fps_bandwithlimit_analysis.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "801232fe",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "import os\n",
    "import re"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\baseline0\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\baseline1\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\baseline2\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\record0\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\record1\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\record2\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay0.0\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay0.1\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay0.2\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay0.3\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay0.4\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay0.5\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay0.6\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay0.7\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay0.8\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay0.9\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay1.0\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay1.1\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay1.2\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay1.3\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay1.4\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay1.5\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay1.6\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay1.7\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay1.8\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay1.9\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay2.0\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay2.1\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay2.2\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay2.3\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay2.4\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay2.5\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay2.6\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay2.7\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay2.8\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-2\\BeatSaber\\PopStars-Medium\\replay2.9\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\baseline0\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\baseline1\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\baseline2\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\record0\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\record1\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\record2\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay0.0\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay0.1\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay0.2\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay0.3\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay0.4\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay0.5\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay0.6\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay0.7\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay0.8\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay0.9\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay1.0\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay1.1\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay1.2\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay1.3\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay1.4\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay1.5\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay1.6\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay1.7\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay1.8\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay1.9\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay2.0\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay2.1\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay2.2\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay2.3\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay2.4\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay2.5\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay2.6\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay2.7\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay2.8\\logcat_VrApi.log\n",
      "./overhead-validation\\Quest-Pro\\BeatSaber\\PopStars-Medium\\replay2.9\\logcat_VrApi.log\n"
     ]
    }
   ],
   "source": [
    "def reset_timestamp(df):\n",
    "    df['timestamp'] = df['timestamp'] - df['timestamp'][0]\n",
    "    return df\n",
    "\n",
    "def process_line(line):\n",
    "    timestamp = line.split(' ')[1]\n",
    "    content = line.split('  : ')[1]\n",
    "    pairs = content.split(',')\n",
    "\n",
    "    d = {}\n",
    "\n",
    "    cpugpu = re.search(\"CPU4/GPU=\\d+/\\d+,\\d+/\\d+\", content).group()\n",
    "    cpugpu = cpugpu.split('=')[1]\n",
    "    cpugpu = cpugpu.split('/')\n",
    "\n",
    "    d['CPU Level'] = int(cpugpu[0])\n",
    "    d['GPU Level'] = int(cpugpu[1].split(',')[0])\n",
    "    d['CPU Frequency (MHz)'] = int(cpugpu[1].split(',')[1])\n",
    "    d['GPU Frequency (MHz)'] = int(cpugpu[2])\n",
    "    \n",
    "    for x in pairs:\n",
    "        tmp = x.split('=')\n",
    "        if tmp[0] == 'CPU4/GPU' or len(tmp) < 2:\n",
    "            continue\n",
    "        if tmp[0] == 'FPS':\n",
    "            d[tmp[0]] = int(tmp[1].split('/')[0])\n",
    "        elif tmp[0] == 'Free':\n",
    "            d[tmp[0]] = int(tmp[1][:-2])\n",
    "        else:\n",
    "            d[tmp[0]] = tmp[1].strip()\n",
    "\n",
    "    d[\"timestamp\"] = timestamp\n",
    "\n",
    "    return pd.DataFrame(d, index=[0])\n",
    "\n",
    "def add_to_dataframe(df, timestamp, key, value):\n",
    "    if key not in df.columns:\n",
    "        df[key] = np.nan\n",
    "    df.loc[timestamp, key] = value\n",
    "    return df\n",
    "\n",
    "def create_dataframe(file_path):\n",
    "    df = pd.DataFrame()\n",
    "    for line in open(file_path, 'r', encoding='utf-16'):\n",
    "        if len(line) < 300:\n",
    "            continue\n",
    "        new_d = process_line(line)\n",
    "        df = pd.concat([df, new_d])\n",
    "\n",
    "    folder = file_path.split('\\\\')\n",
    "    folder[0] = folder[0][19:-4]\n",
    "\n",
    "    df['experiment'] = folder[0]\n",
    "    df['device'] = folder[1]\n",
    "    df['repetition'] = folder[4]\n",
    "\n",
    "    df.to_csv(f'.{file_path.strip(\".log\")}.csv', index=False)\n",
    "\n",
    "    return df\n",
    "\n",
    "def process_logcat():\n",
    "    df_aggregated = pd.DataFrame()\n",
    "    for root, dirs, files in os.walk(\"./\", topdown=False):\n",
    "        for name in files:\n",
    "            if(name == 'logcat_VrApi.log') and (\"clumsy-bandwidth\" in os.path.join(root, name)):\n",
    "                # print(os.path.join(root, name))\n",
    "                df_aggregated = pd.concat([df_aggregated, create_dataframe(os.path.join(root, name))])\n",
    "\n",
    "            if (name == \"logcat_VrApi.log\") and (\"overhead-validation\" in os.path.join(root, name)):\n",
    "                print(os.path.join(root, name))\n",
    "                df_aggregated = pd.concat([df_aggregated, create_dataframe(os.path.join(root, name))])\n",
    "\n",
    "\n",
    "    return df_aggregated\n",
    "\n",
    "df_aggregate = process_logcat()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "ddf74dba",
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.rcParams['font.family'] = 'Arial'\n",
    "custom_params = {\"axes.spines.right\": False, \n",
    "                 \"axes.spines.top\": False,\n",
    "                 \"axes.spines.left\": True,\n",
    "                 \"axes.spines.bottom\": True,\n",
    "                #create dashes next to ticks\n",
    "                 \"xtick.bottom\": True,\n",
    "                 \"ytick.left\": True,\n",
    "                \n",
    "                 \"axes.edgecolor\": \"black\",\n",
    "\n",
    "                 \"axes.grid\": True,\n",
    "                 \"axes.linewidth\": 1.5, \n",
    "                 \"axes.facecolor\": \"white\", \n",
    "                 \"grid.color\": \"lightgray\",\n",
    "                 \n",
    "                 'axes.labelsize': 14, \n",
    "                 'axes.titlesize': 13,\n",
    "                 'xtick.labelsize': 13, \n",
    "                 'ytick.labelsize': 13\n",
    "                 }\n",
    "# custom_params = {}\n",
    "\n",
    "# sns.set_theme(style=\"whitegrid\", rc=custom_params)\n",
    "pallet = sns.set_palette(\"deep\")\n",
    "\n",
    "custom_mean = {\"marker\":\"d\",\n",
    "                \"markerfacecolor\":\"white\", \n",
    "                \"markeredgecolor\":\"black\",\n",
    "                \"markersize\":\"7\"}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "51f50cbc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "70.0\n",
      "72.0\n",
      "25.0\n",
      "25.885\n",
      "25.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 460.8x180 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "def to_mbit(bytes):\n",
    "    return (bytes * 8) / 1048576\n",
    "\n",
    "# depricated the timestamp was at a second rate anyway.\n",
    "def dataframe_add_timestamp(file_path : str, df : pd.DataFrame, col_name :str = 'time'):\n",
    "    with open(file_path, 'r', encoding='utf-16') as file:\n",
    "        data = []\n",
    "        for line in file:\n",
    "            data.append(int(float(line.split(' ')[0])))\n",
    "\n",
    "        df.insert(0, col_name, data)\n",
    "\n",
    "\n",
    "def cumulative_figure():\n",
    "    df_aggregate['experiment'] = df_aggregate['experiment'].replace('', '400')\n",
    "    df_some = df_aggregate[df_aggregate['device'] == 'Quest-2']\n",
    "    experiment = sorted(df_some['experiment'].unique(), key=lambda x: -int(x)) \n",
    "\n",
    "    plt.figure(figsize=(6.4, 2.5))\n",
    "    # plt.title(\"Quest-Pro FPS bandwidth probability density FPS\")\n",
    "    plt.xlim((0, 75))\n",
    "    plt.xlabel(\"Frames per second\")\n",
    "    plt.ylabel(\"Probability Density\")\n",
    "\n",
    "    colors = sns.color_palette(\"deep\", n_colors=100)\n",
    "    line_styles = ['-', '--', '-.', ':']\n",
    "    i = 9\n",
    "    for e in experiment:\n",
    "        df_x = df_some[df_some['experiment'] == e]\n",
    "        df_x = df_x[df_x['repetition'].str.contains(\"replay\", case=False)]\n",
    "        fig = sns.ecdfplot(df_x[\"FPS\"].to_list(), label=e + \"\\u2009Mbps\", color=colors[i], )\n",
    "        print(np.percentile(df_x[\"FPS\"], 0.5))\n",
    "        i += 1\n",
    "        if i == 13: i+=1\n",
    "\n",
    "    \n",
    "    fig.lines[0].set_linestyle(line_styles[1])\n",
    "    fig.lines[1].set_linestyle(line_styles[0])\n",
    "    fig.lines[2].set_linestyle(line_styles[2])\n",
    "    fig.lines[3].set_linestyle(line_styles[3])\n",
    "    fig.lines[4].set_linestyle(line_styles[1])\n",
    "\n",
    "    for e in range(5):\n",
    "        fig.lines[e].set_linewidth(2)\n",
    "    \n",
    "    fig.set_yscale(\"log\")\n",
    "    fig.set_yticks([1e0, 1e-1, 1e-2, 1e-3])\n",
    "    fig.axhline(1e-2, color=\"red\", linestyle=':')\n",
    "    fig.legend()\n",
    "\n",
    "    plt.tight_layout()\n",
    "    plt.savefig(\"./figures/fps-maddness.pdf\")\n",
    "\n",
    "cumulative_figure()\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.10.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}