reports/RQ2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# RQ 2 (Size)\n",
    "\n",
    "> What is the size of unsafe blocks that programmers write?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# import libraries\n",
    "import itertools\n",
    "import functools\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn import tree\n",
    "\n",
    "import utils\n",
    "import top_crates\n",
    "from whitelists import *\n",
    "from drawing import *\n",
    "\n",
    "# Configure pyplot: enlarge plots \n",
    "plt.rcParams['figure.figsize'] = [15/1.5, 8/1.4]\n",
    "# Configure pandas: show all columns when printing a dataframe\n",
    "pd.set_option('display.max_columns', None)\n",
    "# Configure pandas: show more rows when printing a dataframe\n",
    "pd.set_option('display.min_rows', 20)\n",
    "# Font for paper\n",
    "plt.rc('font', family='serif')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load tables\n",
    "unsafe_blocks = utils.load_unsafe_blocks()\n",
    "unsafe_block_sizes = utils.load_unsafe_block_sizes()\n",
    "\n",
    "# Preprocessing\n",
    "user_written_unsafe_blocks = unsafe_blocks[unsafe_blocks.check_mode == \"UnsafeBlockUserProvided\"].copy()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Query 3\n",
    "\n",
    "> how many MIR statements does the compiler generate for user-written unsafe blocks\n",
    "\n",
    "#### Results from the paper for comparison (Fig. 3)\n",
    "\n",
    "![Fig. 3](pics/unsafe_blocks_size_cumulative.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>scope</th>\n",
       "      <th>mir_statements_count</th>\n",
       "      <th>mir_terminators_count</th>\n",
       "      <th>mir_items_count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>104159490</td>\n",
       "      <td>30.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>34.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>104195047</td>\n",
       "      <td>12.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>14.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>104196217</td>\n",
       "      <td>57.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>69.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>104152718</td>\n",
       "      <td>6.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>104163892</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>104153998</td>\n",
       "      <td>17.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>17.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>104163922</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>104198456</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>104196212</td>\n",
       "      <td>60.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>72.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>104213536</td>\n",
       "      <td>21.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>25.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>674967</th>\n",
       "      <td>156422394</td>\n",
       "      <td>26.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>32.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>674968</th>\n",
       "      <td>156422398</td>\n",
       "      <td>26.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>32.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>674969</th>\n",
       "      <td>156422402</td>\n",
       "      <td>26.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>32.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>674970</th>\n",
       "      <td>156422406</td>\n",
       "      <td>24.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>30.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>674971</th>\n",
       "      <td>156422421</td>\n",
       "      <td>32.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>39.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>674972</th>\n",
       "      <td>156422412</td>\n",
       "      <td>23.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>29.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>674973</th>\n",
       "      <td>156422426</td>\n",
       "      <td>54.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>64.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>674974</th>\n",
       "      <td>156422432</td>\n",
       "      <td>23.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>29.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>674995</th>\n",
       "      <td>156423802</td>\n",
       "      <td>30.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>33.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>674996</th>\n",
       "      <td>156423808</td>\n",
       "      <td>42.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>48.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>319600 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            scope  mir_statements_count  mir_terminators_count  \\\n",
       "0       104159490                  30.0                    4.0   \n",
       "11      104195047                  12.0                    2.0   \n",
       "16      104196217                  57.0                   12.0   \n",
       "17      104152718                   6.0                    1.0   \n",
       "22      104163892                   0.0                    1.0   \n",
       "23      104153998                  17.0                    0.0   \n",
       "24      104163922                   0.0                    1.0   \n",
       "25      104198456                   6.0                    0.0   \n",
       "26      104196212                  60.0                   12.0   \n",
       "27      104213536                  21.0                    4.0   \n",
       "...           ...                   ...                    ...   \n",
       "674967  156422394                  26.0                    6.0   \n",
       "674968  156422398                  26.0                    6.0   \n",
       "674969  156422402                  26.0                    6.0   \n",
       "674970  156422406                  24.0                    6.0   \n",
       "674971  156422421                  32.0                    7.0   \n",
       "674972  156422412                  23.0                    6.0   \n",
       "674973  156422426                  54.0                   10.0   \n",
       "674974  156422432                  23.0                    6.0   \n",
       "674995  156423802                  30.0                    3.0   \n",
       "674996  156423808                  42.0                    6.0   \n",
       "\n",
       "        mir_items_count  \n",
       "0                  34.0  \n",
       "11                 14.0  \n",
       "16                 69.0  \n",
       "17                  7.0  \n",
       "22                  1.0  \n",
       "23                 17.0  \n",
       "24                  1.0  \n",
       "25                  6.0  \n",
       "26                 72.0  \n",
       "27                 25.0  \n",
       "...                 ...  \n",
       "674967             32.0  \n",
       "674968             32.0  \n",
       "674969             32.0  \n",
       "674970             30.0  \n",
       "674971             39.0  \n",
       "674972             29.0  \n",
       "674973             64.0  \n",
       "674974             29.0  \n",
       "674995             33.0  \n",
       "674996             48.0  \n",
       "\n",
       "[319600 rows x 4 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Count MIR elements\n",
    "sizes_df = user_written_unsafe_blocks[['scope']].join(\n",
    "    unsafe_block_sizes[['unsafe_scope', 'statement_count', 'terminator_count']].set_index('unsafe_scope'),\n",
    "    on='scope',\n",
    "    how='left',\n",
    ")\n",
    "user_written_unsafe_blocks[\"mir_statements_count\"] = sizes_df.statement_count\n",
    "user_written_unsafe_blocks[\"mir_terminators_count\"] = sizes_df.terminator_count\n",
    "user_written_unsafe_blocks[\"mir_items_count\"] = sizes_df.statement_count + sizes_df.terminator_count\n",
    "user_written_unsafe_blocks[[\"scope\", \"mir_statements_count\", \"mir_terminators_count\", \"mir_items_count\"]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiEAAAFACAYAAABqV6zlAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAYhUlEQVR4nO3df5BdZ33f8fdHFpUtY1mALbQUbBEQcUJBTrEn9djBECuEGWfTYPJjhtSZTD0omEACLRgFPB65VVNoIZ06SeNqQpJOwlCKnWKEUqjsdmSimgaZFicUSBgbQ2nDGodKQsKC2t/+sWc7V5td6a60Z889Z9+vmZ2959k9Z7/Pnnu1Hz3Pc85NVSFJkrTS1nRdgCRJWp0MIZIkqROGEEmS1AlDiCRJ6oQhRJIkdcIQIkmSOrG26wLO1EUXXVRbtmzpugxJknQaDz744Deq6uL57b0NIVu2bOHQoUNdlyFJkk4jyaMLtfduOibJdJI9hw8f7roUSZJ0FnoXQqpqb1XtuPDCC7suRZIknYXehRBJkjQMhhBJktSJ3oUQ14RIkjQMvQshrgmRJGkYehdCJEnSMBhCJElSJwwhkiSpE4YQSZLUid6FEK+OkSRpGHoXQuaujvnKt7quRJIknY3ehRBJkjQMhhBJktQJQ4gkSeqEIUSSJHXCECJJkjphCJEkSZ0whEiSpE6s7bqApUoyDUyv3TjVdSmSJOks9G4kZO5mZWvOPb/rUiRJ0lnoXQiRJEnDYAiRJEmdMIRIkqROGEIkSVInDCGSJKkThhBJktQJQ4gkSeqEIUSSJHXCECJJkjoxESEkyZYke5P8dpLXdV2PJElqX2vvHZNkM7Ab2FZVV460bwduAGaAqqrbmy99HngY+NO2apIkSZOjzTewuwa4B7h8riHJeuBO4MVVdSLJ3UmuA+4HdgHfBj4GXN9iXZIkaQK0Nh1TVXcBR+c1XwU8WlUnmu2DzAaOrcCaqipOEYyS7EhyKMmhJ48fbqNsSZK0QtocCVnIJk4OJkeatucAP5/kK8AfLrZzVe0B9gCsm9paLdYpSZJattIhZAa4YGR7AzBTVfcC965wLZIkqUMrfXXMA8ClSdY121cD+5ZygCTTSfY89cSxZS9OkiStnNZCSJJrgRuBqSS3Jjmvqo4DNwN3JNkNPFRV9y3luFW1t6p2rDn3/BaqliRJKyWza0H7I8k0ML1249Trv/vN/9V1OZIk6TSSPFhVV8xvn4iblS2FIyGSJA1D70KIJEkahpW+OuasjUzHdF2KJEk6C70bCXE6RpKkYehdCJEkScPgdIwkSepE70ZCnI6RJGkYehdCJEnSMBhCJElSJ1wTIkmSOtG7kRDXhEiSNAy9CyGSJGkYDCGSJKkThhBJktQJF6ZKkqRO9G4kxIWpkiQNQ+9CiCRJGgZDiCRJ6oQhRJIkdcIQIkmSOuHVMZIkqRO9Gwnx6hhJkoahdyFEkiQNgyFEkiR1whAiSZI6YQiRJEmdMIRIkqROGEIkSVInvE+IJEnqRO9GQrxPiCRJw9C7ECJJkoah1yFky859XZcgSZLOUK9DiCRJ6i9DiCRJ6oQhRJIkdcIQIkmSOmEIkSRJnTCESJKkThhCJElSJyYqhCT5z0mu6boOSZLUvokJIUleBRzrug5JkrQyWnsDuySbgd3Atqq6cqR9O3ADMANUVd2eJMAVwKG26pEkSZOlzXfRvQa4B7h8riHJeuBO4MVVdSLJ3UmuAzYCHwF+usV6JEnSBGkthFTVXUleMa/5KuDRqjrRbB8Erge+BlzM7GjI+Um+WFWPzT9mkh3ADoBzNlzcVumSJGkFrPSakE3A0ZHtI8Cmqnof8HHgKeBJ4PBCO1fVnqq6oqquOGf9ha0XK0mS2tPmdMxCZoALRrY3NG1U1ZeBHz/dAZJMA9NrN061UZ8kSVohKz0S8gBwaZJ1zfbVwL6lHKCq9lbVjjXnnr/sxUmSpJXT5tUx1wI3AlNJbgXeV1XHk9wM3JHkMeChqrpvicd1JESSpAFoc2HqAeDAAu37gf1ncdy9wN51U1tffxblSZKkjk3MzcrO1JadS5rNkSRJE2KlF6aeNadjJEkaht6NhLgwVZKkYehdCJEkScPgdIwkSepE70ZCnI6RJGkYehdCJEnSMBhCJElSJ1wTIkmSOtG7kRDXhEiSNAy9CyGSJGkYDCGSJKkThhBJktQJF6ZKkqRO9G4kxIWpkiQNQ+9CiCRJGgZDiCRJ6oQhRJIkdcIQIkmSOuHVMZIkqRO9Gwnx6hhJkoahdyFEkiQNgyFEkiR1whAiSZI6YQiRJEmdMIRIkqROGEIkSVInvE+IJEnqRO9GQrxPiCRJwzBWCEny8rYLkSRJq8u4IyHvS3JLkme1Wo0kSVo1xl0T8kvA14BfTnIe8PtV9VB7ZUmSpKEbdyTkq1X1FeAAsA34QJLfSPJD7ZUmSZKGbNwQ8gdJPsvsiMh7quolVfUmYLq90sa3Zee+rkuQJElLNO50zAngNVX18FxDkr8B/M1WqpIkSYM37kjIu4ACSPKiJOdU1Xeq6mfbK02SJA3ZuCHkV4DnNY+ngF9tpxxJkrRajBtCPlVV9wNU1QHgm+2VJEmSVoNxQ8iWJGsBms+XtFeSJElaDcZdmPoJ4JEkjwPPBH5xuQtJsg24EtgAbKyq25b7Z0iSpMkxVgipqo8muR94IfClqvo/4+yXZDOwG9hWVVeOtG8HbgBmZg9ft1fVZ5McBd4G/Psl9kOSJPXMUt7ALsyGhg1Jdo25zzXAPc2+swdJ1gN3Am+tql3AS5NcB9BcAnwL8IYl1CVJknporJGQJO8HfpDZEBJm14TsOt1+VXVXklfMa74KeLSqTjTbB4Hrk6ytqk9U1beSXLBIHTuAHQDnbLh4nNIlSdKEGndNyNOr6m/NbSwQLJZiE3B0ZPtI03ZxkncCTwG/t9COVbUH2AOwbmprnUUNkiSpY+OGkE8neXpVfavZfsZZ/MwZYHSkYwMwU1V/MM7OSaaB6bUbp86iBEmS1LVx14S8GZhJ8kiSR4DfPouf+QBwaZJ1zfbVwNhv/lJVe6tqx5pzzz+LEiRJUtfGDSEfrKr1VfX8qno+s4tHTyvJtcCNwFSSW5OcV1XHgZuBO5LsBh6qqvvOqHpJktRb416iuzPJGmbvEfJ4Vb1/zP0OAAcWaN8P7F9KoXOcjpEkaRjGGglJ8irgYeB3gNcl+YVWqzqFxaZjtuwce0ZHkiRNgHGnY6aBy4CDVfUB4AXtlXRqSaaT7HnqiWNdlSBJkpbBuCHkf1bVE8DcZbFj3TG1DS5MlSRpGMa9RPdFSXYClyV5E/DcFmuSJEmrwLgjIW9h9n4eFwGbGfPqGEmSpMWMe3XMUeCdc9tJLgO+0FZRp+LVMZIkDcO47x1z27ymlwPbl7+c06uqvcDedVNbX9/Fz5ckSctj3OmYHwAebT6KjkZBJEnScIy7MHVHVT02t9EsTu2E0zGSJA3DuCMh35fk5c3HjwE/0mZRp+IlupIkDcO4IyH/EvhvQICjwO7WKpIkSavCuCHkDVX1X1utRJIkrSrjTsf8VJLnJblk3se7W61uiXz/GEmS+mPckZBXAz/J7NUxW4BvAEeAS4CdrVS2CBemSpI0DOOOhHwIeH5VXQs8H/i3VfVK4M2tVbYIF6ZKkjQM44aQC6uqAKrqKeA5zeM/aqswSZI0bONOxzw7yW8CfwG8iNn3kZEkSTpj446E3AR8DtjafL6ptYokSdKqMFYIqarvAP8O+C3g31TViVarOkteJSNJ0uQbK4Qk+XvAJ4HbgOuSvKvVqk5dy3SSPU89cWzBrxtAJEnqh3GnY7ZV1fcBD1bVPcD6Fms6Ja+OkSRpGMYNIYebz9V8nujpGEmSNPnGDSHPTnIncHWS9wIXtFjTsnBaRpKkyTZuCHkL8Bngq8AXgXe0VpEkSVoVxr1PyEHgjVW1p81iJEnS6jHuSMhDVfWZuY0kz2qpHkmStEqMG0K+muTVSS5NcglOx0iSpLM07nTMG4EvjGxfAtyy/OVIkqTV4pQhJMmvAfcCv1JVvzfSvr3luk5V0zQwvXbjVFclSJKkZXC66ZgTwH3AtiS/luSFAFV1b+uVLcKblUmSNAynCyFPNO8T8zZgTVV9aQVqkiRJq8DpQkgBVNWTwFNzjUle22ZRkiRp+E63MPVHkzy9efxDSf5Z8/jvAHe3V5YkSRq604WQ7wBzb1f7sZH277ZTjiRJWi1OF0JuqapPz29M8rKW6pEkSavEKdeELBRAmvYH2ylHkiStFuPeMVWSJGlZGUIkSVInBh1Ctuzc13UJkiRpEeO+d0zrkvw4cBnwNODPq+rDHZckSZJa1GoISbIZ2A1sq6orR9q3AzcAM0BV1e3Ag1X10SQXAu8HDCGSJA1Y2yMh1wD3AJfPNSRZD9wJvLiqTiS5O8l1VXVf8y2vAd7bcl2SJKljra4Jqaq7gKPzmq8CHm3ekwbgIHA9QJLrgYeBry10vCQ7khxKcujJ44dbqlqSJK2ELtaEbOLkYHIE2JTkJ4B3AJ8FLgB+dv6OVbUH2AOwbmprtV+qJElqSxdXx8wwGzLmbABmquojVXVVVb2hqv5aAJmTZDrJnqeeOLbYt5zEK2QkSZpMXYSQB4BLk6xrtq8Gxk4KVbW3qnasOff8VoqTJEkro9UQkuRa4EZgKsmtSc6rquPAzcAdSXYDD40sSpUkSatEq2tCquoAcGCB9v3A/jM5ZpJpYHrtxqmzrE6SJHWpd3dMPZPpGNeFSJI0eXoXQiRJ0jBMzG3bx+V0jCRJw9C7kRCvjpEkaRh6F0IkSdIw9C6ELPVmZXNcnCpJ0mTpXQhxOkaSpGHoXQiRJEnDYAiRJEmd8BJdSZLUid6NhLgmRJKkYehdCFkObV4p41U4kiSNZ1WGEEmS1L1VtSbEUQpJkiZH70ZClnNNiKFEkqTu9C6ESJKkYTCESJKkThhCGk7NSJK0slZtCDF0SJLUrVV1dYwkSZocvRsJ8Y6pkiQNQ+9CiCRJGgZDiCRJ6oQhRJIkdcIQIkmSOmEIkSRJnVj1IcT7hUiS1I1VH0IkSVI3vFmZJEnqRO9GQrxZmSRJw9C7ENIG14VIkrTyDCGSJKkThhBJktQJQ4gkSeqEIWSeufUhrhORJKldhpAFGEAkSWqfIUSSJHXCEDIGR0YkSVp+hhBJktSJiQkhSdYmeVeSPV3XIkmS2jcxIQQ4H/g4k1WTJElqSatvYJdkM7Ab2FZVV460bwduAGaAqqrbq+pwksfbrEeSJE2OtkcdrgHuATLXkGQ9cCfw1qraBbw0yXXjHCzJjiSHkhx68vjhNuo9yUILUrfs3OdCVUmSlkGrIaSq7gKOzmu+Cni0qk402weB68c83p6quqKqrjhn/YXLWKkkSVppXay/2MTJweQIsClJgJ8BvjfJ3+6gLkmStIJaXROyiBnggpHtDcBMVRXwnuZjUUmmgem1G6faq3CEUy+SJLWji5GQB4BLk6xrtq8Gxv5LX1V7q2rHmnPPb6U4SZK0MloNIUmuBW4EppLcmuS8qjoO3AzckWQ38FBV3beEY04n2fPUE8eWvd5TjXo4IiJJ0vJqdTqmqg4ABxZo3w/sP8Nj7gX2rpva+vqzLE+SJHXIG4NJkqRO9C6EtDkdczqjUzJtT884/SNJGrrehRAXpkqSNAy9CyGSJGkYehdCupyOWcjctMlKTtVIkjQEvQshTsdIkjQMvQshkiRpGAwhkiSpE70LIZO2JmSpJmW9yKTUIUlavXoXQlwTIknSMPQuhEiSpGEwhEiSpE4YQiRJUid6F0ImcWGqizyHx3MqSe3rXQhxYaokScPQuxAiSZKGwRAiSZI6YQiRJEmdMIRIkqRO9C6ETOLVMQuZu7pi/udx9hm3XZKkPutdCPHqGEmShqF3IUSSJA2DIUSSJHXCECJJkjphCJEkSZ0whEiSpE4YQiRJUid6F0Im+T4h8+/nsdg9Qrbs3HfKe3+Mfm0p9xkZl/cdGTbPr6S+6F0I8T4hkiQNQ+9CiCRJGgZDiCRJ6oQhRJIkdcIQIkmSOmEIkSRJnTCESJKkThhCJElSJwwhkiSpE4YQSZLUibVdFwCQZD2wC/gK8PWq+nC3FUmSpLa1FkKSbAZ2A9uq6sqR9u3ADcAMUFV1e7P96ar6cJKPAIYQSZIGrs2RkGuAe4DL5xqaEY87gRdX1Ykkdye5Dnge8EDzbee1WJMkSZoQra0Jqaq7gKPzmq8CHq2qE832QeB64KvAxU3bt9uqSZIkTY6VXpi6iZODyZGm7Q+BK5O8CfjAYjsn2ZHkUJJDTx4/3G6lLZj/Futbdu77/23zPy/0/adqG+dYi73F+6ne+n053hZ+obpO17fF+rScdbVpKfUt1sdxj3e6n3UmtSyXtp9bZ+JMfu4k9qMv/P3oVM+BlV6YOgNcMLK9AZipquPALafbuar2AHsA1k1trVYqlCRJK2KlR0IeAC5Nsq7ZvhpYUkxOMp1kz1NPHFv24iRJ0sppLYQkuRa4EZhKcmuS85oRj5uBO5LsBh6qqvuWctyq2ltVO9ace34LVUuSpJXS2nRMVR0ADizQvh/Yf6bHTTINTK/dOHUW1UmSpK717o6pjoRIkjQMvQshkiRpGCbitu1L4XSMJEnD0LuREKdjJEkaht6FEEmSNAxOx0iSpE70biTE6RhJkoahdyFEkiQNQ6r6+RYsSY4CX+y6jhVwEfCNrotYAfZzOFZDH8F+Ds1q6GeXfby0qi6e39i7NSEjvlhVV3RdRNuSHLKfw7Ea+rka+gj2c2hWQz8nsY9Ox0iSpE4YQiRJUif6HEL2dF3ACrGfw7Ia+rka+gj2c2hWQz8nro+9XZgqSZL6rc8jIZIkqcd6eXVMku3ADcAMUFV1e8cljS3JC4DdwGeA5wKPV9U/SrILeMXIt/6Tqtrf7PN2YAPwDOA/VtVHm/bLgV8EHgE2AW+rqv+7Ql05rSSfAp5oNp+squuSPBN4N/AwsBV4Z1V9vfn+3vUzyRbgPuCrTdMG4CHgy/T8fCbZzOxzdVtVXdm0Ldv5S3Iu8F7ga82x3l1Vf76CXVysj/8COA58C9gGvKWq/rI51x8H/rLZ/cGq+ofNPhPbx6a+hfq5i2V6jk54P/cBo3e3fCnwHGAzPTyfp/gb0s/XZlX16gNYD3wJWNds3w1c13VdS6j/SuDvjmz/D+BlwK5Fvv8HgT9qHj8N+AtgIxDgz4DNzdfeB9zUdf/m1f7X+gTcCfx083ga+P0+9xN4FrB9ZPt24JohnE/gJ5tzdKiN8wfsBG5pHr8E+OSE9HH3yON3AL/ePN4C/PwCx5joPp6in8v2HJ3wfv7MyOPvAf51n88ni/8N6eVrs4/TMVcBj1bViWb7IHB9h/UsSVV9uqruGWlaAxwDSPKuJG9L8o4k65uv/xjwQLPvd4HPAy9n9sV0XlXNpfhJ/D28pOnLriRztV1P0x9OrrmX/ayqx6vqXoAk64ArquqPm+1en8+qugs4Oq95Oc/f9SP7/CmwLcmG5e/J4hbqY1XdOrK5htkRkTnTSd6e5B8n+f6mbaL72Pzshc7lcj5HJ7afVfWhkc1fAn59ZLt35/MUf0N6+drs43TMJk5+kh1p2nonyWuAT1TVF5J8GPhyVR1L8kZmXyg3Mdu3z4/sNtffx5j838N7qupPkpwD3N/c5Xb0/B0BnpFkLf3u55zXAR9sHg/xfMLynr/FXstHlr/spUuyEXgV8Nqm6THgtqr6XJJnA59K8gOc+t+kSe7jcj5HJ7mfADR/RC+pqj9rmnp/Puf9Denla7OPIyEzwAUj2xuatl5J8krglcBbAarqc1V1rPnyfwJ+uHm8WH8n/vdQVX/SfH4S+CSz/R2tewPwzZpd99Dbfo74KeBDMMzz2VjO8zexfU5yIfCvgL9fVX8FUFXHqupzzeOvA19nds1IL/u4zM/Rie3niJuA35nb6Pv5nP83hJ6+NvsYQh4ALm2GvgGuBvZ1WM+SNVMTPwr8MrA5yVVJ/vnIt2xldt0LwMeYnYKiSbXfD9zP7OKjbzcLsWDCfg9JLkty00jTXJ/20fSHk2vuZT/nNP8g/JdmuJOhnc8Ry3n+9o3s8xLgs1XV+f+ck1wE/Cbw9qp6JMlrm/afa+okydOYXRT4ZXrYR1j25+jE9hMgyRpm/83dN9LW2/O50N8Qevra7OV9QpL8CLMLkB4Dvlv9ujrmZcAB4FDTdD6z/+B9L7OLbmeYXQh0WzWrkZuVzc9oPv5Dnbyy+c3Ao8AzmayrRp7DbL8+w2yKfhrwD5hdEPUeZmt+AbCzTl7B3at+zknyQeDNVfWNZvuf0vPzmeRa4OeAVwO/xezCtfNYpvOX5DxmV+D/b+CFwK/Wyl9psFAfDzI7Vf1XzbcdrarpJD8M/ALw35t6/7iqfrc5zsT2salvoX7exjI9Rye5n1X17SQ/ATy3qn5j5Ht7eT5P8Tfko/TwtdnLECJJkvqvj9MxkiRpAAwhkiSpE4YQSZLUCUOIJEnqhCFEkiR1whAiSZI6YQiRJEmdMIRIkqRO/D8CoPRqQvDyWwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 635.294x384 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 635.294x384 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 635.294x384 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The portion of unsafe blocks covered in the depicted graph: 97.43147684605758 %\n",
      "The portion of unsafe blocks with #21 MIR or less: 75.50657071339174 %\n",
      "Mean size of a unsafe block: 22.005980951187418\n",
      "Median size of a unsafe block: 10.0\n"
     ]
    }
   ],
   "source": [
    "# Plot histograms\n",
    "\n",
    "# Configure pyplot: enlarge plots \n",
    "plt.rcParams['figure.figsize'] = [15/1.7, 8/1.5]\n",
    "\n",
    "user_written_unsafe_blocks.mir_items_count.plot.hist(\n",
    "    bins=500,\n",
    "    log=True,\n",
    "    xlim=(0, user_written_unsafe_blocks.mir_items_count.max())\n",
    ")\n",
    "plt.show()\n",
    "\n",
    "user_written_unsafe_blocks.mir_items_count.describe()\n",
    "\n",
    "user_written_unsafe_blocks[\n",
    "    user_written_unsafe_blocks.mir_items_count <= 100\n",
    "].mir_items_count.plot.hist(\n",
    "    bins=100,\n",
    "    xlim=(0, 100)\n",
    ")\n",
    "plt.show()\n",
    "\n",
    "ax = user_written_unsafe_blocks.mir_items_count.plot.hist(\n",
    "    cumulative=True,\n",
    "    density=1,\n",
    "    bins=int(user_written_unsafe_blocks.mir_items_count.max()),\n",
    "    xlim=(0, 100),\n",
    "    ylim=(0, 1),\n",
    "    histtype=\"step\",\n",
    "    linewidth=2,\n",
    "    fontsize=12,\n",
    "    # title=\"Cumulative distribution\",\n",
    ")\n",
    "ax.title.set_size(20)\n",
    "ax.set_xlabel(\"Unsafe block size (MIR items)\", fontsize=16)\n",
    "ax.set_ylabel(\"Percentage of unsafe blocks\", fontsize=16)\n",
    "ax.set_axisbelow(True)\n",
    "ax.set_yticklabels(['{:,.0%}'.format(x) for x in ax.get_yticks()])\n",
    "ax.set_xticks(range(0, 100, 10), minor=True)\n",
    "ax.set_xticks(range(0, 100, 20))\n",
    "ax.grid(True, linestyle='dotted')\n",
    "plt.show()\n",
    "print(\"The portion of unsafe blocks covered in the depicted graph:\",\n",
    "    100 * len(user_written_unsafe_blocks.query(\"mir_items_count <= 100\"))\n",
    "    /len(user_written_unsafe_blocks), \"%\"\n",
    ")\n",
    "print(\"The portion of unsafe blocks with #21 MIR or less:\",\n",
    "    100 * len(user_written_unsafe_blocks.query(\"mir_items_count <= 21\"))\n",
    "    /len(user_written_unsafe_blocks), \"%\"\n",
    ")\n",
    "print(\"Mean size of a unsafe block:\", user_written_unsafe_blocks.mir_items_count.mean())\n",
    "print(\"Median size of a unsafe block:\", user_written_unsafe_blocks.mir_items_count.median())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Follow-ups"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> there is a significant share, namely 14.4% of tiny unsafe blocks that either wrap an expression (without function calls) or call a single unsafe function without changing any of its arguments"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5.1% of the user-written unsafe blocks just call an unsafe functions, whithout evaluationg expressions for the arguments\n",
      "9.2% of the user-written unsafe blocks just evaluates an expression, whithout calling any function\n",
      "sum: 14.4%\n"
     ]
    }
   ],
   "source": [
    "print(\"{:.1%} of the user-written unsafe blocks just call an unsafe functions, whithout evaluationg expressions for the arguments\".format(\n",
    "    sum(user_written_unsafe_blocks.mir_statements_count == 0) / len(user_written_unsafe_blocks)\n",
    "))\n",
    "print(\"{:.1%} of the user-written unsafe blocks just evaluates an expression, whithout calling any function\".format(\n",
    "    sum(user_written_unsafe_blocks.mir_terminators_count == 0) / len(user_written_unsafe_blocks)\n",
    "))\n",
    "print(\"sum: {:.1%}\".format(\n",
    "    (sum(user_written_unsafe_blocks.mir_statements_count == 0) +\n",
    "     sum(user_written_unsafe_blocks.mir_terminators_count == 0)) /\n",
    "    len(user_written_unsafe_blocks)\n",
    "))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Conversely, there is a small number (78 or 0.02%) of huge outliers whose size ranges from 2000 to 21306 #MIR.\n",
    "\n",
    "Note: 0.02% of unsafe blocks corresponds to 78 of which\n",
    "- 28 are generated by macros, \n",
    "- 16 are generated at compile time, and\n",
    "- 34 appear to be user written."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Count of huge outliers: 78\n",
      "Percentage: 0.024405506883604506\n",
      "Largest: 21306.0\n"
     ]
    }
   ],
   "source": [
    "large_blocks = user_written_unsafe_blocks.query(\"mir_items_count >= 2000\")\n",
    "print(\"Count of huge outliers:\", len(large_blocks))\n",
    "print(\"Percentage:\", 100*len(large_blocks)/len(user_written_unsafe_blocks))\n",
    "print(\"Largest:\", user_written_unsafe_blocks.mir_items_count.max())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "gen by macros: 28\n",
      "numext_fixed_uint_core    13\n",
      "wdg_base32                 8\n",
      "x11_dl                     3\n",
      "wayland_kbd                1\n",
      "udev_dl                    1\n",
      "xkbcommon_dl               1\n",
      "smithay_client_toolkit     1\n",
      "Name: crate, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "data = user_written_unsafe_blocks[\n",
    "    (user_written_unsafe_blocks.mir_items_count >= 2000)\n",
    "    & (user_written_unsafe_blocks.expansion_kind != \"Root\")\n",
    "]\n",
    "print(\"gen by macros:\", len(data))\n",
    "print(data.crate.value_counts())\n",
    "del data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "gen at compile time: 16\n",
      "gdnative_bindings    8\n",
      "gdnative_common      3\n",
      "gdnative_physics     2\n",
      "gdnative_graphics    2\n",
      "gdnative_ui          1\n",
      "Name: crate, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "data = user_written_unsafe_blocks[\n",
    "    (user_written_unsafe_blocks.mir_items_count >= 2000)\n",
    "    & (user_written_unsafe_blocks.expansion_kind == \"Root\")\n",
    "    & (user_written_unsafe_blocks.file_name.str.contains(\"out/\"))\n",
    "]\n",
    "print(\"gen at compile time:\", len(data))\n",
    "print(data.crate.value_counts())\n",
    "del data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "not gen at compile time or by macros: 34\n",
      "aesni               6\n",
      "wdg_base32          3\n",
      "lava                2\n",
      "rgsl                2\n",
      "egaku2d_core        2\n",
      "tract_linalg        1\n",
      "pkcs11              1\n",
      "makepad_render      1\n",
      "j4rs                1\n",
      "rust_smt            1\n",
      "nobs_vk             1\n",
      "tcrab_console_gl    1\n",
      "thorin              1\n",
      "basalt              1\n",
      "parity_daemonize    1\n",
      "bigbro              1\n",
      "adi                 1\n",
      "dvk                 1\n",
      "asi_vulkan          1\n",
      "sanakirja           1\n",
      "milter              1\n",
      "shout               1\n",
      "daemonize_rs        1\n",
      "laz                 1\n",
      "Name: crate, dtype: int64\n",
      "top cases:\n"
     ]
    },
    {
     "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>build</th>\n",
       "      <th>crate</th>\n",
       "      <th>crate_hash</th>\n",
       "      <th>def_path</th>\n",
       "      <th>def_path_hash</th>\n",
       "      <th>summary_key</th>\n",
       "      <th>scope</th>\n",
       "      <th>check_mode</th>\n",
       "      <th>span</th>\n",
       "      <th>expansion_kind</th>\n",
       "      <th>expansion_kind_descr</th>\n",
       "      <th>file_name</th>\n",
       "      <th>line</th>\n",
       "      <th>col</th>\n",
       "      <th>mir_statements_count</th>\n",
       "      <th>mir_terminators_count</th>\n",
       "      <th>mir_items_count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>54156</th>\n",
       "      <td>8020</td>\n",
       "      <td>dvk</td>\n",
       "      <td>446adaf7f589a7b9</td>\n",
       "      <td>dvk[461f]::core[0]::{{impl}}[6]::load[0]</td>\n",
       "      <td>fff9c3790cbd610bc4288608ab1300e5</td>\n",
       "      <td>dvk.core.implement_core.load</td>\n",
       "      <td>12830172</td>\n",
       "      <td>UnsafeBlockUserProvided</td>\n",
       "      <td>11534306</td>\n",
       "      <td>Root</td>\n",
       "      <td>{{root}}</td>\n",
       "      <td>src/lib.rs</td>\n",
       "      <td>3653</td>\n",
       "      <td>12</td>\n",
       "      <td>17688.0</td>\n",
       "      <td>3618.0</td>\n",
       "      <td>21306.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>615997</th>\n",
       "      <td>91191</td>\n",
       "      <td>adi</td>\n",
       "      <td>91fa5fc858510baf</td>\n",
       "      <td>adi[6ba1]::screen[0]::ffi[0]::render[0]::vulka...</td>\n",
       "      <td>1690dec4e692f1503a2fd2c87cfe29e9</td>\n",
       "      <td>adi.screen.ffi.render.vulkan.asi.gpu.implement...</td>\n",
       "      <td>144662276</td>\n",
       "      <td>UnsafeBlockUserProvided</td>\n",
       "      <td>129686113</td>\n",
       "      <td>Root</td>\n",
       "      <td>{{root}}</td>\n",
       "      <td>src/screen/ffi/render/vulkan/asi/gpu/mod.rs</td>\n",
       "      <td>411</td>\n",
       "      <td>8</td>\n",
       "      <td>13023.0</td>\n",
       "      <td>2700.0</td>\n",
       "      <td>15723.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>669908</th>\n",
       "      <td>98256</td>\n",
       "      <td>nobs_vk</td>\n",
       "      <td>e29cff3e36238251</td>\n",
       "      <td>nobs_vk[20e1]::{{impl}}[1]::with_feature[0]</td>\n",
       "      <td>40c01d0e70185b1e84f6313dd6d63467</td>\n",
       "      <td>nobs_vk.implement_1.with_feature</td>\n",
       "      <td>155327699</td>\n",
       "      <td>UnsafeBlockUserProvided</td>\n",
       "      <td>139324513</td>\n",
       "      <td>Root</td>\n",
       "      <td>{{root}}</td>\n",
       "      <td>src/lib.rs</td>\n",
       "      <td>8382</td>\n",
       "      <td>4</td>\n",
       "      <td>10413.0</td>\n",
       "      <td>2194.0</td>\n",
       "      <td>12607.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        build    crate        crate_hash  \\\n",
       "54156    8020      dvk  446adaf7f589a7b9   \n",
       "615997  91191      adi  91fa5fc858510baf   \n",
       "669908  98256  nobs_vk  e29cff3e36238251   \n",
       "\n",
       "                                                 def_path  \\\n",
       "54156            dvk[461f]::core[0]::{{impl}}[6]::load[0]   \n",
       "615997  adi[6ba1]::screen[0]::ffi[0]::render[0]::vulka...   \n",
       "669908        nobs_vk[20e1]::{{impl}}[1]::with_feature[0]   \n",
       "\n",
       "                           def_path_hash  \\\n",
       "54156   fff9c3790cbd610bc4288608ab1300e5   \n",
       "615997  1690dec4e692f1503a2fd2c87cfe29e9   \n",
       "669908  40c01d0e70185b1e84f6313dd6d63467   \n",
       "\n",
       "                                              summary_key      scope  \\\n",
       "54156                        dvk.core.implement_core.load   12830172   \n",
       "615997  adi.screen.ffi.render.vulkan.asi.gpu.implement...  144662276   \n",
       "669908                   nobs_vk.implement_1.with_feature  155327699   \n",
       "\n",
       "                     check_mode       span expansion_kind  \\\n",
       "54156   UnsafeBlockUserProvided   11534306           Root   \n",
       "615997  UnsafeBlockUserProvided  129686113           Root   \n",
       "669908  UnsafeBlockUserProvided  139324513           Root   \n",
       "\n",
       "       expansion_kind_descr                                    file_name  \\\n",
       "54156              {{root}}                                   src/lib.rs   \n",
       "615997             {{root}}  src/screen/ffi/render/vulkan/asi/gpu/mod.rs   \n",
       "669908             {{root}}                                   src/lib.rs   \n",
       "\n",
       "        line  col  mir_statements_count  mir_terminators_count  \\\n",
       "54156   3653   12               17688.0                 3618.0   \n",
       "615997   411    8               13023.0                 2700.0   \n",
       "669908  8382    4               10413.0                 2194.0   \n",
       "\n",
       "        mir_items_count  \n",
       "54156           21306.0  \n",
       "615997          15723.0  \n",
       "669908          12607.0  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = user_written_unsafe_blocks[\n",
    "    (user_written_unsafe_blocks.mir_items_count >= 2000)\n",
    "    & (user_written_unsafe_blocks.expansion_kind == \"Root\")\n",
    "    & (~user_written_unsafe_blocks.file_name.str.contains(\"out/\"))\n",
    "]\n",
    "print(\"not gen at compile time or by macros:\", len(data))\n",
    "print(data.crate.value_counts())\n",
    "print(\"top cases:\")\n",
    "data[\n",
    "    data.mir_items_count >= 12000\n",
    "]"
   ]
  }
 ],
 "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}