz32ZWVkZ0qSQqGQPjkR3VpSADCdXvnvv9BPfvwXun27N9mljGrOnGz96z/+rh755lfj+r4NDQ0qLCycloW3CTmYUULBkN7/h3f1XuPrUR0//76lerD0q5q/cmmCKwOA+Prpn7+oDz/oSHYZ4/rpwRfjEnJM05Rpmqqvr5dpmmpsbJx6cVEg5CDpOi9e15s/bVaneS2q4wt/Z7vWfWOrnBn89QWQup585p/rJz9+YUb35Dz59BNxea+9e/fKNE1J0lNPPRUZo5No/JbAtAuHwzr3ixM68fNfRXV8dv5cPfT7O7W0aCUDhgGkjUe++dW43wqaqY4ePSpJam9v1/79+7V3795xx+vECyEH06Lz4nW91/i6Pmn/KKrj7/nc/Xro93cqyz0nsYUBAKZNYWGh6urqVFxcrPr6eu3bty+h1yPkICGCAwGdefVttf/db6I63pnp0kP/sljGQ0zGBwDpzDAMSVJra2vCr0XIQdxcPHFO7/7sl1E/CbW08F499C+LlTOPyfgAIN34fD4dOXJEpaWlw/a3t7dLknbs2JHwGgg5mLRA34DOHn1H7X8fXW9NljtHG//Jw7p/R6HszFkDAGnN7XarqalJTU1NKikpUVFRkUzT1P79+1VYWDgi/CQCIQcx8V+5pbf++qiuvX8xquONh9Zqyz/7srLz5ya4MgDATHP48GHV19eroaFBVVVVMgxDjz/+OGtXYWYIh8Myj5/RG4ebozp+zkK3tj/1KPPWAAAkSfv27Uv4AOOxEHIwjBW2dP2DS3rnxV+o6+L1qM5Z9eXN2rB7G2NrAAAzCiEH6r7Wpbf+5qiunjajOt5ms+kL3ynhSSgAwIxGyJmFwuGwLp34QG/+1asK9g1Edc6iNcv1wBNfU/7yhQmuDgCA+CDkzAKWZen6uYs63fyWLrd9FNU5efcs0PpdD6lg23rZ7PTWAABSDyEnTVlhSx+/cUpv/uUrUZ+z5bEva9VXPieHk8e7AQCpj5CTRnpu+vX2f/4HXW79MKrj3cvma+vv79Si1csTXBkAANOPkJPCrLClj35zUu/8518oFAhOeHxGTqYe/Odf14oHVzMZHwAg7RFyUtCV0x365b9rjOrYRWuW66F/WazcJfMSXBUAADMLISfF/Pon/3XC21Gf+/aXtObrW+itAQDMaoScFPKz//Hfjbp/wf3LtO1/+Ibcy+ZPb0EAAMxghJwUYFmWfv4/1Y3Y/+Wyb2npxnuTUBEAADMfIWeG831yU01/+lcj9v9e7TPKnJudhIoAAEgNhJwZzHz7rI7Vvzxi/z/7SZnsDnsSKgIAIHUQcmaot/7mqM7/um3E/sf/4/envxgAAFIQIWcGGm2AcdHvebRx9+envxgAAFIUIWeGGS3gfKV8j5ZsKJj+YgAASGGEnBki0Degl77/H0bs/+aP9mrOwrwkVAQAQGoj5MwAXRevq/m5vx6x/9v//nsslgkAwCQRcpLs3Z/9Uu//w7sj9j/2H8pls9uSUBEAAOmBkJNEo42/cS+dr5I//YPpLwYAgDRDyEmCcCik/+9f/98j9n/u21/Sum9sTUJFAACkH0LONAv09uulP/x/Ruz/nf9zn7Lz5iShIgAA0hMhZxr5r9zSkT/5yxH7GX8DAED8EXKmyUe/OaU3f9o8Yj8zGAMAkBiEnGngu3xzRMDJcufod59/OkkVAQCQ/gg50+C9l14f9vWmf+rRhhKWaAAAIJFYynoaXD1jRrY3fvMLBBwAAKYBISfBLMtSsD8Q+Xr9Iw8lsRoAAGYPQk6ChUPhYV87MrhDCADAdCDkJNiVUx2RbVdWhmw2HhUHAGA6EHIS7JO2jyLb+QWLk1cIAACzDCEnwT5+81Rke8UDq5NYCQAAswshJ8ECvQOR7eWfW5XESgAAmF0IOQlkha1hX2fnz01SJQAAzD6EnATyX7k17GvWpwIAYPoQchLoo9+cjGznzM9NYiUAAMw+hJwEMt86G9le8eCaJFYCAMDsQ8hJoNs3fJHtlds3JrESAABmH0JOgljW8EHH7qXzk1QJAACzEyEnQbqvdg772u7gjxoAgOnEb94E+fjN05HtrLw5SawEAIDZiZCTIBdPfBDZNrYy6BgAgOlGyEmQrovXI9sF29YnsRIAAGYnQs40mFewKNklAAAw6xByEqC36/awr+0OR5IqAQBg9nIm8+Kmaaqmpkb5+fmSpNzcXFVWVkZ1rtfrVW1trdrb22UYhvbt26fS0tIEVhu9i++ei2w7MpL6RwwAwKyVtJ4c0zRVXFysRx99VNXV1aqurpZpmiorK5vw3KamJtXW1mr37t0qLS1VV1eXqqqqVFNTMw2VT+zCXSFn5cNMAggAQDIkLeRUVVXJMAyVlJRE9lVWVqq5uVler3fccw8ePKjGxkbt27dP1dXVeu211+R2u3Xo0KFElx2Vq2fMyPayopXJKwQAgFksKSHH5/PJ6/XK4/EM228Yhtxut+rr68c8t729XRUVFcP2ud1u7d69W9JgD9FMsnh9QbJLAABgVkrKgJGhnprCwsIRrxmGMW5PzmjnSIPjeYbOj0YoFFIoFIrq2FgE+gaGfW1z2BJyHURn6M+eNpg9aPPUdfdiOKGwFVUbjtXeDh74gJIUcsbrbcnLy5M02Nvjdrujfs+TJ09q165dUR9/9uzZiQ+ahK6TlyPbNpdDJ06cSMh1EJvW1tZkl4BpRpunnmAgENk+c+a0LnxyMepzP9veW7dujVtdSF1JCTmdnZ2S7gSauw31yHR1dUUdckzTVFtbm1577bWoa1i7dq1ycnKiPj5ab7zzSWR7+eb7tWXLlrhfA9ELhUJqbW3Vpk2b+JfdLEGbp67fdNwJNevWrdfCpRPPMUZ7YzxJCTkFBWOPU/H7/ZJGD0BjKS8v109/+tOYen4cDkdCPhDXzl6IbBsPruVDN0Mkqr0xc9Hmqcd217bDboup/WhvjCYpA4+LiookDfbWfNbQvmgDS01NjUpLS8ccqzPd+v29ke0lG6IbHwQAAOIvKSFnaHBwR0fHiNdM0xzx1NVYmpqaVFBQMGMmAbQsa9jXruzMJFUCAACSEnLcbrc8Ho+am5uH7TdNUz6fL6rQ4vV6ZZrmiGMnmmMnkW59fCWy7chwymazjXM0AABIpKRNBjg0w/HdoaSmpkYej2fYBIGmaWrbtm3D5s5pb29XVVWVJKm+vj7yX01NjV588cXp+yY+4+KJDyLb7qXzk1YHAABI4tpVhmHo6NGjqqmpUUtLi/x+vwzDGHPtqqExOqZpas+ePZKk2traEcfV1dUlrugJXHzvfGR78XrG4wAAkExJXT3SMAwdOHBgwmOOHz8+7OszZ84kurRJ8V26Edk2HlyTxEoAAEDSblelu7zlC5NdAgAAsxohJ06CA4FhXztcSe0kAwBg1iPkxMnV03eWqnBmZSSxEgAAIBFy4ubKqTtz/izdwMrjAAAkGyEnTsx33o9sL1x9TxIrAQAAEiEnbvq6bke2l39uVRIrAQAAEiEnIbLn5Sa7BAAAZj1CThz4r9wa9rXdwR8rAADJxm/jOOh462xkOztvThIrAQAAQwg5cXD3k1X3MB4HAIAZgZATB10XrkW2ja0s5wAAwExAyImDQN9AZHv+yqVJrAQAAAwh5MSBzW6LbDszXUmsBAAADCHkxIEVtiRJ+caiJFcCAACGEHKmKBQIRrZtdv44AQCYKfitPEW3Oq5Gtj87Xw4AAEgeZ7ILSHWWZUW2l268N4mVAEDqsCxLViikcDCgYLdfkrR+Fev+Ib4IOVPku3wzsj1ngTuJlQDA9LJCIYUDAwoHgwoH+iXZFOrvlcKWQv19CgcGZLM7FOzpliMzS5ZlKTzQL9nskhWe4N1tE7wOTIyQM0W9nd2R7YGe/iRWAgDxYVmWwv19CoeCCvX2KuDvkmVZCt72y2azywqHYn7PUH/fXRcYP+DcvOVT/hJGU2DqCDlTZHc4IttL1q9IYiUAMLqhHpSA3yfLCivU1yu7w6lgX4+Ct7tld7oku13hu4PImO8Ve8C5m83pkhUMSHa7bHaH7BkZskIhOTKz9P82NKv1xCl53zypV479fErXASRCzpR13jXbsTMzI4mVAJithkJM8Hb3YO9LT48Gum7K7spQODAw4fnRHDOCzSZZkt3llM3pkis3T1YoKGd2jizLkiMjUzanS3anUzaHM6qnT099cFi/8rbGXgswBkLOFPk+uTMmx8bq4wCmKBwKygoGFQ4GZQUGFA4NTlMR7u+XbDYNdN4cHOvyaXAYL6DEGl5sTqes4J1pMZw5c+TImSu7wyFHVrYc2Tmyu1yy2fhZh9RAyJmibPcc+S7dkCQtYEkHAOMIBwODA3IHBhTs6Vaw57ZCvT1R97jczQoFFeudI0d2jpxZOXLluiW7PXKbyuHKkO2uW+9AuiDkTFHXpeuRbXpyAFjhwTEvob5eBW77Feq5PXzQ7SgmdbvoU3ZXhmSzDT61JClr4RI5snNkczjkzJ4ju4ulZjB7EXKmqM/XE9m22XjkEUhnlmXJCgZkhUKDvTD9fZLNpr7rVwZ7Y6IYuDshm012V4bsrgzZ7HbZMzLkyMySzeGUZH263zEYZPiZA4yLkDNFNrtdVnjwcUgW5wTSg2WFFQ4ENNB1S8FuvwL+rgnPiSXgZOTNkz0zS3ZXhpxZ2XJkZ8tm53YREG+EnCkaCjgszgmklsEnkgYUHuhT4Ha3woEBDdy6Ebf3d2RlKyN/vhyZWZ8O2M2g5wWYZoScKRjovTP5Hz+8gJnEGhwT09crKxhQ/60bCg/0y+ZySWFLVig48VuMw5GZJUfOHCkclnOuW46sLNldmXJkMI0EMJMQcqbAd/HOv/runi8HwPSwLEvhQEDhQL8Cvk6F+voU8HfpPpfU/cHpkccHAjG9vyNnjhyZWcqcv0hObikBKYeQEyf3f3FTsksA0lI4EFCor1eWFVaw57b6rl6e8nvaM7MG54OxLNkzMuXIzpbd6RocK5ORKbuTH41AOuCTPAV3Pz7uyOCPEpgMy7KkcFih/r5P/+tVwO9XqPf2lN/bleuWPSNTzuw5cmTnyJGVza1lYBbhN/MUXDltRrYDLM4JjGBZlkK9tz+9pRQYXD/pdrdCvbcnNQHeWFy5eQoP9Csjf77s2Tlqf/+cPrflATmY4A6Y1Qg5U3D1zJ2Qs3TjvUmsBEgOKxRSaKBf4cCAwoGBwa/7+xXoujXhStWxBhxX3jy5cubIkuTKmStHzpxRe2VCoZAs0VsDgJAzJQO378yLsXi9kcRKgPgLBwMK3u7+dBmCfgX7ehXqGbyFFFlJOk7sGZmDK1Fn58iZlS175p1bTNxeAjBZhJxJCofCw77OmJOVpEqAyRmavXegq1Oh/l4Fb3fL5nAoeLt74nNjDDj2jMzBQb2uDNldrsHBvoyPAZBghJxJunH+zhMeNruNH9aYUaxwePA2Un+fZFmDvTC9txXs7Y1bD4w9M0tWKCQrGFDm/IWyQuHBp5QyMmWz2eTIzJI9M4vPBkYIhULq7emT39et7u7b+uTSVXV8dEH/0Px6sktDmiHkTNKV0x2R7cXruFWF6WdZlsL9/eq7cSVyW0mWFffruPLmyTXXPbiGkitjMMTYWYx2Nui53aPe3n5du3Jdfl+3LnRckt3hUCAQUE93jzo+uqhFixcoGAwqEAjqjZa3df+alQoMBPT2G7/VzRudsjvsmjcvT4FAQMFgSH7fxD2FEhOsIj4IOZNkvn02sj3v3iVJrATp6O5J7sKBgEI9tzXg64ysNB1PNqdL9owMZS1cMjgehhAzI1iWpVAopGAwpGAgqGAwqG5/jwIDA+rvH9CN67dks9kUCAQ00B+Q2XFRc+bk6Mrla7p545bmzp2jc+9/pFAwpLz8XIVCYYXDYYVCIYUj22G90fK21qy/X3a7XYGBgM6f+3hKdbe/d2bY16FQSFevXB/j6LHNXzhvSnUAEiFn0vyf3Ipsr9iyKomVIBVZlqVgT7dCvT2fDuwdUKivN26PVEsaXCvJ6ZIjI0POnDmSJOecuXJkZsvGo9UJ0dfXr96eXnV1+tXf16/+/gFdv3oj8toHZz/SrZudCgVDesP7jlavvU+BQECBQFBv/eaEVq4q0ED/gC5d+GRa637/9PmEX2NFwT1yupxyuZzq6+3Xjes39YUdW3X+3Mfa8ZXPq6e7R8uNZbrHWKpHHv0qPTmIC0JOHLjvWZDsEjCDWOGwAt1+WeHQYHDp75cVDilw2y+FwxO/QQxsDofsrgy55rqVuWDRYLChF2ZClmUpGAjK77+t/r5++X3dGhgYUGAgqEAgoL7efl27ekMdH11QVnaWXv/HN7R4yUK9/ovfaN3G1RroH9Dpk+c0b0G+wqGwBgYC6u3pjbmOz4aZjz7oGOPI6TE3d45cLmfktlL+vDzluudq5f2G+nr7tHbjauXkZGtFwTK5XC7Z7DZZYUvzF86Tw+GQw+lQdnam8uflKTMzQ9k52ZozNyep3xNmN0LOJPR3D/9h5sxwJakSJJMVDg+uXN15U71XLiXkGvaMTDk/XT/J5nTJkZUt5xjzw6Qzy7Lk93Wrp6c30gtw5fI1hUNhnW5/X3nz3JExIUbBPXrz2DvKyspSbl6uQsGQzp/7WH29fcqZk61AIKjAwOQHX7/37snI9q0bnXH47kZatnyJLl+8oqX3LJbxaQ+I0+lUV6dPlqT1G1fLsix13vJpfeEauVxOOV1Odftv6/7VBbLZbMrLz1Oue46cTqdy3XNld9jlsNtldzjkcNhlt9vlcNhls9uVkeFi4kSkJULOJPiv3pr4IKSkcDCg8MDgxHZD88PYHE4F/F2yOV0Kdvskm12y4tMjY8/IUNaCJbJnZAw+Xp2RIZvDmRYhxrIs9fcPqK+3T523fJ/2lATU+u4phcIhhcOWAoGAbl7v1LUr1zVvQb4CAwG1/PJNrd2wSv/4aoskyelyKhiIftXwd4+3jvlaz+3Ye1vGMjd3jrr9g/MGrVp7n1wupz65dFULFs3XyvtW6OOPLujzngeVmZmhy5euaNOWjcrKylQoGNLS5Yu1eMlCzV84T7m5cyMhxU4vHBBXhJwpWvO1LckuAVEauo0UHhgcA2NZlvpvXpessGwOh6zQ+DP0Sr1DbzThtWx2hxw5OcrImyfJJkdW1mCIcbpS8naSZVm6ef2Wurt7dO3qDZ09eU4dH13Q+6fPa8k9i3X+/Y/VeatLixYvUNtvTys04Z/l+O6+jRNLwBmN0+mQw+GQzW5XX2+f1m9cLVdmhlwup86cPKcHHtqk/Pl5+uiDDm1+sFDZ2VmRYOV0ObVg4TytXnuf3Pm5WrR4gdx5ufR6ACmCkDNFNnvq/4s7XQw+Ut2ncGSCuz6F+/t0n8tSV/u74587iV/KNodTVjgkm9MpZ1bO4NNJc3NTqhfG1+XXtSs3dOtWl65cuqrLl67oxNttCgwEdPw3J2IOGIkcMHv/6ntlWZYKN6/T7du9cufNVU5OttZtXKNQKKRly5fI5XLJleHUtetXtO3zD2ne/PyUag8A8UXIQUqwwmFZoeCnt5H6Bwf0Dgwo2NczOOFdnNicLtmdzsGemOwc2V2uwUntXBmyORyyffraTP/FGQwGddH8JNLb8vF5U++f/VChYEin2s4q1z036vlKYuV0OeWw29XfP/ik2MNf3KpgMKRQMKR1G1crI8Old95q1RN/8E+Vk5MjV4ZTlmUp1z1Xc+bkKCPDpazsLM3NnaM5c3Ni7jUJhUI6cSKgvHz3jG8nAIlFyJmE6+cSM8h0NrNCg08i9XfelBUKKtTXq1Bf/MZP3M2V65Zrrlu2T5cYcGRmye5MrcHjlmXp1s0umR9fVOuJU7poXtYvXnldfl93VOFlMgFn8ZKFWm4s05J7FskmmzIzM/SVYo+WG8uUPy9PWdmZynXP5VYOgBmDkDMJd69bFeiL37wm6c4KhxXsua1g720Fu30K3O6WTbYJV6uOic0mR1a2XLl5kiTnXLfaz5zR5i0PpOQv33A4rHeOt+rsyXN6rfnXan33pPr64jch4KLFC7Rg4Ty5MjO0/YsPKRQO6Z7lS2Xcu1wbN61VrnsuvSEAUhYhZxLu/pl/z6b7klfIDGaFw+q7fkUDt24MLjcw1nGxvKndLtdct0J9vXK58z5drfrTAb2ujFF/GYdCIYU1s39JX/nkmn72n/6rwlZYb7z+ti5fuqob127G5b09X96mgpUrtGrtSt173wotW75UixbPV84c5i4BkP4IOVPFv3JlWZYC/i71fnJxSreYhhZ2dOXmKSN/3qcz9jpls6Xe00gTqf/3/0lHj/xKp9rOTnzwOBYtXqCNm9epcPM6bdm6SQUrl2vRkoVyufhoAwA/CSehvzt+A11TzdByBD0XO2R3uRTw+2I635GVLZvdLlfePDmzcuScM0c2e+rdRpqMcDisJ/7J0zrd/n5M581bkK8FC+fpXzy5RytX3avNWzYoIzMjQVUCQPog5EzCxRMf3Pki/os+zyhWKKT+W9fVe+WyrNDwx4mj6bVx5eYpZ9kK2TOzZuXYDr+vW83//Req/l9rJzz26e/9gb68c7ty3XO1wlgmFzNpA8CUEHImYe7iPN2+3iVJyluxMMnVxF84FNTtjvMx99LY7A7l3LNCGXnzZ/UCkDdvdOqvX/i5Dv3kryc89uVfv6jlxtJZGQABINEIOVOUMScr2SXERbCvVz0XPlawJ7pHi+2ZWcqct0CZ8xbI5nTN6l/Sv3i1RX/7s5f1i1dej+r4lasK9LdH/5Ip/AEgwQg5s5gVCun2JVMDt65PeGzmgkXKnLdQjuycWR1opMEntt59q01P//4fRT0j8B889biKd39FWx4qSnB1AIAhhJxJuHKyI9klTFpooF/+82cVHph4rpXMBYuVs2xFSq61FE+BgYD+y4t/r3+7/99Ffc6WrUX6wo4H9cS/2qMFC+clrjgAwJgIOVNkd8zsAGCFQuq7fkW9V6KbpTnnngJlzl84q4PNQP+A/tt/adI/HvXqV68di+ocp8up2p/8qXZ85fPKzMpMcIUAgGgQcibB7nQoHBycpdc5Q5+ACfh98n8Y3RwsGfMWas7yglkTbEKhkG539+jDcx/r/TPn9UbLO3K5nOru7tE/vtoS9fts2VqkJ595Ql975Iuz/hYeAMxEhJwpyDcWJbuEYSzLku/cKYV6eyY81uXO19yC+9Mu2IRCIQ0MBBQYCOjqles6e+oD/ee/+i8KDoTU/t6ZKb33qjUr9fxP/kRr1t0fp2oBAIlEyElhgW6f+m/d0EDnLckKj3tsRv4CZS+9R46M1LuVYlmWbt7o1HvvtOvK5WsKhkIaGBhQ67unNGdujk61ndX7p8/H/brrN67W/v+9QhuK1sjp5KMCAKmGn9wxsiwrcqtq2q8dDqv3yiX1Xfsk6nNyV62Ta05uAquKL8uy9Nu32/Vv9/9fOn3yXEKvtWb9/Vq2fIk2blqn9RtXa+OmdcrMypA7LzclF/MEAAxHyIlR18U7j1v3+ya+LRQPA1231P3xBxMf+ClHVrbcazamxDiRgf4BnTv7oV7+26P6q0M/i8t7rt+4WgsWzZfNZlNfX7/yF+Tqi19+WDu++gUtXrIwJf5cAABTR8iJUWjgzrwombnZCbmGZVka6Lyp2+aHUR0/995Vcs7Nld2RGs1580anHt/9XV29MvH8PEMcDodWrV2pBQvnaWfJlzVvfr4yMl1yOBxaes9iLVw0X3n57hEBJhQK6cSJE9qyZQu9MwAwy6TGb8UZatFaI67vZ4XDutX2zoTHObLnaK6xUo6sxISsRPmvPz+i/RX/R1TH5szJ1v/23Pe1+3eLWVEbADAp/PaYAfpuXFXPxQkmGLTblXvvKrly86anqDj68NzH+r2dfzDuMd949Cva8tAm/d4/K5E7L3XGEAEAZi5CThL1Xrk0/iR9Npvcq9bLmTNn+oqKg4H+AV2+dFUnW8/o33yveszj/vgH/7P+1dOl01gZAGA2SWrIMU1TNTU1ys/PlyTl5uaqsrIy4ecmW7f50bjrRTmyc+RevWHGDZC93d2jWze7dO3KdZ09/YE+PPex2t87o7Onz6u3pzfq92n2/kzLli9JYKUAACQx5JimqeLiYtXV1amkpESSVFZWprKyMh04cCBh5ybTRLMQ56/fLHtGxrTVY1mWenv7dPP6LV2+eEU9Pb26fvWmLnRc0j++2qIlyxbJ+6vjWrR4ga5dvTHl6/3Zf3xOxbu/HIfKAQCYWNJCTlVVlQzDiIQUSaqsrFRxcbG8Xq88Hk9Czk0GKxTSrfZ3x3w9f8Nm2V3Dw00oFFIgEFQwEFQwOPh/v69bvb19CgwENDAQ0NVPrsvpcmhgIKCPz19QVnamLMtSIBDUR+c7dOXyNd3u7lH+vDz5fd061XZWKwru+fTcaxPW/cH7H0lSzAHngYc26cMPOlS4eZ3mzM3R7z22W1/62sMxvQcAAFOVlJDj8/nk9XpVWjp8PIZhGHK73aqvrx8zqEzl3EQIBwIKB/oVvN2tQM9thXp7NBAMKzjQryynXaGwJYd99NtO/8u//Wu9eexdzVuQL5vNplAwpGAwqNvdiZt/50JHdAt1jiZnTrbuW1WgjIwMbdi0VrnuuVq7fpXuvW+F7r1vBQtTAgBmlKSEHK/XK0kqLCwc8ZphGJHX431uXFiW1nju071blkuSOk/9dsQhTklO5+CaUKMFnO/9m3+v1pN35sC5daMzIaV+lt1uVzgcVnZOthYumq9rV64rb55bue65WlFwj0LBkL6w40HZbDYtWbZIK1cV6L5V9/IINwAgJSXlt5dpmmO+lpc3+Ii0z+eT2+2O67mSFA4PrvF0+/ZthUKxL8/QH+7X0s8tV78V86l657fv6z8e/jtJ0r33L/90r03S4JstWbZYDqdDDodDwUBQly9+ok1bNsjhcMrhdMhus+nixU/0uQcK5XI5ZXfY5ev06977DDldTvX39WnhkoVyOZ1yOh2yO+2aPz9fefl5cmW45HDEvhhnX1+v+vpi/15niqH27u7ulj3NFiPF6Gjz2WWs9nY4HMrKyuLvwCyXlJDT2dkp6U4ouVtu7uAcKV1dXaMGlamcK0n9/f2SpI6OCealGdfknnpaWrhWf1r7x1O4buy6/J3q8ndO6zVnonPnErsOFmYe2nx2Ga29N2zYoJycnCRUg5kiKSGnoKBgzNf8fr+k0UPMVM8dem3lypXKzMwk4QNAGsvKykp2CUiypIScoqIiSYM9Lp81tG+snpipnCtJTqdTCxYsiK1gAACQcpLSlWEYg2s+jXbLyDTNcZ+Omsq5AABg9khKyHG73fJ4PGpubh623zRN+Xy+EY+Hx+tcAAAweyRtUEp1dbVM0xz2yHdNTY08Hs+wSf5M09S2bdtUX18f87kAAGD2StoEKIZh6OjRo6qpqVFLS4v8fr8Mwxhz/am7x9nEei4AAJh9bJZlTWLGF2BmMU0zMl4LQPrjM49oMJUtUk5xcfGISSGPHz8+7OtoV6lP5dXs01FDQ4NaWlrGXGg33u1K+yffRG0u8ZnH5NGTg5TS1NSkhoYGbdy4MbKvoKBg2IDzsVaplzTsB2m0xyHxvF5vpG0LCwvV2Ng44ph4tyvtn1zRtLnEZx5TZAEp5Mknn7S6uromPGbnzp3D9nV0dFhr1661WlpaYj4O0+ehhx6yvvWtb436WrzblfafGcZrc8viM4+pYcpfpAyv1yuv16va2lo1NTXJ5/ONOGZolfrPzpd09yr1sRyHmSHe7Ur7pwY+85gqQg5SRktLi6TBe/jl5eXatm2bGhoahh0T7Sr1SV/NHjGJd7vS/qmBzzymipCDlFFZWakzZ86osbExcj++qqpq2A+9aFepj/Y4zAzxblfaPzXwmcdUEXKQcgoLC1VdXa3Gxka53W5VVVVFXot2lfpoj8PMEO92pf1TC595TBYhBymrsLBQFRUVku78ay7aVeqnupo9ple825X2T0185hErQg5S2tAgwqFu5mhXqZ/qavaYXvFuV9o/dfGZRywIOUgLQ4MJo12lntXsU0u825X2T3185hENQg5Smtfr1a5duyJfR7tKPavZp5Z4tyvtn7r4zCMmyZ6oB4hGW1ub9a1vfcs6ePDgsH2jTRQ22uRe3/ve96wnn3xyUsdh+ow3MVy825X2nxnGanM+84gHlnVASvD5fCovL1dbW5vy8vK0a9euEVO7321ofRrDMOT3+ydcx2ai45BY7e3tamhoiDwaXFFRIY/HM2JOk3i3K+2fPBO1OZ95xAMhBwAApCXG5AAAgLREyAEAAGmJkAMAANISIQcAAKQlQg4AAEhLhBwAAJCWCDkAACAtEXIAAEBaIuQAaaK+vl7r1q1TU1NTsksBgBmBkAOkCbfbLcMw5Ha7k10KAMwILOsAAADSEj05AAAgLRFyAABAWiLkACnCNE3t3btX69at07p161RcXKyamprI601NTVq3bp3a29sjxw8dO9p/Xq83cq7P51NZWZm2bdumbdu2qaqqatq/PwCIN2eyCwAQnT179sgwDNXV1UmSWltb1dzcrMrKylGPNwxDhw8fjnydl5cn0zRVXl6uwsJCeTweSYMBZ+fOncrLy1NFRYUkqba2VqZpDjsfAFINIQdIAV6vVz6fT88995wKCwslSSUlJWMGnCFDQWbI/v37JSkSlCTp2WefVV5eno4ePTrsvOLiYpmmKcMw4vVtAMC04nYVkAKGgsaf//mfyzTNSb1HfX292tvbVV1dPSy4NDc3a9euXTJNM/KfJBUWFjLnDoCURk8OkAIMw9BTTz2lQ4cOqbm5WYZhyOPxqKKiIqp5cdrb21VbWyuPx6PS0tJh+yXp0KFDOnTo0IjzioqK4vdNAMA0I+QAKaKyslKPPvqoXn75ZTU3N6uhoUFer3fYbaaxlJeXy+12D7tNdbfGxsbIbTAASBfcrgJSSGFhoSorK3X06FHV1dXJNM1hT0mNpqqqSqZpqq6ubkSvz1Cwmeg9ACAVEXKAFODz+eTz+YbtGxpUPN4YHa/Xq4aGBpWWlo4YhDxk165dOnjw4Ij39/l8kx7/AwAzAbergBTg9XpVXl6u0tLSSO9LQ0ODJGn37t2jnuPz+VReXi5p8Emsz/bWFBUVye1260c/+pH27NmjnTt3qqKiIvKo+cGDB/X0009r3759CfzOACBxCDlAChgaZOz1enXkyBFJg4ORGxsbxxx4fOTIkUjvzN69e0e8XldXp5KSErndbh09elRVVVWqr6+XaZoqLCxURUXFsEHKAJBqWKATAACkJcbkAACAtETIAQAAaYmQAwAA0hIhBwAApCVCDgAASEuEHAAAkJYIOQAAIC39/y9oD4jyWczHAAAAAElFTkSuQmCC",
                        "text/plain": [
                            "<Figure size 592.147x500 with 1 Axes>"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                }
            ],
            "source": [
                "datadir = Path(\"ssh://fischer/home/alex/memento/data/ns3/\")\n",
                "\n",
                "dist_frames = pd.concat([\n",
                "    load_frame(datadir / \"w1_0.csv\"),\n",
                "    load_frame(datadir / \"w2_0.csv\"),\n",
                "    load_frame(datadir / \"w3_0.csv\")\n",
                "], ignore_index=True)\n",
                "\n",
                "sns.displot(\n",
                "    # +1 to have 1-based workloads index.\n",
                "    data=dist_frames.assign(workload = lambda df: df['workload'] + 1),\n",
                "    kind=\"ecdf\",\n",
                "    x=\"size\",\n",
                "    hue=\"workload\",\n",
                ").set(\n",
                "    #ylim=(0, 0.2),  # To zoom in on w1/w3.\n",
                ");"
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "# Data Loading\n",
                "\n",
                "Functions to load results and merge them into a single frame.\n",
                "We save many more metrics than we visualize here. To help with plotting the metrics we care about, the helper includes arguments to fetch the required metrics."
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "For aggregate accuracy, consider performance only on classes seen so far."
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 7,
            "metadata": {},
            "outputs": [],
            "source": [
                "evaluation_resultdir = resultdir\n",
                "\n",
                "def load_workload_results(experiments: dict,\n",
                "                          resultdir=evaluation_resultdir,\n",
                "                          aggregate=True):\n",
                "    frames = []\n",
                "    for label, filename in experiments.items():\n",
                "        frame = eh.data.read_csv(resultdir / filename / 'results.csv.gz')\n",
                "\n",
                "        if aggregate: # Compute balanced accuracy on classes observed until now.\n",
                "            frame = (\n",
                "                frame\n",
                "                #.loc[lambda df: df['observed']]\n",
                "                #.drop(columns=['observed', 'observed_now'])\n",
                "                .groupby(\"iteration\", as_index=False)\n",
                "                .mean()\n",
                "            )\n",
                "\n",
                "        frames.append(\n",
                "            frame\n",
                "            .assign(experiment=label, iteration=lambda df: df['iteration'] + 1)\n",
                "        )\n",
                "\n",
                "    return pd.concat(frames, ignore_index=True)\n",
                "\n",
                "def load_transtime_results(experiments: dict,\n",
                "                           metric=0.99, experiment='all',\n",
                "                           resultdir=evaluation_resultdir):\n",
                "    frames = []\n",
                "    for label, filename in experiments.items():\n",
                "        frame = eh.data.read_csv(resultdir / filename / 'results.csv.gz')\n",
                "        frames.append(\n",
                "            frame\n",
                "            [(frame['metric'] == metric)  &\n",
                "             (frame['experiment'] == experiment)]\n",
                "            .assign(\n",
                "                experiment=label,\n",
                "                iteration=lambda df: df['iteration'] + 1,\n",
                "                value=lambda df: df['value'] * 1000, # s -> ms\n",
                "            )\n",
                "        )\n",
                "\n",
                "    return pd.concat(frames, ignore_index=True)\n",
                "\n",
                "def load_stats(experiments: dict,\n",
                "                        resultdir=evaluation_resultdir):\n",
                "    frames = []\n",
                "    for label, filename in experiments.items():\n",
                "        frame = eh.data.read_csv(resultdir / filename / \"stats.csv.gz\" )\n",
                "        frames.append(\n",
                "            frame\n",
                "            .assign(experiment=label, iteration=lambda df: df['iteration'] + 1)\n",
                "        )\n",
                "\n",
                "    return pd.concat(frames, ignore_index=True)"
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "# Figure 11a: Periodic Patterns"
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 8,
            "metadata": {},
            "outputs": [
                {
                    "name": "stderr",
                    "output_type": "stream",
                    "text": [
                        "/tmp/ipykernel_6105/1076462417.py:16: FutureWarning: The default value of numeric_only in DataFrameGroupBy.mean is deprecated. In a future version, numeric_only will default to False. Either specify numeric_only or select only columns which should be valid for the function.\n",
                        "  .mean()\n",
                        "/tmp/ipykernel_6105/1076462417.py:16: FutureWarning: The default value of numeric_only in DataFrameGroupBy.mean is deprecated. In a future version, numeric_only will default to False. Either specify numeric_only or select only columns which should be valid for the function.\n",
                        "  .mean()\n",
                        "/tmp/ipykernel_6105/1076462417.py:16: FutureWarning: The default value of numeric_only in DataFrameGroupBy.mean is deprecated. In a future version, numeric_only will default to False. Either specify numeric_only or select only columns which should be valid for the function.\n",
                        "  .mean()\n",
                        "/tmp/ipykernel_6105/1076462417.py:16: FutureWarning: The default value of numeric_only in DataFrameGroupBy.mean is deprecated. In a future version, numeric_only will default to False. Either specify numeric_only or select only columns which should be valid for the function.\n",
                        "  .mean()\n"
                    ]
                },
                {
                    "data": {
                        "image/png": "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",
                        "text/plain": [
                            "<Figure size 456x322.441 with 1 Axes>"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                }
            ],
            "source": [
                "periodic_markers = [(1, \"W1+3\"), (6, \"W1\"), (11, \"W1+3\"), (16, \"W1\")]\n",
                "frame = (\n",
                "    load_workload_results({\n",
                "        \"Memento\": \"periodic/memento\",\n",
                "        #\"Memento(T=0)\": \"periodic/memento-deterministic\",\n",
                "        \"Random\": \"periodic/random\",\n",
                "        \"FIFO\": \"periodic/fifo\",\n",
                "        \"LARS\": \"periodic/lars\",\n",
                "    })\n",
                "    .rename(columns={'experiment': \"Memory\"})\n",
                ")\n",
                "sns.relplot(\n",
                "    data=frame,\n",
                "    x='iteration',\n",
                "    y='accuracy',\n",
                "    hue='Memory',\n",
                "    style=\"Memory\",\n",
                "    kind='line',\n",
                "    palette=palette,\n",
                "    **eh.plot_utils.opts_onethird()\n",
                ").set(\n",
                "    xlabel=\"Iteration\",\n",
                "    ylabel=\"Accuracy\",\n",
                "    ylim=(0, 1.0),\n",
                "    xlim=(0.9, frame['iteration'].max()),\n",
                ")\n",
                "#plt.gca().get_legend()._set_loc(3)\n",
                "plt.legend(\n",
                "    title=None, loc=\"lower left\",\n",
                "    ncol=2, columnspacing=1.5, handletextpad=0.5, handlelength=1.5,\n",
                ")\n",
                "\n",
                "plot_markers(*periodic_markers)\n",
                "plt.savefig(plotdir / \"periodic_peformance.pdf\")\n",
                "plt.show()"
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "# Figure 11b: Incremental Learning"
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 9,
            "metadata": {},
            "outputs": [
                {
                    "name": "stderr",
                    "output_type": "stream",
                    "text": [
                        "/tmp/ipykernel_6105/1076462417.py:16: FutureWarning: The default value of numeric_only in DataFrameGroupBy.mean is deprecated. In a future version, numeric_only will default to False. Either specify numeric_only or select only columns which should be valid for the function.\n",
                        "  .mean()\n",
                        "/tmp/ipykernel_6105/1076462417.py:16: FutureWarning: The default value of numeric_only in DataFrameGroupBy.mean is deprecated. In a future version, numeric_only will default to False. Either specify numeric_only or select only columns which should be valid for the function.\n",
                        "  .mean()\n",
                        "/tmp/ipykernel_6105/1076462417.py:16: FutureWarning: The default value of numeric_only in DataFrameGroupBy.mean is deprecated. In a future version, numeric_only will default to False. Either specify numeric_only or select only columns which should be valid for the function.\n",
                        "  .mean()\n",
                        "/tmp/ipykernel_6105/1076462417.py:16: FutureWarning: The default value of numeric_only in DataFrameGroupBy.mean is deprecated. In a future version, numeric_only will default to False. Either specify numeric_only or select only columns which should be valid for the function.\n",
                        "  .mean()\n"
                    ]
                },
                {
                    "data": {
                        "image/png": "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",
                        "text/plain": [
                            "<Figure size 456x322.441 with 1 Axes>"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                }
            ],
            "source": [
                "incremental_markers = [(1, \"W1\"), (11, \"W2\"), (21, \"W3\"),]\n",
                "frame = (\n",
                "    load_workload_results({\n",
                "        \"Memento\": \"incremental/memento\",\n",
                "        \"Random\": \"incremental/random\",\n",
                "        #\"Memento(T=0)\": \"incremental/memento-deterministic\",\n",
                "        \"FIFO\": \"incremental/fifo\",\n",
                "        \"LARS\": \"incremental/lars\",\n",
                "    })\n",
                "    .rename(columns={'experiment': \"Memory\"})\n",
                ")\n",
                "sns.relplot(\n",
                "    data=frame,\n",
                "    x='iteration',\n",
                "    y='accuracy',\n",
                "    hue='Memory',\n",
                "    style=\"Memory\",\n",
                "    kind='line',\n",
                "    palette=palette,\n",
                "    #estimator=np.min,\n",
                "    errorbar=None,  # Not useful for max three entries per iteration.\n",
                "    **eh.plot_utils.opts_onethird()\n",
                ").set(\n",
                "    xlabel=\"Iteration\",\n",
                "    ylabel=\"Accuracy\",\n",
                "    ylim=(0, 1.0),\n",
                "    xlim=(0.9, frame['iteration'].max()),\n",
                ")\n",
                "plt.legend(\n",
                "    title=None, loc=\"lower left\",\n",
                "    ncol=2, columnspacing=1.5, handletextpad=0.5, handlelength=1.5,\n",
                ")\n",
                "plot_markers(*incremental_markers)\n",
                "plt.savefig(plotdir / \"sequential_peformance.pdf\")\n",
                "plt.show()"
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "Some additional insights into how the algorithms pick up new samples:\n",
                "- as soon as w2 appears, Memento selects enough samples such that there is roughly a 50:50 split between w1 and w2 in the memroy.\n",
                "- as soon as w3 appears, Memento selects samples such that about 45% of Memory are covered by w3. It removes more samples from w2 than from w1 to make space."
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 10,
            "metadata": {},
            "outputs": [
                {
                    "data": {
                        "image/png": "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",
                        "text/plain": [
                            "<Figure size 645.206x500 with 1 Axes>"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                }
            ],
            "source": [
                "sns.relplot(\n",
                "    data=load_stats({\n",
                "        \"Memento\": \"incremental/memento\",\n",
                "    }).melt(id_vars=['iteration', 'experiment'], value_vars=[\"w1\", \"w2\", \"w3\"]),\n",
                "    kind=\"line\",\n",
                "    x='iteration',\n",
                "    y='value',\n",
                "    hue='variable',\n",
                "    style='experiment',\n",
                ");"
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "In terms of samples, Memento does not go for an even split. Yet, the select data allows to learn all workloads well.\n",
                "To gain more insight into Memento's decision, let's look at the estimated coverage of samples space.\n",
                "\n",
                "Memento does not base it's decision on the sample count, but sample _density_.\n",
                "Memento reports changes in the space covered by the selected samples by ntegrating (approximately) over the estimated sample space density.\n",
                "Below we can see the following:\n",
                "- After w2, Memento reports an increase of 50%\n",
                "- After w3, an increase of 33%\n",
                "\n",
                "Note that the increase is computed with respect to the sample space _after_ the update; an increase of 50% means that 50% of the sample space was not covered before.\n",
                "As soon as w2 appears, Memento discoveres that the sample space is twice as big as it knew before, and selects samples such that both the areas of w1 and w2 are equally covered.\n",
                "\n",
                "Similarly, as soon as w3 appears, Memento learns that there is a third of the sample space (the area of w3) that it did not know about before, and selects just enough samples to fill it. It appears that more than 33% of samples are needed to cover this 33% of sample space, maybe because samples of w3 are more diverse?\n"
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 11,
            "metadata": {},
            "outputs": [
                {
                    "data": {
                        "image/png": "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",
                        "text/plain": [
                            "<Figure size 637.2x450.568 with 1 Axes>"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                }
            ],
            "source": [
                "sns.relplot(\n",
                "    data=load_stats({\n",
                "        \"Memento\": \"incremental/memento\",\n",
                "    }),\n",
                "    kind=\"line\",\n",
                "    x='iteration',\n",
                "    y='coverage_increase',\n",
                "    hue='experiment',\n",
                "    style='experiment',\n",
                "    **eh.plot_utils.default_opts\n",
                ").set(\n",
                "    xlabel=\"Iteration\",\n",
                "    ylabel=\"Coverage Increase\",\n",
                "    ylim=(0, 1.0),\n",
                "    xlim=(0.9, frame['iteration'].max()),\n",
                ");\n",
                "# fig = plt.gcf()\n",
                "# fig.suptitle(\n",
                "#     \"Density increase for incremental learning.\"\n",
                "# )\n",
                "# fig.subplots_adjust(top=0.9)\n",
                "plot_markers(*incremental_markers);"
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "# Memory overview\n",
                "\n",
                "At the end of both experiments, how many samples of each workload has Memento ultimately selected?\n",
                "\n",
                "It manages to preseve rare or past patterns better than the alternatives."
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 12,
            "metadata": {},
            "outputs": [
                {
                    "name": "stdout",
                    "output_type": "stream",
                    "text": [
                        "\\begin{tabular}{lrrr}\n",
                        "\\toprule\n",
                        "{} &    W1 &     W2 &    W3 \\\\\n",
                        "\\midrule\n",
                        "\\textit{Rare Patterns} (\\cref{ssec:sim_tail_perf})\\\\\n",
                        "Memento &  5696 &   9053 &  5219 \\\\\n",
                        "Random  &   252 &  19634 &   114 \\\\\n",
                        "LARS    &  3159 &  14231 &  2610 \\\\\n",
                        "FIFO    &     0 &  20000 &     0 \\\\\n",
                        "\\midrule\n",
                        "\\textit{Incremental Learning} (\\cref{ssec:sim_no_forgetting})\\\\\n",
                        "Memento &  5376 &   5888 &   8704 \\\\\n",
                        "Random  &  1266 &  17836 &    898 \\\\\n",
                        "LARS    &  9540 &   9557 &    903 \\\\\n",
                        "FIFO    &     0 &      0 &  20000 \\\\\n",
                        "\\bottomrule\n",
                        "\\end{tabular}\n",
                        "\n"
                    ]
                },
                {
                    "name": "stderr",
                    "output_type": "stream",
                    "text": [
                        "/tmp/ipykernel_6105/1956666400.py:34: FutureWarning: In future versions `DataFrame.to_latex` is expected to utilise the base implementation of `Styler.to_latex` for formatting and rendering. The arguments signature may therefore change. It is recommended instead to use `DataFrame.style.to_latex` which also contains additional functionality.\n",
                        "  .to_latex()\n",
                        "/tmp/ipykernel_6105/1956666400.py:41: FutureWarning: In future versions `DataFrame.to_latex` is expected to utilise the base implementation of `Styler.to_latex` for formatting and rendering. The arguments signature may therefore change. It is recommended instead to use `DataFrame.style.to_latex` which also contains additional functionality.\n",
                        "  .to_latex()\n"
                    ]
                }
            ],
            "source": [
                "def mem_contents(experiments):\n",
                "    return (\n",
                "        load_stats(experiments)\n",
                "        .loc[\n",
                "            lambda df: df['iteration'] == df['iteration'].max(),\n",
                "            ['experiment', 'w1', 'w2', 'w3']\n",
                "        ]\n",
                "        # Make pretty!\n",
                "        .set_index(\"experiment\")\n",
                "        .rename_axis(index=None)\n",
                "        .rename(columns={'w1': \"W1\", 'w2': \"W2\", 'w3': \"W3\"})\n",
                "    )\n",
                "\n",
                "mem_periodic = mem_contents({\n",
                "    \"Memento\": \"periodic/memento\",\n",
                "    #\"Memento(T=0)\": \"periodic/memento-deterministic\",\n",
                "    \"Random\": \"periodic/random\",\n",
                "    \"LARS\": \"periodic/lars\",\n",
                "    \"FIFO\": \"periodic/fifo\",\n",
                "})\n",
                "mem_incremental = mem_contents({\n",
                "    \"Memento\": \"incremental/memento\",\n",
                "    #\"Memento(T=0)\": \"incremental/memento-deterministic\",\n",
                "    \"Random\": \"incremental/random\",\n",
                "    \"LARS\": \"incremental/lars\",\n",
                "    \"FIFO\": \"incremental/fifo\",\n",
                "})\n",
                "\n",
                "# Stitch both together in a nice latex table.\n",
                "# Also add some intermediate headers.\n",
                "header_periodic = r\"\\textit{Rare Patterns} (\\cref{ssec:sim_tail_perf})\\\\\"\n",
                "latex_periodic = (\n",
                "    mem_periodic\n",
                "    .to_latex()\n",
                "    .replace(\"\\midrule\", f\"\\midrule\\n{header_periodic}\")\n",
                ")\n",
                "header_incremental = \\\n",
                "    r\"\\textit{Incremental Learning} (\\cref{ssec:sim_no_forgetting})\\\\\"\n",
                "latex_incremental = (\n",
                "    mem_incremental\n",
                "    .to_latex()\n",
                "    .replace(\"\\midrule\", f\"\\midrule\\n{header_incremental}\")\n",
                ")\n",
                "latex_combined = (\n",
                "    latex_periodic.split(r'\\bottomrule')[0] + # cut off end\n",
                "    '\\midrule' + latex_incremental.split('\\midrule')[1] # cut off head\n",
                ")\n",
                "\n",
                "print(latex_combined)\n",
                "with open(plotdir / \"final_mem.tex\", \"w\") as file:\n",
                "    file.write(latex_combined)"
            ]
        },
        {
            "cell_type": "markdown",
            "metadata": {},
            "source": [
                "# Figure 11c: Regression"
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 13,
            "metadata": {},
            "outputs": [
                {
                    "data": {
                        "image/png": "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",
                        "text/plain": [
                            "<Figure size 456x322.441 with 1 Axes>"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                }
            ],
            "source": [
                "congestion_markers = [\n",
                "    (1, \"66\\%\"), (11, \"100\\%\"), (21, \"133\\%\"),\n",
                "]\n",
                "frame = (\n",
                "    load_transtime_results({\n",
                "        \"Memento\": \"congestion/memento\",\n",
                "        #\"Memento(T=0)\": \"congestion/memento-deterministic\",\n",
                "        \"Random\": \"congestion/random\",\n",
                "        \"FIFO\": \"congestion/fifo\",\n",
                "    }, metric='0.99', experiment=\"all\")\n",
                "    .rename(columns={'experiment': \"Memory\"})\n",
                ")\n",
                "sns.relplot(\n",
                "    data=frame,\n",
                "    x='iteration',\n",
                "    y='value',\n",
                "    hue='Memory',\n",
                "    style=\"Memory\",\n",
                "    kind='line',\n",
                "    palette=palette,\n",
                "    #drawstyle='steps-post',\n",
                "    #estimator=lambda x: np.quantile(x, 0.99),\n",
                "    errorbar=None,  # Does not work with steps.\n",
                "    **eh.plot_utils.opts_onethird()\n",
                ").set(\n",
                "    xlabel=\"Iteration\",\n",
                "    ylabel=\"99th-percentile Error [ms]\",\n",
                "    #yscale='log',\n",
                "    ylim=(10, 35),\n",
                "    xlim=(0.9, frame['iteration'].max()),\n",
                ")\n",
                "plt.legend(\n",
                "    title=None, ncol=3, loc=\"lower left\",\n",
                "    columnspacing=0.75, handletextpad=0.45, handlelength=1.25\n",
                ")\n",
                "plot_markers(*congestion_markers)\n",
                "plt.savefig(plotdir / \"congestion_performance.pdf\")\n",
                "plt.show()"
            ]
        }
    ],
    "metadata": {
        "interpreter": {
            "hash": "f78f8e3aabf3906a5ed8d5dc64bda1959ee71737771fe5fec1226ba5e4787360"
        },
        "kernelspec": {
            "display_name": "Python 3.9.7 64-bit ('env': venv)",
            "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": "3.10.11"
        },
        "orig_nbformat": 2
    },
    "nbformat": 4,
    "nbformat_minor": 2
}