Paper.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "57d744dd-1f36-45e7-aefd-3a2ccedb13e1",
   "metadata": {},
   "outputs": [],
   "source": [
    "import glob\n",
    "import json\n",
    "import re\n",
    "from collections import Counter, defaultdict\n",
    "from itertools import combinations\n",
    "from pprint import pprint\n",
    "\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.ticker as mtick\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "from scipy.stats import chisquare, chi2_contingency\n",
    "from statsmodels.stats.inter_rater import aggregate_raters, fleiss_kappa\n",
    "from scipy.stats import bootstrap\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b267a960-404f-40c1-a17d-6359598df720",
   "metadata": {},
   "source": [
    "# Load Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "b71452bc-bce1-4bb9-91a4-072741f4e638",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_bug_id(bug):\n",
    "    # build bug_id identyfier\n",
    "    \n",
    "    bug_id = \"\"\n",
    "    \n",
    "    if 'cve' in bug:\n",
    "        bug_id = bug['cve']['id']\n",
    "        \n",
    "    if not bug_id or \"<missing\" in bug_id:\n",
    "        bug_id = bug['projectName']\n",
    "        \n",
    "    return bug_id\n",
    "\n",
    "\n",
    "def to_user(x):\n",
    "    x = x.replace(\"auto_\", \"\")\n",
    "    if x[0:3] in ['A_1', 'A_4', 'A_7']:\n",
    "        return 'E1'#'PP'\n",
    "    elif x[0:3] in ['A_2', 'A_5']:\n",
    "        return 'E2'#'MF'\n",
    "    elif x[0:3] in ['A_3', 'A_6']:\n",
    "        return 'E3'#'JN'\n",
    "    elif x[0:3] in ['A_7']:\n",
    "         return 'E*1'#'PP'\n",
    "    elif x[0] == 'B':\n",
    "        return 'U1'#'RW'\n",
    "    elif x[0] == 'C':\n",
    "        return 'U2' #'AK'\n",
    "    elif x[0] == 'D':\n",
    "        return 'U3' #'ŁH'\n",
    "    else:\n",
    "        return 'auto'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "bc36189f-4a57-46f9-accf-0d3655fe20ef",
   "metadata": {},
   "outputs": [],
   "source": [
    "bug_count = Counter()\n",
    "\n",
    "reviewer_count = Counter()\n",
    "\n",
    "bugs2 = defaultdict(dict)\n",
    "\n",
    "for fjson in glob.glob(\"2023-10-23_113026/*.json\"):\n",
    "    \n",
    "    with open(fjson) as f:\n",
    "        data = json.load(f)\n",
    "    \n",
    "        if not data:\n",
    "            continue\n",
    "            \n",
    "        for bug in data:\n",
    "            # Do not count model annotations (example annotations shown to users)\n",
    "            if not 'datasetName' in bug['data']: \n",
    "                continue\n",
    "                \n",
    "            # get user id\n",
    "            user_id = fjson.split(\"/\")[1].split(\".\")[0]\n",
    "            \n",
    "            # Count number of Tasks done by rev\n",
    "            reviewer_count[user_id] += 1\n",
    "\n",
    "            bug_id = get_bug_id(bug['data'])\n",
    "                \n",
    "            bug_count[bug_id] += 1\n",
    "            \n",
    "            \n",
    "            if 'datasetName' in bug['data']: \n",
    "                ds = bug['data']['datasetName'] + \"_\"\n",
    "    \n",
    "            bugs2[ds + bug_id][user_id] =  bug\n",
    "\n",
    "        \n",
    "df_bug_count = pd.DataFrame.from_dict(bug_count, orient='index')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "ead52de0-f93a-42c5-8147-6f935f1cbb1c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'E1': {0: 199},\n",
       " 'E2': {0: 211},\n",
       " 'E3': {0: 188},\n",
       " 'U1': {0: 594},\n",
       " 'U2': {0: 596},\n",
       " 'U3': {0: 594}}"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rdf = pd.DataFrame.from_dict(reviewer_count, orient='index')\n",
    "rdf['u'] = rdf.index.map(to_user)\n",
    "rdf.groupby('u').sum().T.to_dict()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "1c1e4147-16df-4857-bb4d-9589dc708293",
   "metadata": {},
   "outputs": [],
   "source": [
    "ab = len(df_bug_count.index.unique())\n",
    "\n",
    "ab3 = len(df_bug_count[df_bug_count[0] >= 3].index.unique())\n",
    "ab2 = len(df_bug_count[df_bug_count[0] >= 2].index.unique())\n",
    "ab1 = len(df_bug_count[df_bug_count[0] >= 1].index.unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "1d7a4ea1-3371-452a-a4e4-40324efb29ad",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "labels = ('unique', 'R >= 2', 'R >= 3')\n",
    "bugs_count = {\n",
    "    'all': np.array([800, 800, 800]),\n",
    "    'reviewd': np.array([ab, ab2, ab3]),\n",
    "\n",
    "}\n",
    "width = 0.6  # the width of the bars: can also be len(x) sequence\n",
    "\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "\n",
    "for rtype, r_count in bugs_count.items():\n",
    "    p = ax.bar(labels, r_count, width, label=labels)\n",
    "    ax.bar_label(p, label_type='center')\n",
    "\n",
    "ax.set_title('Reviewed bugs')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "05bb94d7-d508-4d68-a090-6af6786fd226",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Oznaczenia kategorii dla plików \n",
    "\n",
    "files_categories = defaultdict(lambda: defaultdict(list))\n",
    "#lines_categories = defaultdict(lambda: defaultdict(lambda: defaultdict(list)))\n",
    "lines_categories_an = defaultdict(lambda: defaultdict(lambda : defaultdict(lambda: defaultdict(lambda: defaultdict(list)))))\n",
    "\n",
    "anwsers = defaultdict(lambda: defaultdict(list))\n",
    "\n",
    "records = []\n",
    "records_initial = []\n",
    "\n",
    "\n",
    "def extract_annotated_lines(files, user, fid):\n",
    "    #enum_files = defaultdict(lambda: defaultdict(Counter))\n",
    "    ret = []\n",
    "    for id, file in enumerate(files):\n",
    "        fname = file['fileName']            \n",
    "        fcat = file['category']\n",
    "\n",
    "        for image in file['lines']:\n",
    "            for line in file['lines'][image]:\n",
    "                #enum_files[fname][image][line['lineNumber']] += 1\n",
    "                #lid = enum_files[fname][image][line['lineNumber']] \n",
    "                ret.append((bug, user, fname, fid, fcat, image, line['lineNumber'], line['category']))\n",
    "    return ret\n",
    "    \n",
    "\n",
    "for bug in bugs2:\n",
    "    if len(bugs2[bug]) > 0:\n",
    "        for user in bugs2[bug]:\n",
    "            \n",
    "            anotation = bugs2[bug][user]\n",
    "\n",
    "            for fid, files in enumerate(bugs2[bug][user][\"data\"]['initialAnnotation']['diffsFiles']):\n",
    "                records_initial.extend(extract_annotated_lines(files, \"auto_\"+ user, fid))\n",
    "\n",
    "            for res in anotation['annotations']: \n",
    "                # 0 bug type, 1 rev is sure, 2 - problematic\n",
    "                for fid, files in enumerate(res['result'][3]['value']['diffsFiles']):\n",
    "                    records.extend(extract_annotated_lines(files, user, fid))\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "bbc11728-f04c-4c9b-a8e8-e143aa7112b0",
   "metadata": {},
   "outputs": [],
   "source": [
    "records.extend(records_initial)\n",
    "df = pd.DataFrame(records)\n",
    "df_initial = pd.DataFrame(records_initial)\n",
    "\n",
    "df.columns = ['id', 'bundle', 'file', 'fileid', 'fcat', 'image', 'line', 'annotation']\n",
    "\n",
    "df['user'] = df['bundle'].apply(to_user)\n",
    "df['auto'] = df['bundle'].apply(lambda x: 'auto' in x)\n",
    "\n",
    "df['ds'] = df['id'].apply(lambda x: x.split(\"_\")[0])\n",
    "df['bug'] = df['id'].apply(lambda x: x.split(\"_\")[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "ef54766a-e6f9-48a0-8ad8-acf117d71ce4",
   "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>id</th>\n",
       "      <th>bundle</th>\n",
       "      <th>file</th>\n",
       "      <th>fileid</th>\n",
       "      <th>fcat</th>\n",
       "      <th>image</th>\n",
       "      <th>line</th>\n",
       "      <th>annotation</th>\n",
       "      <th>user</th>\n",
       "      <th>auto</th>\n",
       "      <th>ds</th>\n",
       "      <th>bug</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>B_6_13</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>programming</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>103</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>U1</td>\n",
       "      <td>False</td>\n",
       "      <td>cve</td>\n",
       "      <td>CVE-2020-10289</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>B_6_13</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>programming</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>137</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>U1</td>\n",
       "      <td>False</td>\n",
       "      <td>cve</td>\n",
       "      <td>CVE-2020-10289</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>B_6_13</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>programming</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>103</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>U1</td>\n",
       "      <td>False</td>\n",
       "      <td>cve</td>\n",
       "      <td>CVE-2020-10289</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>B_6_13</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>programming</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>137</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>U1</td>\n",
       "      <td>False</td>\n",
       "      <td>cve</td>\n",
       "      <td>CVE-2020-10289</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>C_4_9</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>programming</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>103</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>U2</td>\n",
       "      <td>False</td>\n",
       "      <td>cve</td>\n",
       "      <td>CVE-2020-10289</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>407713</th>\n",
       "      <td>cve_CVE-2018-16876</td>\n",
       "      <td>auto_C_5_8</td>\n",
       "      <td>lib/ansible/plugins/connection/ssh.py</td>\n",
       "      <td>0</td>\n",
       "      <td>programming</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>361</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>U2</td>\n",
       "      <td>True</td>\n",
       "      <td>cve</td>\n",
       "      <td>CVE-2018-16876</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>407714</th>\n",
       "      <td>cve_CVE-2018-16876</td>\n",
       "      <td>auto_C_5_8</td>\n",
       "      <td>lib/ansible/plugins/connection/ssh.py</td>\n",
       "      <td>0</td>\n",
       "      <td>programming</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>362</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>U2</td>\n",
       "      <td>True</td>\n",
       "      <td>cve</td>\n",
       "      <td>CVE-2018-16876</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>407715</th>\n",
       "      <td>cve_CVE-2018-16876</td>\n",
       "      <td>auto_C_5_8</td>\n",
       "      <td>lib/ansible/plugins/connection/ssh.py</td>\n",
       "      <td>0</td>\n",
       "      <td>programming</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>363</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>U2</td>\n",
       "      <td>True</td>\n",
       "      <td>cve</td>\n",
       "      <td>CVE-2018-16876</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>407716</th>\n",
       "      <td>cve_CVE-2018-16876</td>\n",
       "      <td>auto_C_5_8</td>\n",
       "      <td>lib/ansible/plugins/connection/ssh.py</td>\n",
       "      <td>0</td>\n",
       "      <td>programming</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>364</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>U2</td>\n",
       "      <td>True</td>\n",
       "      <td>cve</td>\n",
       "      <td>CVE-2018-16876</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>407717</th>\n",
       "      <td>cve_CVE-2018-16876</td>\n",
       "      <td>auto_C_5_8</td>\n",
       "      <td>lib/ansible/plugins/connection/ssh.py</td>\n",
       "      <td>0</td>\n",
       "      <td>programming</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>365</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>U2</td>\n",
       "      <td>True</td>\n",
       "      <td>cve</td>\n",
       "      <td>CVE-2018-16876</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>407718 rows × 12 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                        id      bundle                                   file  \\\n",
       "0       cve_CVE-2020-10289      B_6_13     actionlib_tools/scripts/library.py   \n",
       "1       cve_CVE-2020-10289      B_6_13     actionlib_tools/scripts/library.py   \n",
       "2       cve_CVE-2020-10289      B_6_13     actionlib_tools/scripts/library.py   \n",
       "3       cve_CVE-2020-10289      B_6_13     actionlib_tools/scripts/library.py   \n",
       "4       cve_CVE-2020-10289       C_4_9     actionlib_tools/scripts/library.py   \n",
       "...                    ...         ...                                    ...   \n",
       "407713  cve_CVE-2018-16876  auto_C_5_8  lib/ansible/plugins/connection/ssh.py   \n",
       "407714  cve_CVE-2018-16876  auto_C_5_8  lib/ansible/plugins/connection/ssh.py   \n",
       "407715  cve_CVE-2018-16876  auto_C_5_8  lib/ansible/plugins/connection/ssh.py   \n",
       "407716  cve_CVE-2018-16876  auto_C_5_8  lib/ansible/plugins/connection/ssh.py   \n",
       "407717  cve_CVE-2018-16876  auto_C_5_8  lib/ansible/plugins/connection/ssh.py   \n",
       "\n",
       "        fileid         fcat         image  line annotation user   auto   ds  \\\n",
       "0            0  programming   afterChange   103   bug(fix)   U1  False  cve   \n",
       "1            0  programming   afterChange   137   bug(fix)   U1  False  cve   \n",
       "2            0  programming  beforeChange   103   bug(fix)   U1  False  cve   \n",
       "3            0  programming  beforeChange   137   bug(fix)   U1  False  cve   \n",
       "4            0  programming   afterChange   103   bug(fix)   U2  False  cve   \n",
       "...        ...          ...           ...   ...        ...  ...    ...  ...   \n",
       "407713       0  programming   afterChange   361   bug(fix)   U2   True  cve   \n",
       "407714       0  programming   afterChange   362   bug(fix)   U2   True  cve   \n",
       "407715       0  programming   afterChange   363   bug(fix)   U2   True  cve   \n",
       "407716       0  programming   afterChange   364   bug(fix)   U2   True  cve   \n",
       "407717       0  programming   afterChange   365   bug(fix)   U2   True  cve   \n",
       "\n",
       "                   bug  \n",
       "0       CVE-2020-10289  \n",
       "1       CVE-2020-10289  \n",
       "2       CVE-2020-10289  \n",
       "3       CVE-2020-10289  \n",
       "4       CVE-2020-10289  \n",
       "...                ...  \n",
       "407713  CVE-2018-16876  \n",
       "407714  CVE-2018-16876  \n",
       "407715  CVE-2018-16876  \n",
       "407716  CVE-2018-16876  \n",
       "407717  CVE-2018-16876  \n",
       "\n",
       "[407718 rows x 12 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dbb71219-93b6-47f0-8c03-252b1da8f4c0",
   "metadata": {},
   "source": [
    "# Analysis friendly ds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "54a758f3-c2c8-473f-923d-0dd954a6c7ab",
   "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>id</th>\n",
       "      <th>bundle</th>\n",
       "      <th>file</th>\n",
       "      <th>fileid</th>\n",
       "      <th>fcat</th>\n",
       "      <th>image</th>\n",
       "      <th>line</th>\n",
       "      <th>annotation</th>\n",
       "      <th>user</th>\n",
       "      <th>auto</th>\n",
       "      <th>ds</th>\n",
       "      <th>bug</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>B_6_13</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>programming</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>103</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>U1</td>\n",
       "      <td>False</td>\n",
       "      <td>cve</td>\n",
       "      <td>CVE-2020-10289</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>B_6_13</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>programming</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>137</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>U1</td>\n",
       "      <td>False</td>\n",
       "      <td>cve</td>\n",
       "      <td>CVE-2020-10289</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>B_6_13</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>programming</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>103</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>U1</td>\n",
       "      <td>False</td>\n",
       "      <td>cve</td>\n",
       "      <td>CVE-2020-10289</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>B_6_13</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>programming</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>137</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>U1</td>\n",
       "      <td>False</td>\n",
       "      <td>cve</td>\n",
       "      <td>CVE-2020-10289</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>C_4_9</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>programming</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>103</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>U2</td>\n",
       "      <td>False</td>\n",
       "      <td>cve</td>\n",
       "      <td>CVE-2020-10289</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>407713</th>\n",
       "      <td>cve_CVE-2018-16876</td>\n",
       "      <td>auto_C_5_8</td>\n",
       "      <td>lib/ansible/plugins/connection/ssh.py</td>\n",
       "      <td>0</td>\n",
       "      <td>programming</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>361</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>U2</td>\n",
       "      <td>True</td>\n",
       "      <td>cve</td>\n",
       "      <td>CVE-2018-16876</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>407714</th>\n",
       "      <td>cve_CVE-2018-16876</td>\n",
       "      <td>auto_C_5_8</td>\n",
       "      <td>lib/ansible/plugins/connection/ssh.py</td>\n",
       "      <td>0</td>\n",
       "      <td>programming</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>362</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>U2</td>\n",
       "      <td>True</td>\n",
       "      <td>cve</td>\n",
       "      <td>CVE-2018-16876</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>407715</th>\n",
       "      <td>cve_CVE-2018-16876</td>\n",
       "      <td>auto_C_5_8</td>\n",
       "      <td>lib/ansible/plugins/connection/ssh.py</td>\n",
       "      <td>0</td>\n",
       "      <td>programming</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>363</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>U2</td>\n",
       "      <td>True</td>\n",
       "      <td>cve</td>\n",
       "      <td>CVE-2018-16876</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>407716</th>\n",
       "      <td>cve_CVE-2018-16876</td>\n",
       "      <td>auto_C_5_8</td>\n",
       "      <td>lib/ansible/plugins/connection/ssh.py</td>\n",
       "      <td>0</td>\n",
       "      <td>programming</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>364</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>U2</td>\n",
       "      <td>True</td>\n",
       "      <td>cve</td>\n",
       "      <td>CVE-2018-16876</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>407717</th>\n",
       "      <td>cve_CVE-2018-16876</td>\n",
       "      <td>auto_C_5_8</td>\n",
       "      <td>lib/ansible/plugins/connection/ssh.py</td>\n",
       "      <td>0</td>\n",
       "      <td>programming</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>365</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>U2</td>\n",
       "      <td>True</td>\n",
       "      <td>cve</td>\n",
       "      <td>CVE-2018-16876</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>407718 rows × 12 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                        id      bundle                                   file  \\\n",
       "0       cve_CVE-2020-10289      B_6_13     actionlib_tools/scripts/library.py   \n",
       "1       cve_CVE-2020-10289      B_6_13     actionlib_tools/scripts/library.py   \n",
       "2       cve_CVE-2020-10289      B_6_13     actionlib_tools/scripts/library.py   \n",
       "3       cve_CVE-2020-10289      B_6_13     actionlib_tools/scripts/library.py   \n",
       "4       cve_CVE-2020-10289       C_4_9     actionlib_tools/scripts/library.py   \n",
       "...                    ...         ...                                    ...   \n",
       "407713  cve_CVE-2018-16876  auto_C_5_8  lib/ansible/plugins/connection/ssh.py   \n",
       "407714  cve_CVE-2018-16876  auto_C_5_8  lib/ansible/plugins/connection/ssh.py   \n",
       "407715  cve_CVE-2018-16876  auto_C_5_8  lib/ansible/plugins/connection/ssh.py   \n",
       "407716  cve_CVE-2018-16876  auto_C_5_8  lib/ansible/plugins/connection/ssh.py   \n",
       "407717  cve_CVE-2018-16876  auto_C_5_8  lib/ansible/plugins/connection/ssh.py   \n",
       "\n",
       "        fileid         fcat         image  line annotation user   auto   ds  \\\n",
       "0            0  programming   afterChange   103   bug(fix)   U1  False  cve   \n",
       "1            0  programming   afterChange   137   bug(fix)   U1  False  cve   \n",
       "2            0  programming  beforeChange   103   bug(fix)   U1  False  cve   \n",
       "3            0  programming  beforeChange   137   bug(fix)   U1  False  cve   \n",
       "4            0  programming   afterChange   103   bug(fix)   U2  False  cve   \n",
       "...        ...          ...           ...   ...        ...  ...    ...  ...   \n",
       "407713       0  programming   afterChange   361   bug(fix)   U2   True  cve   \n",
       "407714       0  programming   afterChange   362   bug(fix)   U2   True  cve   \n",
       "407715       0  programming   afterChange   363   bug(fix)   U2   True  cve   \n",
       "407716       0  programming   afterChange   364   bug(fix)   U2   True  cve   \n",
       "407717       0  programming   afterChange   365   bug(fix)   U2   True  cve   \n",
       "\n",
       "                   bug  \n",
       "0       CVE-2020-10289  \n",
       "1       CVE-2020-10289  \n",
       "2       CVE-2020-10289  \n",
       "3       CVE-2020-10289  \n",
       "4       CVE-2020-10289  \n",
       "...                ...  \n",
       "407713  CVE-2018-16876  \n",
       "407714  CVE-2018-16876  \n",
       "407715  CVE-2018-16876  \n",
       "407716  CVE-2018-16876  \n",
       "407717  CVE-2018-16876  \n",
       "\n",
       "[407718 rows x 12 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "array(['U1', 'U2', 'U3', 'E2', 'E3', 'E1'], dtype=object)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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>ds</th>\n",
       "      <th>fcat</th>\n",
       "      <th>id</th>\n",
       "      <th>file</th>\n",
       "      <th>fileid</th>\n",
       "      <th>bundle</th>\n",
       "      <th>image</th>\n",
       "      <th>line</th>\n",
       "      <th>annotation</th>\n",
       "      <th>user</th>\n",
       "      <th>auto</th>\n",
       "      <th>bug</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>bugs-in-py</td>\n",
       "      <td>data</td>\n",
       "      <td>11</td>\n",
       "      <td>11</td>\n",
       "      <td>11</td>\n",
       "      <td>11</td>\n",
       "      <td>11</td>\n",
       "      <td>11</td>\n",
       "      <td>11</td>\n",
       "      <td>11</td>\n",
       "      <td>11</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>bugs-in-py</td>\n",
       "      <td>documentation</td>\n",
       "      <td>180</td>\n",
       "      <td>180</td>\n",
       "      <td>180</td>\n",
       "      <td>180</td>\n",
       "      <td>180</td>\n",
       "      <td>180</td>\n",
       "      <td>180</td>\n",
       "      <td>180</td>\n",
       "      <td>180</td>\n",
       "      <td>180</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>bugs-in-py</td>\n",
       "      <td>other</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>bugs-in-py</td>\n",
       "      <td>programming</td>\n",
       "      <td>607</td>\n",
       "      <td>607</td>\n",
       "      <td>607</td>\n",
       "      <td>607</td>\n",
       "      <td>607</td>\n",
       "      <td>607</td>\n",
       "      <td>607</td>\n",
       "      <td>607</td>\n",
       "      <td>607</td>\n",
       "      <td>607</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>bugs-in-py</td>\n",
       "      <td>project</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>bugs-in-py</td>\n",
       "      <td>test</td>\n",
       "      <td>600</td>\n",
       "      <td>600</td>\n",
       "      <td>600</td>\n",
       "      <td>600</td>\n",
       "      <td>600</td>\n",
       "      <td>600</td>\n",
       "      <td>600</td>\n",
       "      <td>600</td>\n",
       "      <td>600</td>\n",
       "      <td>600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>crawl</td>\n",
       "      <td>data</td>\n",
       "      <td>55</td>\n",
       "      <td>55</td>\n",
       "      <td>55</td>\n",
       "      <td>55</td>\n",
       "      <td>55</td>\n",
       "      <td>55</td>\n",
       "      <td>55</td>\n",
       "      <td>55</td>\n",
       "      <td>55</td>\n",
       "      <td>55</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>crawl</td>\n",
       "      <td>documentation</td>\n",
       "      <td>542</td>\n",
       "      <td>542</td>\n",
       "      <td>542</td>\n",
       "      <td>542</td>\n",
       "      <td>542</td>\n",
       "      <td>542</td>\n",
       "      <td>542</td>\n",
       "      <td>542</td>\n",
       "      <td>542</td>\n",
       "      <td>542</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>crawl</td>\n",
       "      <td>markup</td>\n",
       "      <td>13</td>\n",
       "      <td>13</td>\n",
       "      <td>13</td>\n",
       "      <td>13</td>\n",
       "      <td>13</td>\n",
       "      <td>13</td>\n",
       "      <td>13</td>\n",
       "      <td>13</td>\n",
       "      <td>13</td>\n",
       "      <td>13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>crawl</td>\n",
       "      <td>other</td>\n",
       "      <td>40</td>\n",
       "      <td>40</td>\n",
       "      <td>40</td>\n",
       "      <td>40</td>\n",
       "      <td>40</td>\n",
       "      <td>40</td>\n",
       "      <td>40</td>\n",
       "      <td>40</td>\n",
       "      <td>40</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>crawl</td>\n",
       "      <td>programming</td>\n",
       "      <td>446</td>\n",
       "      <td>446</td>\n",
       "      <td>446</td>\n",
       "      <td>446</td>\n",
       "      <td>446</td>\n",
       "      <td>446</td>\n",
       "      <td>446</td>\n",
       "      <td>446</td>\n",
       "      <td>446</td>\n",
       "      <td>446</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>crawl</td>\n",
       "      <td>project</td>\n",
       "      <td>113</td>\n",
       "      <td>113</td>\n",
       "      <td>113</td>\n",
       "      <td>113</td>\n",
       "      <td>113</td>\n",
       "      <td>113</td>\n",
       "      <td>113</td>\n",
       "      <td>113</td>\n",
       "      <td>113</td>\n",
       "      <td>113</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>crawl</td>\n",
       "      <td>test</td>\n",
       "      <td>394</td>\n",
       "      <td>394</td>\n",
       "      <td>394</td>\n",
       "      <td>394</td>\n",
       "      <td>394</td>\n",
       "      <td>394</td>\n",
       "      <td>394</td>\n",
       "      <td>394</td>\n",
       "      <td>394</td>\n",
       "      <td>394</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>cve</td>\n",
       "      <td>data</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>cve</td>\n",
       "      <td>documentation</td>\n",
       "      <td>96</td>\n",
       "      <td>96</td>\n",
       "      <td>96</td>\n",
       "      <td>96</td>\n",
       "      <td>96</td>\n",
       "      <td>96</td>\n",
       "      <td>96</td>\n",
       "      <td>96</td>\n",
       "      <td>96</td>\n",
       "      <td>96</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>cve</td>\n",
       "      <td>markup</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>cve</td>\n",
       "      <td>other</td>\n",
       "      <td>25</td>\n",
       "      <td>25</td>\n",
       "      <td>25</td>\n",
       "      <td>25</td>\n",
       "      <td>25</td>\n",
       "      <td>25</td>\n",
       "      <td>25</td>\n",
       "      <td>25</td>\n",
       "      <td>25</td>\n",
       "      <td>25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>cve</td>\n",
       "      <td>programming</td>\n",
       "      <td>352</td>\n",
       "      <td>352</td>\n",
       "      <td>352</td>\n",
       "      <td>352</td>\n",
       "      <td>352</td>\n",
       "      <td>352</td>\n",
       "      <td>352</td>\n",
       "      <td>352</td>\n",
       "      <td>352</td>\n",
       "      <td>352</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>cve</td>\n",
       "      <td>project</td>\n",
       "      <td>40</td>\n",
       "      <td>40</td>\n",
       "      <td>40</td>\n",
       "      <td>40</td>\n",
       "      <td>40</td>\n",
       "      <td>40</td>\n",
       "      <td>40</td>\n",
       "      <td>40</td>\n",
       "      <td>40</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>cve</td>\n",
       "      <td>test</td>\n",
       "      <td>199</td>\n",
       "      <td>199</td>\n",
       "      <td>199</td>\n",
       "      <td>199</td>\n",
       "      <td>199</td>\n",
       "      <td>199</td>\n",
       "      <td>199</td>\n",
       "      <td>199</td>\n",
       "      <td>199</td>\n",
       "      <td>199</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            ds           fcat   id  file  fileid  bundle  image  line  \\\n",
       "0   bugs-in-py           data   11    11      11      11     11    11   \n",
       "1   bugs-in-py  documentation  180   180     180     180    180   180   \n",
       "2   bugs-in-py          other    4     4       4       4      4     4   \n",
       "3   bugs-in-py    programming  607   607     607     607    607   607   \n",
       "4   bugs-in-py        project    5     5       5       5      5     5   \n",
       "5   bugs-in-py           test  600   600     600     600    600   600   \n",
       "6        crawl           data   55    55      55      55     55    55   \n",
       "7        crawl  documentation  542   542     542     542    542   542   \n",
       "8        crawl         markup   13    13      13      13     13    13   \n",
       "9        crawl          other   40    40      40      40     40    40   \n",
       "10       crawl    programming  446   446     446     446    446   446   \n",
       "11       crawl        project  113   113     113     113    113   113   \n",
       "12       crawl           test  394   394     394     394    394   394   \n",
       "13         cve           data    7     7       7       7      7     7   \n",
       "14         cve  documentation   96    96      96      96     96    96   \n",
       "15         cve         markup    8     8       8       8      8     8   \n",
       "16         cve          other   25    25      25      25     25    25   \n",
       "17         cve    programming  352   352     352     352    352   352   \n",
       "18         cve        project   40    40      40      40     40    40   \n",
       "19         cve           test  199   199     199     199    199   199   \n",
       "\n",
       "    annotation  user  auto  bug  \n",
       "0           11    11    11   11  \n",
       "1          180   180   180  180  \n",
       "2            4     4     4    4  \n",
       "3          607   607   607  607  \n",
       "4            5     5     5    5  \n",
       "5          600   600   600  600  \n",
       "6           55    55    55   55  \n",
       "7          542   542   542  542  \n",
       "8           13    13    13   13  \n",
       "9           40    40    40   40  \n",
       "10         446   446   446  446  \n",
       "11         113   113   113  113  \n",
       "12         394   394   394  394  \n",
       "13           7     7     7    7  \n",
       "14          96    96    96   96  \n",
       "15           8     8     8    8  \n",
       "16          25    25    25   25  \n",
       "17         352   352   352  352  \n",
       "18          40    40    40   40  \n",
       "19         199   199   199  199  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "users = df['user'].unique()\n",
    "\n",
    "display(df)\n",
    "\n",
    "files = df\\\n",
    "    .groupby(['id', 'ds', 'file', 'fileid', 'fcat']).count().reset_index()\\\n",
    "    .groupby(['ds', 'fcat']).count().reset_index()\n",
    "\n",
    "display(users, files)\n",
    "\n",
    "\n",
    "#files.to_dict()\n",
    "df.to_csv(\"collective.csv\")\n",
    "df.to_json(\"collective.json\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0a490b06-37df-4d21-9969-ec8c2afb0139",
   "metadata": {},
   "source": [
    "# Analysis"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e67e9e04-a614-41dc-b976-2c902cc2f24c",
   "metadata": {},
   "source": [
    "## Flow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b18ae350-9c4a-4c08-9948-b657bd9cd0a1",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "f18c5e94-ed45-4327-8909-16e917fce9a9",
   "metadata": {},
   "outputs": [],
   "source": [
    "ret_df = [] \n",
    "per_line_match = []\n",
    "\n",
    "ret_flow = []\n",
    "\n",
    "users = df['user'].unique()\n",
    "\n",
    "def get_stats(z, u1, u2):\n",
    "    \n",
    "    z['match'] = (z['annotation_x'] == z['annotation_y'])\n",
    "\n",
    "    out = z.groupby(['ds', 'id', 'file', 'fileid']).agg(All=('match', 'count'), Match=('match', 'sum'))\n",
    "    \n",
    "    flow = z.reset_index().groupby(\n",
    "        ['ds', 'id','annotation_y', 'annotation_x', 'user_x','file', 'fileid',  'line', 'image']\n",
    "    ).agg(Count=('match', 'count'))\n",
    "    \n",
    "    out['u1'] = u1\n",
    "    out['u2'] = u2\n",
    "    \n",
    "    return out, flow\n",
    "\n",
    "for u in users:\n",
    "    \n",
    "    cond1 = (df['user'] == u) & ~(df['auto'])\n",
    "    cond2 = (df['user'] == u) & (df['auto'])\n",
    "    \n",
    "    z = df[cond1].merge(df[cond2], how='inner', on=['id','file', 'fileid', 'image','line','ds'])\n",
    "    \n",
    "    out, flow_auto = get_stats(z, u, 'AUTO')\n",
    "    \n",
    "    ret_flow.append(flow_auto)\n",
    "    ret_df.append(out)\n",
    "    \n",
    "\n",
    "    \n",
    "for u1, u2 in combinations(users,2):\n",
    "    \n",
    "    cond1 = (df['user'] == u1) & ~(df['auto'])\n",
    "    cond2 = (df['user'] == u2) & ~(df['auto'])\n",
    "\n",
    "    z = df[cond1].merge(df[cond2], how='inner', on=['id','file','fileid', 'image','line','ds'])\n",
    "\n",
    "    out, flow = get_stats(z, u1, u2)\n",
    "    \n",
    "    ret_df.append(out)\n",
    "\n",
    "ret_df = pd.concat(ret_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "a35b1018-dc6a-4c2d-b5bf-909306d95731",
   "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></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>Count</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ds</th>\n",
       "      <th>id</th>\n",
       "      <th>annotation_y</th>\n",
       "      <th>annotation_x</th>\n",
       "      <th>user_x</th>\n",
       "      <th>file</th>\n",
       "      <th>fileid</th>\n",
       "      <th>line</th>\n",
       "      <th>image</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">bugs-in-py</th>\n",
       "      <th rowspan=\"5\" valign=\"top\">bugs-in-py_PySnooper-1</th>\n",
       "      <th rowspan=\"5\" valign=\"top\">bug(fix)</th>\n",
       "      <th rowspan=\"5\" valign=\"top\">bug(fix)</th>\n",
       "      <th rowspan=\"5\" valign=\"top\">E1</th>\n",
       "      <th>pysnooper/pycompat.py</th>\n",
       "      <th>0</th>\n",
       "      <th>11</th>\n",
       "      <th>afterChange</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"4\" valign=\"top\">pysnooper/tracer.py</th>\n",
       "      <th rowspan=\"4\" valign=\"top\">0</th>\n",
       "      <th>17</th>\n",
       "      <th>afterChange</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <th>afterChange</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87</th>\n",
       "      <th>beforeChange</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>89</th>\n",
       "      <th>afterChange</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <th>...</th>\n",
       "      <th>...</th>\n",
       "      <th>...</th>\n",
       "      <th>...</th>\n",
       "      <th>...</th>\n",
       "      <th>...</th>\n",
       "      <th>...</th>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">cve</th>\n",
       "      <th rowspan=\"5\" valign=\"top\">cve_CVE-2022-31015</th>\n",
       "      <th>bug(fix)</th>\n",
       "      <th>bug(fix) + refactoring</th>\n",
       "      <th>E1</th>\n",
       "      <th>src/waitress/channel.py</th>\n",
       "      <th>0</th>\n",
       "      <th>379</th>\n",
       "      <th>afterChange</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"4\" valign=\"top\">test</th>\n",
       "      <th rowspan=\"4\" valign=\"top\">test</th>\n",
       "      <th rowspan=\"4\" valign=\"top\">E1</th>\n",
       "      <th rowspan=\"4\" valign=\"top\">tests/test_channel.py</th>\n",
       "      <th rowspan=\"4\" valign=\"top\">0</th>\n",
       "      <th rowspan=\"2\" valign=\"top\">379</th>\n",
       "      <th>afterChange</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>beforeChange</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">456</th>\n",
       "      <th>afterChange</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>beforeChange</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>16047 rows × 1 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                                                                                               Count\n",
       "ds         id                     annotation_y annotation_x           user_x file                    fileid line image              \n",
       "bugs-in-py bugs-in-py_PySnooper-1 bug(fix)     bug(fix)               E1     pysnooper/pycompat.py   0      11   afterChange       1\n",
       "                                                                             pysnooper/tracer.py     0      17   afterChange       1\n",
       "                                                                                                            18   afterChange       1\n",
       "                                                                                                            87   beforeChange      1\n",
       "                                                                                                            89   afterChange       1\n",
       "...                                                                                                                              ...\n",
       "cve        cve_CVE-2022-31015     bug(fix)     bug(fix) + refactoring E1     src/waitress/channel.py 0      379  afterChange       1\n",
       "                                  test         test                   E1     tests/test_channel.py   0      379  afterChange       1\n",
       "                                                                                                                 beforeChange      1\n",
       "                                                                                                            456  afterChange       1\n",
       "                                                                                                                 beforeChange      1\n",
       "\n",
       "[16047 rows x 1 columns]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "flow_auto"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "8661a045-cfcf-4824-946b-efa80474b34a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "auto,user,S\n",
      "A: bug(fix),U: bug(fix),26.403257057366396\n",
      "A: bug(fix),U: bug(fix) + refactoring,2.2848453633531998\n",
      "A: bug(fix),U: documentation,3.2967895420989377\n",
      "A: bug(fix),U: other,0.8265273587913571\n",
      "A: bug(fix),U: refactoring,3.303166311039168\n",
      "A: bug(fix),U: test,0.06082456527604052\n",
      "A: documentation,U: bug(fix),0.2972555367522625\n",
      "A: documentation,U: bug(fix) + refactoring,0.006867289627940059\n",
      "A: documentation,U: documentation,14.887793392686335\n",
      "A: documentation,U: other,0.11919652711353101\n",
      "A: documentation,U: refactoring,0.026488117136340226\n",
      "A: documentation,U: test,0.009319893066490078\n",
      "A: documentation,U: test + refactoring,0.004414686189390038\n",
      "A: test,U: bug(fix),0.0941799720403208\n",
      "A: test,U: bug(fix) + refactoring,0.00833885169107007\n",
      "A: test,U: documentation,1.4499791528707724\n",
      "A: test,U: other,0.03924165501680033\n",
      "A: test,U: refactoring,0.8363377725455571\n",
      "A: test,U: test,44.620214357540526\n",
      "A: test,U: test + refactoring,1.4249625977975622\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_232579/783009.py:2: FutureWarning: The default value of numeric_only in DataFrameGroupBy.sum 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",
      "  tmp = flow_auto.reset_index().groupby(['annotation_y', 'annotation_x']).sum().reset_index()\n"
     ]
    }
   ],
   "source": [
    "flow_auto = pd.concat(ret_flow)\n",
    "tmp = flow_auto.reset_index().groupby(['annotation_y', 'annotation_x']).sum().reset_index()\n",
    "S = tmp['Count'].sum()\n",
    "tmp['S'] = (tmp['Count'] / S) * 100\n",
    "#tmp\n",
    "tmp['auto'] = \"A: \" + tmp['annotation_y']\n",
    "tmp['user'] = \"U: \" + tmp['annotation_x']\n",
    "\n",
    "print(tmp[['auto', 'user', 'S']].to_csv(index=False))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9ad6d053-f26d-4d70-aa23-80404c3a8616",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "362f52e9-90d6-4f77-a9a1-de1623b9ee50",
   "metadata": {},
   "source": [
    "# Tangling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "dad49a5d-157b-425b-a9c3-15cc898a2613",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_232579/27409827.py:108: UserWarning: Tight layout not applied. tight_layout cannot make axes width small enough to accommodate all axes decorations\n",
      "  fig.tight_layout()\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1000x200 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1000x200 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "\n",
    "def export_legend(legend, filename=\"legend.svg\"):\n",
    "    fig  = legend.figure\n",
    "    fig.canvas.draw()\n",
    "    bbox  = legend.get_window_extent().transformed(fig.dpi_scale_trans.inverted())\n",
    "    fig.savefig(filename, dpi=\"figure\", bbox_inches=bbox)\n",
    "\n",
    "\n",
    "\n",
    "def plot_percentage(df, cols, cols2, legend=True, filename=None):\n",
    "    \n",
    "    fig, axes = plt.subplots(figsize=(10,2), ncols=2, sharey=True)\n",
    "\n",
    "\n",
    "    # plot in groups\n",
    "    x_var, y_var = cols\n",
    "    df_grouped = df.groupby(x_var)[y_var].value_counts(normalize=True).unstack(y_var)\n",
    "    cmap = sns.color_palette(\"Set2\", as_cmap=True)\n",
    "    ax = df_grouped.plot.barh(stacked=True, width=0.9, ax=axes[1], colormap=cmap)\n",
    "    \n",
    "    \n",
    "    # Plot total count\n",
    "    x_var2, y_var2 = cols2\n",
    "    df_grouped2 = df.groupby(x_var).count()[y_var]\n",
    "    \n",
    "    df_grouped2 /= df_grouped2.sum()\n",
    "    cmap = sns.color_palette(\"Set1\", as_cmap=True)\n",
    "\n",
    "    df_grouped2.plot.barh(stacked=False, width=0.9, ax=axes[0], colormap=cmap)\n",
    "\n",
    "    if legend and filename:\n",
    "        legend = plt.legend(\n",
    "            bbox_to_anchor=(0.5, 1.20),\n",
    "            loc=\"lower center\",\n",
    "            borderaxespad=0,\n",
    "            frameon=False,\n",
    "            ncol=8,\n",
    "        )\n",
    "        export_legend(legend, \"legend_\" + filename )\n",
    "    else:\n",
    "        ax.get_legend().remove()\n",
    "        \n",
    "    df_grouped.fillna(0)\n",
    "    for ix, row in df_grouped.reset_index(drop=True).iterrows():\n",
    "        #print(ix, row, )\n",
    "        cumulative = 0\n",
    "        for element in row:\n",
    "            if element == element and element >= 0.05:\n",
    "                #print(\">>\", element, ix)\n",
    "                plt.text(\n",
    "                    cumulative + element/3,\n",
    "                    ix,\n",
    "                    f\"{int(element * 100)} %\",\n",
    "                    va=\"center\",\n",
    "                    ha=\"left\",\n",
    "                    fontsize=8\n",
    "\n",
    "                )\n",
    "                cumulative += element\n",
    "    \n",
    "\n",
    "    #for ax in axes.flat:\n",
    "    #    ax.margins(0.1)\n",
    "        \n",
    "    #plt.autoscale(enable=True, axis='both', tight=True) # tight layout\n",
    "    #fig.subplots_adjust(wspace=0.7)\n",
    "\n",
    "    axes[0].xaxis.set_major_formatter(lambda x,y: str(int(x*100))+'%')\n",
    "    axes[1].xaxis.set_major_formatter(lambda x,y: str(int(x*100))+'%')\n",
    "\n",
    "    #ax.tick_params(axis='x', which='major', labelsize=8)\n",
    "    \n",
    "    # If you have positive numbers and want to invert the x-axis of the left plot\n",
    "    axes[0].invert_xaxis() \n",
    "\n",
    "    \n",
    "    #plt.subplots_adjust(wspace=0, top=0.85, bottom=0.1, left=0.18, right=0.95)\n",
    "\n",
    "    #plt.setp(axes[0].yaxis.get_majorticklabels(), ha='center')\n",
    "    \n",
    "    #axes[0].set(yticks=y, yticklabels=job_r)\n",
    "\n",
    "    \n",
    "    axes[0].yaxis.tick_right()\n",
    "    plt.setp(axes[0].yaxis.get_majorticklabels(), ha='center')\n",
    "    axes[0].tick_params(axis=u'both', which=u'both',length=0)\n",
    "    \n",
    "    # Create offset transform by some points in x direction\n",
    "    dx = 35/72.\n",
    "    dy = 0 / 72.\n",
    "    offset = matplotlib.transforms.ScaledTranslation(dx, dy, fig.dpi_scale_trans)\n",
    "    # apply offset transform to all y ticklabels.\n",
    "    for label in axes[0].yaxis.get_majorticklabels():\n",
    "        label.set_transform(label.get_transform() + offset)\n",
    "\n",
    "    #fig[0][\"right\"].major_ticklabels.set_ha(\"center\")\n",
    "\n",
    "    fig.subplots_adjust(wspace=0.5)\n",
    "\n",
    "    sns.despine(left=True, right=True)\n",
    "    axes[0].set_title(\"% of lines\")\n",
    "    axes[1].set_title(\"% of annotations\")\n",
    "    axes[0].set(ylabel=None)\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "    plt.tight_layout()\n",
    "    fig.tight_layout()\n",
    "    if filename:\n",
    "        fig.savefig(filename)\n",
    "\n",
    "    \n",
    "\n",
    "plot_percentage(df[~df['auto']], ['fcat', 'annotation'], ['fcat', 'annotation'], legend = True, filename=\"fcat_annotations.svg\")\n",
    "plot_percentage(df[~df['auto']], ['fcat', 'annotation'], ['fcat', 'annotation'], legend = False, filename=\"fcat_annotations.svg\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ce5c5e0b-61ea-488b-8562-253b31eb8c93",
   "metadata": {},
   "source": [
    "# Consensus"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "685c2f9e-b47c-4ca8-9632-289bfe8de459",
   "metadata": {},
   "outputs": [],
   "source": [
    "user_l = {}\n",
    "df_inter_rater = None\n",
    "users_ids = ['U1', 'U2', 'U3', 'E1', 'E2', 'E3']\n",
    "cols = ['id', 'file', 'fileid', 'image', 'line', 'annotation']\n",
    "colsj = cols[:-1]\n",
    "\n",
    "for user in users_ids:\n",
    "    filter = (df['user'] == user) & (df['auto'] == False)\n",
    "    if df_inter_rater is None:\n",
    "        df_inter_rater = df[filter][cols].copy()\n",
    "    else:\n",
    "        df_inter_rater = df_inter_rater.merge(df[filter][cols], left_on = colsj, right_on = colsj, suffixes=('', \"_\"+user), how='outer')\n",
    "\n",
    "    #df_inter_rater = df_inter_rater.drop_duplicates(subset=['id', 'file', 'image', 'line', 'annotation'])\n",
    "    \n",
    "    #tmp_g_debug = df_inter_rater.groupby(colsj).count().reset_index()\n",
    "    #display(tmp_g_debug[tmp_g_debug['annotation'] > 1])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "6b90a086-d459-4182-9aba-433e2cd05de4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(67955.0, 67951.0)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "filter = (df['auto'] == False)\n",
    "len(df[filter])/3, len(df[~filter])/3, "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "c4bb9527-10ca-4c49-b224-83213931ced2",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_inter_rater.columns = [ c if c != 'annotation' else 'annotation_U1' for c in df_inter_rater.columns]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "f74b71f3-7d9c-4bd8-9743-1ed1e0a5327d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['bug(fix)',\n",
       " 'bug(fix) + refactoring',\n",
       " 'documentation',\n",
       " 'refactoring',\n",
       " 'other',\n",
       " 'test + refactoring',\n",
       " 'test']"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "possible_line_annotations = df['annotation'].unique().tolist()\n",
    "possible_line_annotations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "688e74d8-33c0-4b78-9e20-02198be9dabd",
   "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>id</th>\n",
       "      <th>file</th>\n",
       "      <th>fileid</th>\n",
       "      <th>image</th>\n",
       "      <th>line</th>\n",
       "      <th>annotation_U1</th>\n",
       "      <th>annotation_U2</th>\n",
       "      <th>annotation_U3</th>\n",
       "      <th>annotation_E1</th>\n",
       "      <th>annotation_E2</th>\n",
       "      <th>annotation_E3</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>103</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>137</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>103</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>137</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>cve_CVE-2016-10516</td>\n",
       "      <td>werkzeug/debug/tbtools.py</td>\n",
       "      <td>0</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>353</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix) + refactoring</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67958</th>\n",
       "      <td>bugs-in-py_pandas-88</td>\n",
       "      <td>pandas/tests/reshape/test_pivot.py</td>\n",
       "      <td>0</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>900</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67959</th>\n",
       "      <td>bugs-in-py_pandas-88</td>\n",
       "      <td>pandas/tests/reshape/test_pivot.py</td>\n",
       "      <td>0</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>901</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67960</th>\n",
       "      <td>bugs-in-py_pandas-88</td>\n",
       "      <td>pandas/tests/reshape/test_pivot.py</td>\n",
       "      <td>0</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>902</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67961</th>\n",
       "      <td>bugs-in-py_pandas-88</td>\n",
       "      <td>pandas/tests/reshape/test_pivot.py</td>\n",
       "      <td>0</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>903</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67962</th>\n",
       "      <td>bugs-in-py_pandas-88</td>\n",
       "      <td>pandas/tests/reshape/test_pivot.py</td>\n",
       "      <td>0</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>904</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>67963 rows × 11 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                         id                                file  fileid  \\\n",
       "0        cve_CVE-2020-10289  actionlib_tools/scripts/library.py       0   \n",
       "1        cve_CVE-2020-10289  actionlib_tools/scripts/library.py       0   \n",
       "2        cve_CVE-2020-10289  actionlib_tools/scripts/library.py       0   \n",
       "3        cve_CVE-2020-10289  actionlib_tools/scripts/library.py       0   \n",
       "4        cve_CVE-2016-10516           werkzeug/debug/tbtools.py       0   \n",
       "...                     ...                                 ...     ...   \n",
       "67958  bugs-in-py_pandas-88  pandas/tests/reshape/test_pivot.py       0   \n",
       "67959  bugs-in-py_pandas-88  pandas/tests/reshape/test_pivot.py       0   \n",
       "67960  bugs-in-py_pandas-88  pandas/tests/reshape/test_pivot.py       0   \n",
       "67961  bugs-in-py_pandas-88  pandas/tests/reshape/test_pivot.py       0   \n",
       "67962  bugs-in-py_pandas-88  pandas/tests/reshape/test_pivot.py       0   \n",
       "\n",
       "              image  line annotation_U1           annotation_U2 annotation_U3  \\\n",
       "0       afterChange   103      bug(fix)                bug(fix)      bug(fix)   \n",
       "1       afterChange   137      bug(fix)                bug(fix)      bug(fix)   \n",
       "2      beforeChange   103      bug(fix)                bug(fix)      bug(fix)   \n",
       "3      beforeChange   137      bug(fix)                bug(fix)      bug(fix)   \n",
       "4       afterChange   353      bug(fix)  bug(fix) + refactoring           NaN   \n",
       "...             ...   ...           ...                     ...           ...   \n",
       "67958  beforeChange   900           NaN                    test          test   \n",
       "67959  beforeChange   901           NaN                    test          test   \n",
       "67960  beforeChange   902           NaN                    test          test   \n",
       "67961  beforeChange   903           NaN                    test          test   \n",
       "67962  beforeChange   904           NaN                    test          test   \n",
       "\n",
       "      annotation_E1 annotation_E2 annotation_E3  \n",
       "0               NaN           NaN           NaN  \n",
       "1               NaN           NaN           NaN  \n",
       "2               NaN           NaN           NaN  \n",
       "3               NaN           NaN           NaN  \n",
       "4               NaN      bug(fix)           NaN  \n",
       "...             ...           ...           ...  \n",
       "67958          test           NaN           NaN  \n",
       "67959          test           NaN           NaN  \n",
       "67960          test           NaN           NaN  \n",
       "67961          test           NaN           NaN  \n",
       "67962          test           NaN           NaN  \n",
       "\n",
       "[67963 rows x 11 columns]"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_inter_rater"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "acea65e3-dd3d-4bde-957d-86d9c25af000",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "48f67489-472f-42be-a3f0-c1377ae71b56",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f544a7f0-372b-445c-8862-5490aecb1a36",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "44f72cd2-6969-4386-830e-e2a3e59b6f57",
   "metadata": {},
   "outputs": [],
   "source": [
    "def consensus(row):\n",
    "    ret = {k:0 for k in possible_line_annotations}\n",
    "    val = None\n",
    "    for u in users_ids:\n",
    "        u_val = row['annotation_' + u]\n",
    "        if not pd.isnull(u_val):\n",
    "            ret[u_val] += 1\n",
    "            val = u_val\n",
    "            \n",
    "    if sum(ret.values()) < 3: # Check if all 3 reviewers annotated line\n",
    "        print(\"Error\")\n",
    "\n",
    "    return Counter(ret).most_common(1)[0][0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "1939dda7-6098-4e0e-bdca-fcf2015fd6d3",
   "metadata": {},
   "outputs": [],
   "source": [
    "tmp = pd.DataFrame(df_inter_rater.apply(consensus, axis=1).tolist())\n",
    "G = df_inter_rater.copy()\n",
    "G['consensus']= tmp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "1fbc76e1-b9ae-4dc0-8940-eb1c97dc5356",
   "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>id</th>\n",
       "      <th>file</th>\n",
       "      <th>fileid</th>\n",
       "      <th>image</th>\n",
       "      <th>line</th>\n",
       "      <th>annotation_U1</th>\n",
       "      <th>annotation_U2</th>\n",
       "      <th>annotation_U3</th>\n",
       "      <th>annotation_E1</th>\n",
       "      <th>annotation_E2</th>\n",
       "      <th>annotation_E3</th>\n",
       "      <th>consensus</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>103</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bug(fix)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>137</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bug(fix)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>103</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bug(fix)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>137</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bug(fix)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>cve_CVE-2016-10516</td>\n",
       "      <td>werkzeug/debug/tbtools.py</td>\n",
       "      <td>0</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>353</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix) + refactoring</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bug(fix)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67958</th>\n",
       "      <td>bugs-in-py_pandas-88</td>\n",
       "      <td>pandas/tests/reshape/test_pivot.py</td>\n",
       "      <td>0</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>900</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67959</th>\n",
       "      <td>bugs-in-py_pandas-88</td>\n",
       "      <td>pandas/tests/reshape/test_pivot.py</td>\n",
       "      <td>0</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>901</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67960</th>\n",
       "      <td>bugs-in-py_pandas-88</td>\n",
       "      <td>pandas/tests/reshape/test_pivot.py</td>\n",
       "      <td>0</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>902</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67961</th>\n",
       "      <td>bugs-in-py_pandas-88</td>\n",
       "      <td>pandas/tests/reshape/test_pivot.py</td>\n",
       "      <td>0</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>903</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67962</th>\n",
       "      <td>bugs-in-py_pandas-88</td>\n",
       "      <td>pandas/tests/reshape/test_pivot.py</td>\n",
       "      <td>0</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>904</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>67963 rows × 12 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                         id                                file  fileid  \\\n",
       "0        cve_CVE-2020-10289  actionlib_tools/scripts/library.py       0   \n",
       "1        cve_CVE-2020-10289  actionlib_tools/scripts/library.py       0   \n",
       "2        cve_CVE-2020-10289  actionlib_tools/scripts/library.py       0   \n",
       "3        cve_CVE-2020-10289  actionlib_tools/scripts/library.py       0   \n",
       "4        cve_CVE-2016-10516           werkzeug/debug/tbtools.py       0   \n",
       "...                     ...                                 ...     ...   \n",
       "67958  bugs-in-py_pandas-88  pandas/tests/reshape/test_pivot.py       0   \n",
       "67959  bugs-in-py_pandas-88  pandas/tests/reshape/test_pivot.py       0   \n",
       "67960  bugs-in-py_pandas-88  pandas/tests/reshape/test_pivot.py       0   \n",
       "67961  bugs-in-py_pandas-88  pandas/tests/reshape/test_pivot.py       0   \n",
       "67962  bugs-in-py_pandas-88  pandas/tests/reshape/test_pivot.py       0   \n",
       "\n",
       "              image  line annotation_U1           annotation_U2 annotation_U3  \\\n",
       "0       afterChange   103      bug(fix)                bug(fix)      bug(fix)   \n",
       "1       afterChange   137      bug(fix)                bug(fix)      bug(fix)   \n",
       "2      beforeChange   103      bug(fix)                bug(fix)      bug(fix)   \n",
       "3      beforeChange   137      bug(fix)                bug(fix)      bug(fix)   \n",
       "4       afterChange   353      bug(fix)  bug(fix) + refactoring           NaN   \n",
       "...             ...   ...           ...                     ...           ...   \n",
       "67958  beforeChange   900           NaN                    test          test   \n",
       "67959  beforeChange   901           NaN                    test          test   \n",
       "67960  beforeChange   902           NaN                    test          test   \n",
       "67961  beforeChange   903           NaN                    test          test   \n",
       "67962  beforeChange   904           NaN                    test          test   \n",
       "\n",
       "      annotation_E1 annotation_E2 annotation_E3 consensus  \n",
       "0               NaN           NaN           NaN  bug(fix)  \n",
       "1               NaN           NaN           NaN  bug(fix)  \n",
       "2               NaN           NaN           NaN  bug(fix)  \n",
       "3               NaN           NaN           NaN  bug(fix)  \n",
       "4               NaN      bug(fix)           NaN  bug(fix)  \n",
       "...             ...           ...           ...       ...  \n",
       "67958          test           NaN           NaN      test  \n",
       "67959          test           NaN           NaN      test  \n",
       "67960          test           NaN           NaN      test  \n",
       "67961          test           NaN           NaN      test  \n",
       "67962          test           NaN           NaN      test  \n",
       "\n",
       "[67963 rows x 12 columns]"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "G"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aaf2eaf1-3579-4018-bb39-828c23896e14",
   "metadata": {},
   "source": [
    "## Consensus vs Auto"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "993615ae-fcf6-471c-b694-cea33a3fbb6c",
   "metadata": {},
   "outputs": [],
   "source": [
    "Z = df[df['auto'] == True].groupby(cols).count().reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "128c583e-237f-49bf-b4c8-18da2a474714",
   "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>id</th>\n",
       "      <th>file</th>\n",
       "      <th>fileid</th>\n",
       "      <th>image</th>\n",
       "      <th>line</th>\n",
       "      <th>annotation</th>\n",
       "      <th>bundle</th>\n",
       "      <th>fcat</th>\n",
       "      <th>user</th>\n",
       "      <th>auto</th>\n",
       "      <th>ds</th>\n",
       "      <th>bug</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>bugs-in-py_PySnooper-1</td>\n",
       "      <td>pysnooper/pycompat.py</td>\n",
       "      <td>0</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>11</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>bugs-in-py_PySnooper-1</td>\n",
       "      <td>pysnooper/tracer.py</td>\n",
       "      <td>0</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>17</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>bugs-in-py_PySnooper-1</td>\n",
       "      <td>pysnooper/tracer.py</td>\n",
       "      <td>0</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>18</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>bugs-in-py_PySnooper-1</td>\n",
       "      <td>pysnooper/tracer.py</td>\n",
       "      <td>0</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>89</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>bugs-in-py_PySnooper-1</td>\n",
       "      <td>pysnooper/tracer.py</td>\n",
       "      <td>0</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>135</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67946</th>\n",
       "      <td>cve_CVE-2022-31124</td>\n",
       "      <td>tests/fuzzer/fuzz_valid_magic.py</td>\n",
       "      <td>1</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>27</td>\n",
       "      <td>test</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67947</th>\n",
       "      <td>cve_CVE-2022-31124</td>\n",
       "      <td>tests/fuzzer/requirements.txt</td>\n",
       "      <td>1</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>1</td>\n",
       "      <td>test</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67948</th>\n",
       "      <td>cve_CVE-2022-31124</td>\n",
       "      <td>tests/fuzzer/requirements.txt</td>\n",
       "      <td>1</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>2</td>\n",
       "      <td>test</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67949</th>\n",
       "      <td>cve_CVE-2022-31124</td>\n",
       "      <td>tests/fuzzer/requirements.txt</td>\n",
       "      <td>1</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>3</td>\n",
       "      <td>test</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67950</th>\n",
       "      <td>cve_CVE-2022-31124</td>\n",
       "      <td>tests/fuzzer/requirements.txt</td>\n",
       "      <td>1</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>4</td>\n",
       "      <td>test</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>67951 rows × 12 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                           id                              file  fileid  \\\n",
       "0      bugs-in-py_PySnooper-1             pysnooper/pycompat.py       0   \n",
       "1      bugs-in-py_PySnooper-1               pysnooper/tracer.py       0   \n",
       "2      bugs-in-py_PySnooper-1               pysnooper/tracer.py       0   \n",
       "3      bugs-in-py_PySnooper-1               pysnooper/tracer.py       0   \n",
       "4      bugs-in-py_PySnooper-1               pysnooper/tracer.py       0   \n",
       "...                       ...                               ...     ...   \n",
       "67946      cve_CVE-2022-31124  tests/fuzzer/fuzz_valid_magic.py       1   \n",
       "67947      cve_CVE-2022-31124     tests/fuzzer/requirements.txt       1   \n",
       "67948      cve_CVE-2022-31124     tests/fuzzer/requirements.txt       1   \n",
       "67949      cve_CVE-2022-31124     tests/fuzzer/requirements.txt       1   \n",
       "67950      cve_CVE-2022-31124     tests/fuzzer/requirements.txt       1   \n",
       "\n",
       "             image  line annotation  bundle  fcat  user  auto  ds  bug  \n",
       "0      afterChange    11   bug(fix)       3     3     3     3   3    3  \n",
       "1      afterChange    17   bug(fix)       3     3     3     3   3    3  \n",
       "2      afterChange    18   bug(fix)       3     3     3     3   3    3  \n",
       "3      afterChange    89   bug(fix)       3     3     3     3   3    3  \n",
       "4      afterChange   135   bug(fix)       3     3     3     3   3    3  \n",
       "...            ...   ...        ...     ...   ...   ...   ...  ..  ...  \n",
       "67946  afterChange    27       test       3     3     3     3   3    3  \n",
       "67947  afterChange     1       test       3     3     3     3   3    3  \n",
       "67948  afterChange     2       test       3     3     3     3   3    3  \n",
       "67949  afterChange     3       test       3     3     3     3   3    3  \n",
       "67950  afterChange     4       test       3     3     3     3   3    3  \n",
       "\n",
       "[67951 rows x 12 columns]"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Z"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "671422d7-53b3-4640-bc6c-c61112ccebc8",
   "metadata": {},
   "outputs": [],
   "source": [
    "R = Z.merge(G, left_on=colsj, right_on=colsj)\n",
    "#R = R[['consensus', 'annotation']]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "8a1e362e-47d6-496e-991d-21bf4c226077",
   "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>id</th>\n",
       "      <th>file</th>\n",
       "      <th>fileid</th>\n",
       "      <th>image</th>\n",
       "      <th>line</th>\n",
       "      <th>annotation</th>\n",
       "      <th>bundle</th>\n",
       "      <th>fcat</th>\n",
       "      <th>user</th>\n",
       "      <th>auto</th>\n",
       "      <th>ds</th>\n",
       "      <th>bug</th>\n",
       "      <th>annotation_U1</th>\n",
       "      <th>annotation_U2</th>\n",
       "      <th>annotation_U3</th>\n",
       "      <th>annotation_E1</th>\n",
       "      <th>annotation_E2</th>\n",
       "      <th>annotation_E3</th>\n",
       "      <th>consensus</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>bugs-in-py_PySnooper-1</td>\n",
       "      <td>pysnooper/pycompat.py</td>\n",
       "      <td>0</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>11</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>other</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bug(fix)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>bugs-in-py_PySnooper-1</td>\n",
       "      <td>pysnooper/tracer.py</td>\n",
       "      <td>0</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>17</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>bug(fix) + refactoring</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bug(fix)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>bugs-in-py_PySnooper-1</td>\n",
       "      <td>pysnooper/tracer.py</td>\n",
       "      <td>0</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>18</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>bug(fix) + refactoring</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bug(fix)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>bugs-in-py_PySnooper-1</td>\n",
       "      <td>pysnooper/tracer.py</td>\n",
       "      <td>0</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>89</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bug(fix)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>bugs-in-py_PySnooper-1</td>\n",
       "      <td>pysnooper/tracer.py</td>\n",
       "      <td>0</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>135</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>bug(fix) + refactoring</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bug(fix) + refactoring</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bug(fix) + refactoring</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67958</th>\n",
       "      <td>cve_CVE-2022-31124</td>\n",
       "      <td>tests/fuzzer/fuzz_valid_magic.py</td>\n",
       "      <td>1</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>27</td>\n",
       "      <td>test</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67959</th>\n",
       "      <td>cve_CVE-2022-31124</td>\n",
       "      <td>tests/fuzzer/requirements.txt</td>\n",
       "      <td>1</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>1</td>\n",
       "      <td>test</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>other</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>other</td>\n",
       "      <td>NaN</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67960</th>\n",
       "      <td>cve_CVE-2022-31124</td>\n",
       "      <td>tests/fuzzer/requirements.txt</td>\n",
       "      <td>1</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>2</td>\n",
       "      <td>test</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>other</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>other</td>\n",
       "      <td>NaN</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67961</th>\n",
       "      <td>cve_CVE-2022-31124</td>\n",
       "      <td>tests/fuzzer/requirements.txt</td>\n",
       "      <td>1</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>3</td>\n",
       "      <td>test</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>other</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>other</td>\n",
       "      <td>NaN</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67962</th>\n",
       "      <td>cve_CVE-2022-31124</td>\n",
       "      <td>tests/fuzzer/requirements.txt</td>\n",
       "      <td>1</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>4</td>\n",
       "      <td>test</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>other</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>other</td>\n",
       "      <td>NaN</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>67963 rows × 19 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                           id                              file  fileid  \\\n",
       "0      bugs-in-py_PySnooper-1             pysnooper/pycompat.py       0   \n",
       "1      bugs-in-py_PySnooper-1               pysnooper/tracer.py       0   \n",
       "2      bugs-in-py_PySnooper-1               pysnooper/tracer.py       0   \n",
       "3      bugs-in-py_PySnooper-1               pysnooper/tracer.py       0   \n",
       "4      bugs-in-py_PySnooper-1               pysnooper/tracer.py       0   \n",
       "...                       ...                               ...     ...   \n",
       "67958      cve_CVE-2022-31124  tests/fuzzer/fuzz_valid_magic.py       1   \n",
       "67959      cve_CVE-2022-31124     tests/fuzzer/requirements.txt       1   \n",
       "67960      cve_CVE-2022-31124     tests/fuzzer/requirements.txt       1   \n",
       "67961      cve_CVE-2022-31124     tests/fuzzer/requirements.txt       1   \n",
       "67962      cve_CVE-2022-31124     tests/fuzzer/requirements.txt       1   \n",
       "\n",
       "             image  line annotation  bundle  fcat  user  auto  ds  bug  \\\n",
       "0      afterChange    11   bug(fix)       3     3     3     3   3    3   \n",
       "1      afterChange    17   bug(fix)       3     3     3     3   3    3   \n",
       "2      afterChange    18   bug(fix)       3     3     3     3   3    3   \n",
       "3      afterChange    89   bug(fix)       3     3     3     3   3    3   \n",
       "4      afterChange   135   bug(fix)       3     3     3     3   3    3   \n",
       "...            ...   ...        ...     ...   ...   ...   ...  ..  ...   \n",
       "67958  afterChange    27       test       3     3     3     3   3    3   \n",
       "67959  afterChange     1       test       3     3     3     3   3    3   \n",
       "67960  afterChange     2       test       3     3     3     3   3    3   \n",
       "67961  afterChange     3       test       3     3     3     3   3    3   \n",
       "67962  afterChange     4       test       3     3     3     3   3    3   \n",
       "\n",
       "                annotation_U1 annotation_U2           annotation_U3  \\\n",
       "0                       other           NaN                bug(fix)   \n",
       "1      bug(fix) + refactoring           NaN                bug(fix)   \n",
       "2      bug(fix) + refactoring           NaN                bug(fix)   \n",
       "3                    bug(fix)           NaN                bug(fix)   \n",
       "4      bug(fix) + refactoring           NaN  bug(fix) + refactoring   \n",
       "...                       ...           ...                     ...   \n",
       "67958                    test          test                     NaN   \n",
       "67959                   other      bug(fix)                     NaN   \n",
       "67960                   other      bug(fix)                     NaN   \n",
       "67961                   other      bug(fix)                     NaN   \n",
       "67962                   other      bug(fix)                     NaN   \n",
       "\n",
       "      annotation_E1 annotation_E2 annotation_E3               consensus  \n",
       "0          bug(fix)           NaN           NaN                bug(fix)  \n",
       "1          bug(fix)           NaN           NaN                bug(fix)  \n",
       "2          bug(fix)           NaN           NaN                bug(fix)  \n",
       "3          bug(fix)           NaN           NaN                bug(fix)  \n",
       "4          bug(fix)           NaN           NaN  bug(fix) + refactoring  \n",
       "...             ...           ...           ...                     ...  \n",
       "67958           NaN          test           NaN                    test  \n",
       "67959           NaN         other           NaN                   other  \n",
       "67960           NaN         other           NaN                   other  \n",
       "67961           NaN         other           NaN                   other  \n",
       "67962           NaN         other           NaN                   other  \n",
       "\n",
       "[67963 rows x 19 columns]"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "R"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "d80bbb95-75cf-44ca-a28d-512c98938a9f",
   "metadata": {},
   "outputs": [],
   "source": [
    "def users_vs_auto(row):\n",
    "    ret = Counter()\n",
    "    for u in users_ids:\n",
    "        u_val = row['annotation_' + u]\n",
    "        if not pd.isnull(u_val):\n",
    "            ret[u_val] += 1\n",
    "            val = u_val\n",
    "            \n",
    "    if sum(ret.values()) < 3: # Check if all 3 reviewers annotated line\n",
    "        print(\"Error\")\n",
    "\n",
    "    if ret.most_common(1)[0][1] > 1:\n",
    "        consensus_type =  \"%dxU\" % (ret.most_common(1)[0][1])\n",
    "    else:\n",
    "        return \"other\"\n",
    "    \n",
    "    if ret.most_common(1)[0][0] == row[\"annotation\"]:\n",
    "        consensus_type += \" == A\"\n",
    "    else:\n",
    "        consensus_type += \" != A\"\n",
    "    \n",
    "    return consensus_type"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "9e5444ce-e877-4646-9028-df757be14066",
   "metadata": {},
   "outputs": [],
   "source": [
    "R['UvsA'] = R.apply(users_vs_auto, axis=1)\n",
    "R['ds'] = R['id'].map(lambda x: x.split(\"_\")[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "e34f543a-64b3-4c5f-82ea-2dc2a74ab7f2",
   "metadata": {},
   "outputs": [],
   "source": [
    "cuser_vs_auto = R.groupby(['ds', 'UvsA']).count().reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "a547b151-cc67-41ef-a00e-b2a43227c482",
   "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>ds</th>\n",
       "      <th>UvsA</th>\n",
       "      <th>id</th>\n",
       "      <th>file</th>\n",
       "      <th>fileid</th>\n",
       "      <th>image</th>\n",
       "      <th>line</th>\n",
       "      <th>annotation</th>\n",
       "      <th>bundle</th>\n",
       "      <th>fcat</th>\n",
       "      <th>user</th>\n",
       "      <th>auto</th>\n",
       "      <th>bug</th>\n",
       "      <th>annotation_U1</th>\n",
       "      <th>annotation_U2</th>\n",
       "      <th>annotation_U3</th>\n",
       "      <th>annotation_E1</th>\n",
       "      <th>annotation_E2</th>\n",
       "      <th>annotation_E3</th>\n",
       "      <th>consensus</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>bugs-in-py</td>\n",
       "      <td>2xU != A</td>\n",
       "      <td>706</td>\n",
       "      <td>706</td>\n",
       "      <td>706</td>\n",
       "      <td>706</td>\n",
       "      <td>706</td>\n",
       "      <td>706</td>\n",
       "      <td>706</td>\n",
       "      <td>706</td>\n",
       "      <td>706</td>\n",
       "      <td>706</td>\n",
       "      <td>706</td>\n",
       "      <td>529</td>\n",
       "      <td>581</td>\n",
       "      <td>541</td>\n",
       "      <td>263</td>\n",
       "      <td>107</td>\n",
       "      <td>97</td>\n",
       "      <td>706</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>bugs-in-py</td>\n",
       "      <td>2xU == A</td>\n",
       "      <td>2133</td>\n",
       "      <td>2133</td>\n",
       "      <td>2133</td>\n",
       "      <td>2133</td>\n",
       "      <td>2133</td>\n",
       "      <td>2133</td>\n",
       "      <td>2133</td>\n",
       "      <td>2133</td>\n",
       "      <td>2133</td>\n",
       "      <td>2133</td>\n",
       "      <td>2133</td>\n",
       "      <td>1568</td>\n",
       "      <td>1676</td>\n",
       "      <td>1630</td>\n",
       "      <td>579</td>\n",
       "      <td>435</td>\n",
       "      <td>511</td>\n",
       "      <td>2133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>bugs-in-py</td>\n",
       "      <td>3xU != A</td>\n",
       "      <td>693</td>\n",
       "      <td>693</td>\n",
       "      <td>693</td>\n",
       "      <td>693</td>\n",
       "      <td>693</td>\n",
       "      <td>693</td>\n",
       "      <td>693</td>\n",
       "      <td>693</td>\n",
       "      <td>693</td>\n",
       "      <td>693</td>\n",
       "      <td>693</td>\n",
       "      <td>543</td>\n",
       "      <td>478</td>\n",
       "      <td>612</td>\n",
       "      <td>128</td>\n",
       "      <td>195</td>\n",
       "      <td>123</td>\n",
       "      <td>693</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>bugs-in-py</td>\n",
       "      <td>3xU == A</td>\n",
       "      <td>16331</td>\n",
       "      <td>16331</td>\n",
       "      <td>16331</td>\n",
       "      <td>16331</td>\n",
       "      <td>16331</td>\n",
       "      <td>16331</td>\n",
       "      <td>16331</td>\n",
       "      <td>16331</td>\n",
       "      <td>16331</td>\n",
       "      <td>16331</td>\n",
       "      <td>16331</td>\n",
       "      <td>12016</td>\n",
       "      <td>12684</td>\n",
       "      <td>12510</td>\n",
       "      <td>3216</td>\n",
       "      <td>4434</td>\n",
       "      <td>4133</td>\n",
       "      <td>16331</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>bugs-in-py</td>\n",
       "      <td>other</td>\n",
       "      <td>358</td>\n",
       "      <td>358</td>\n",
       "      <td>358</td>\n",
       "      <td>358</td>\n",
       "      <td>358</td>\n",
       "      <td>358</td>\n",
       "      <td>358</td>\n",
       "      <td>358</td>\n",
       "      <td>358</td>\n",
       "      <td>358</td>\n",
       "      <td>358</td>\n",
       "      <td>218</td>\n",
       "      <td>304</td>\n",
       "      <td>296</td>\n",
       "      <td>162</td>\n",
       "      <td>27</td>\n",
       "      <td>67</td>\n",
       "      <td>358</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>crawl</td>\n",
       "      <td>2xU != A</td>\n",
       "      <td>911</td>\n",
       "      <td>911</td>\n",
       "      <td>911</td>\n",
       "      <td>911</td>\n",
       "      <td>911</td>\n",
       "      <td>911</td>\n",
       "      <td>911</td>\n",
       "      <td>911</td>\n",
       "      <td>911</td>\n",
       "      <td>911</td>\n",
       "      <td>911</td>\n",
       "      <td>568</td>\n",
       "      <td>730</td>\n",
       "      <td>730</td>\n",
       "      <td>380</td>\n",
       "      <td>183</td>\n",
       "      <td>142</td>\n",
       "      <td>911</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>crawl</td>\n",
       "      <td>2xU == A</td>\n",
       "      <td>2942</td>\n",
       "      <td>2942</td>\n",
       "      <td>2942</td>\n",
       "      <td>2942</td>\n",
       "      <td>2942</td>\n",
       "      <td>2942</td>\n",
       "      <td>2942</td>\n",
       "      <td>2942</td>\n",
       "      <td>2942</td>\n",
       "      <td>2942</td>\n",
       "      <td>2942</td>\n",
       "      <td>2397</td>\n",
       "      <td>1993</td>\n",
       "      <td>2041</td>\n",
       "      <td>412</td>\n",
       "      <td>1297</td>\n",
       "      <td>686</td>\n",
       "      <td>2942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>crawl</td>\n",
       "      <td>3xU != A</td>\n",
       "      <td>1523</td>\n",
       "      <td>1523</td>\n",
       "      <td>1523</td>\n",
       "      <td>1523</td>\n",
       "      <td>1523</td>\n",
       "      <td>1523</td>\n",
       "      <td>1523</td>\n",
       "      <td>1523</td>\n",
       "      <td>1523</td>\n",
       "      <td>1523</td>\n",
       "      <td>1523</td>\n",
       "      <td>1316</td>\n",
       "      <td>1235</td>\n",
       "      <td>746</td>\n",
       "      <td>374</td>\n",
       "      <td>661</td>\n",
       "      <td>237</td>\n",
       "      <td>1523</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>crawl</td>\n",
       "      <td>3xU == A</td>\n",
       "      <td>20458</td>\n",
       "      <td>20458</td>\n",
       "      <td>20458</td>\n",
       "      <td>20458</td>\n",
       "      <td>20458</td>\n",
       "      <td>20458</td>\n",
       "      <td>20458</td>\n",
       "      <td>20458</td>\n",
       "      <td>20458</td>\n",
       "      <td>20458</td>\n",
       "      <td>20458</td>\n",
       "      <td>14676</td>\n",
       "      <td>15164</td>\n",
       "      <td>15936</td>\n",
       "      <td>6061</td>\n",
       "      <td>5151</td>\n",
       "      <td>4386</td>\n",
       "      <td>20458</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>crawl</td>\n",
       "      <td>other</td>\n",
       "      <td>268</td>\n",
       "      <td>268</td>\n",
       "      <td>268</td>\n",
       "      <td>268</td>\n",
       "      <td>268</td>\n",
       "      <td>268</td>\n",
       "      <td>268</td>\n",
       "      <td>268</td>\n",
       "      <td>268</td>\n",
       "      <td>268</td>\n",
       "      <td>268</td>\n",
       "      <td>223</td>\n",
       "      <td>190</td>\n",
       "      <td>179</td>\n",
       "      <td>27</td>\n",
       "      <td>77</td>\n",
       "      <td>108</td>\n",
       "      <td>268</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>cve</td>\n",
       "      <td>2xU != A</td>\n",
       "      <td>1578</td>\n",
       "      <td>1578</td>\n",
       "      <td>1578</td>\n",
       "      <td>1578</td>\n",
       "      <td>1578</td>\n",
       "      <td>1578</td>\n",
       "      <td>1578</td>\n",
       "      <td>1578</td>\n",
       "      <td>1578</td>\n",
       "      <td>1578</td>\n",
       "      <td>1578</td>\n",
       "      <td>779</td>\n",
       "      <td>1347</td>\n",
       "      <td>1393</td>\n",
       "      <td>689</td>\n",
       "      <td>227</td>\n",
       "      <td>299</td>\n",
       "      <td>1578</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>cve</td>\n",
       "      <td>2xU == A</td>\n",
       "      <td>2172</td>\n",
       "      <td>2172</td>\n",
       "      <td>2172</td>\n",
       "      <td>2172</td>\n",
       "      <td>2172</td>\n",
       "      <td>2172</td>\n",
       "      <td>2172</td>\n",
       "      <td>2172</td>\n",
       "      <td>2172</td>\n",
       "      <td>2172</td>\n",
       "      <td>2172</td>\n",
       "      <td>1873</td>\n",
       "      <td>1653</td>\n",
       "      <td>1356</td>\n",
       "      <td>370</td>\n",
       "      <td>761</td>\n",
       "      <td>503</td>\n",
       "      <td>2172</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>cve</td>\n",
       "      <td>3xU != A</td>\n",
       "      <td>1943</td>\n",
       "      <td>1943</td>\n",
       "      <td>1943</td>\n",
       "      <td>1943</td>\n",
       "      <td>1943</td>\n",
       "      <td>1943</td>\n",
       "      <td>1943</td>\n",
       "      <td>1943</td>\n",
       "      <td>1943</td>\n",
       "      <td>1943</td>\n",
       "      <td>1943</td>\n",
       "      <td>668</td>\n",
       "      <td>1793</td>\n",
       "      <td>1742</td>\n",
       "      <td>1255</td>\n",
       "      <td>179</td>\n",
       "      <td>192</td>\n",
       "      <td>1943</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>cve</td>\n",
       "      <td>3xU == A</td>\n",
       "      <td>15702</td>\n",
       "      <td>15702</td>\n",
       "      <td>15702</td>\n",
       "      <td>15702</td>\n",
       "      <td>15702</td>\n",
       "      <td>15702</td>\n",
       "      <td>15702</td>\n",
       "      <td>15702</td>\n",
       "      <td>15702</td>\n",
       "      <td>15702</td>\n",
       "      <td>15702</td>\n",
       "      <td>13311</td>\n",
       "      <td>8894</td>\n",
       "      <td>13152</td>\n",
       "      <td>2093</td>\n",
       "      <td>3969</td>\n",
       "      <td>5687</td>\n",
       "      <td>15702</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>cve</td>\n",
       "      <td>other</td>\n",
       "      <td>245</td>\n",
       "      <td>245</td>\n",
       "      <td>245</td>\n",
       "      <td>245</td>\n",
       "      <td>245</td>\n",
       "      <td>245</td>\n",
       "      <td>245</td>\n",
       "      <td>245</td>\n",
       "      <td>245</td>\n",
       "      <td>245</td>\n",
       "      <td>245</td>\n",
       "      <td>195</td>\n",
       "      <td>188</td>\n",
       "      <td>185</td>\n",
       "      <td>50</td>\n",
       "      <td>88</td>\n",
       "      <td>29</td>\n",
       "      <td>245</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            ds      UvsA     id   file  fileid  image   line  annotation  \\\n",
       "0   bugs-in-py  2xU != A    706    706     706    706    706         706   \n",
       "1   bugs-in-py  2xU == A   2133   2133    2133   2133   2133        2133   \n",
       "2   bugs-in-py  3xU != A    693    693     693    693    693         693   \n",
       "3   bugs-in-py  3xU == A  16331  16331   16331  16331  16331       16331   \n",
       "4   bugs-in-py     other    358    358     358    358    358         358   \n",
       "5        crawl  2xU != A    911    911     911    911    911         911   \n",
       "6        crawl  2xU == A   2942   2942    2942   2942   2942        2942   \n",
       "7        crawl  3xU != A   1523   1523    1523   1523   1523        1523   \n",
       "8        crawl  3xU == A  20458  20458   20458  20458  20458       20458   \n",
       "9        crawl     other    268    268     268    268    268         268   \n",
       "10         cve  2xU != A   1578   1578    1578   1578   1578        1578   \n",
       "11         cve  2xU == A   2172   2172    2172   2172   2172        2172   \n",
       "12         cve  3xU != A   1943   1943    1943   1943   1943        1943   \n",
       "13         cve  3xU == A  15702  15702   15702  15702  15702       15702   \n",
       "14         cve     other    245    245     245    245    245         245   \n",
       "\n",
       "    bundle   fcat   user   auto    bug  annotation_U1  annotation_U2  \\\n",
       "0      706    706    706    706    706            529            581   \n",
       "1     2133   2133   2133   2133   2133           1568           1676   \n",
       "2      693    693    693    693    693            543            478   \n",
       "3    16331  16331  16331  16331  16331          12016          12684   \n",
       "4      358    358    358    358    358            218            304   \n",
       "5      911    911    911    911    911            568            730   \n",
       "6     2942   2942   2942   2942   2942           2397           1993   \n",
       "7     1523   1523   1523   1523   1523           1316           1235   \n",
       "8    20458  20458  20458  20458  20458          14676          15164   \n",
       "9      268    268    268    268    268            223            190   \n",
       "10    1578   1578   1578   1578   1578            779           1347   \n",
       "11    2172   2172   2172   2172   2172           1873           1653   \n",
       "12    1943   1943   1943   1943   1943            668           1793   \n",
       "13   15702  15702  15702  15702  15702          13311           8894   \n",
       "14     245    245    245    245    245            195            188   \n",
       "\n",
       "    annotation_U3  annotation_E1  annotation_E2  annotation_E3  consensus  \n",
       "0             541            263            107             97        706  \n",
       "1            1630            579            435            511       2133  \n",
       "2             612            128            195            123        693  \n",
       "3           12510           3216           4434           4133      16331  \n",
       "4             296            162             27             67        358  \n",
       "5             730            380            183            142        911  \n",
       "6            2041            412           1297            686       2942  \n",
       "7             746            374            661            237       1523  \n",
       "8           15936           6061           5151           4386      20458  \n",
       "9             179             27             77            108        268  \n",
       "10           1393            689            227            299       1578  \n",
       "11           1356            370            761            503       2172  \n",
       "12           1742           1255            179            192       1943  \n",
       "13          13152           2093           3969           5687      15702  \n",
       "14            185             50             88             29        245  "
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cuser_vs_auto"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "86409107-502f-451d-a168-4e4f86aed5e9",
   "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>id</th>\n",
       "      <th>file</th>\n",
       "      <th>fileid</th>\n",
       "      <th>image</th>\n",
       "      <th>line</th>\n",
       "      <th>annotation</th>\n",
       "      <th>bundle</th>\n",
       "      <th>fcat</th>\n",
       "      <th>user</th>\n",
       "      <th>auto</th>\n",
       "      <th>bug</th>\n",
       "      <th>annotation_U1</th>\n",
       "      <th>annotation_U2</th>\n",
       "      <th>annotation_U3</th>\n",
       "      <th>annotation_E1</th>\n",
       "      <th>annotation_E2</th>\n",
       "      <th>annotation_E3</th>\n",
       "      <th>consensus</th>\n",
       "      <th>sort</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ds</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>bugs-in-py</th>\n",
       "      <td>20221</td>\n",
       "      <td>20221</td>\n",
       "      <td>20221</td>\n",
       "      <td>20221</td>\n",
       "      <td>20221</td>\n",
       "      <td>20221</td>\n",
       "      <td>20221</td>\n",
       "      <td>20221</td>\n",
       "      <td>20221</td>\n",
       "      <td>20221</td>\n",
       "      <td>20221</td>\n",
       "      <td>14874</td>\n",
       "      <td>15723</td>\n",
       "      <td>15589</td>\n",
       "      <td>4348</td>\n",
       "      <td>5198</td>\n",
       "      <td>4931</td>\n",
       "      <td>20221</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>crawl</th>\n",
       "      <td>26102</td>\n",
       "      <td>26102</td>\n",
       "      <td>26102</td>\n",
       "      <td>26102</td>\n",
       "      <td>26102</td>\n",
       "      <td>26102</td>\n",
       "      <td>26102</td>\n",
       "      <td>26102</td>\n",
       "      <td>26102</td>\n",
       "      <td>26102</td>\n",
       "      <td>26102</td>\n",
       "      <td>19180</td>\n",
       "      <td>19312</td>\n",
       "      <td>19632</td>\n",
       "      <td>7254</td>\n",
       "      <td>7369</td>\n",
       "      <td>5559</td>\n",
       "      <td>26102</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cve</th>\n",
       "      <td>21640</td>\n",
       "      <td>21640</td>\n",
       "      <td>21640</td>\n",
       "      <td>21640</td>\n",
       "      <td>21640</td>\n",
       "      <td>21640</td>\n",
       "      <td>21640</td>\n",
       "      <td>21640</td>\n",
       "      <td>21640</td>\n",
       "      <td>21640</td>\n",
       "      <td>21640</td>\n",
       "      <td>16826</td>\n",
       "      <td>13875</td>\n",
       "      <td>17828</td>\n",
       "      <td>4457</td>\n",
       "      <td>5224</td>\n",
       "      <td>6710</td>\n",
       "      <td>21640</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               id   file  fileid  image   line  annotation  bundle   fcat  \\\n",
       "ds                                                                          \n",
       "bugs-in-py  20221  20221   20221  20221  20221       20221   20221  20221   \n",
       "crawl       26102  26102   26102  26102  26102       26102   26102  26102   \n",
       "cve         21640  21640   21640  21640  21640       21640   21640  21640   \n",
       "\n",
       "             user   auto    bug  annotation_U1  annotation_U2  annotation_U3  \\\n",
       "ds                                                                             \n",
       "bugs-in-py  20221  20221  20221          14874          15723          15589   \n",
       "crawl       26102  26102  26102          19180          19312          19632   \n",
       "cve         21640  21640  21640          16826          13875          17828   \n",
       "\n",
       "            annotation_E1  annotation_E2  annotation_E3  consensus  sort  \n",
       "ds                                                                        \n",
       "bugs-in-py           4348           5198           4931      20221    10  \n",
       "crawl                7254           7369           5559      26102    10  \n",
       "cve                  4457           5224           6710      21640    10  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmYAAAITCAYAAACtw9D4AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjYuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/P9b71AAAACXBIWXMAAA9hAAAPYQGoP6dpAAA9h0lEQVR4nO3deVxU9f7H8fewgyzikpKYZoCk4UqLaWiZebVbUd2sXJJ2M5fSXLiapS2a2WLltUy0NNGumV1vkWZd8YKVueGSpaaipaKpsUSCBuf3hz/nRig6w8B8cV7Px2MeOed8zzmfTyfk3fecOWOzLMsSAAAA3M7L3QUAAADgJIIZAACAIQhmAAAAhiCYAQAAGIJgBgAAYAiCGQAAgCEIZgAAAIYgmAEAABiCYAYAAGAIghkAAIAhCGYAAACGIJgBAAAYgmAGAABgCIIZAACAIQhmAAAAhvBxdwFATdF+xJxqPd66F+9xaHyXLl3Upk0bvfrqq1VTkBPS09N17bXX6pdfflHt2rXdXU612jshrtqOddG4zdV2LFd555139Nhjjyk3N9ctx+/4esdqO9aqwauq7Vio+ZgxA1Blrr76ah04cEBhYWHuLgUAagSCGYAq4+fnp4YNG8pms7m7FFTCiRMn3F0C4DEIZsB55Pfff9egQYNUu3Zt1a1bV2PHjpVlWZIkm82mjz76qMz42rVr65133rG///LLL9WmTRsFBAQoPj5eH330kWw2m7KysiRJv/zyi/r06aP69esrMDBQ0dHRmj179hnrSU9Pl81ms1+ueuedd1S7dm0tW7ZMl156qYKDg/WXv/xFBw4cqLCvLl26aNCgQWfsbcKECYqLK3/psH379ho3btxZ/q15ptLSUr3wwguKioqSv7+/LrroIj333HPKzs6WzWbTP//5T3Xp0kUBAQF67733dOTIEd19992KjIxUUFCQ4uLiNH/+fPv+/v3vf6t27doqLS2VJGVlZclms2nEiBH2MQ8//LDuvvvuau+1JjrT+enQoYNGjx5dZuzPP/8sX19frVixQpJ0/PhxjRw5Uo0aNVKtWrV05ZVXKj093Q1dwBkEM+A88u6778rHx0erV6/Wa6+9pldeeUUzZ848p20LCgp00003KS4uTuvXr9czzzyjUaNGlRnz5JNPauvWrfr000/13Xffafr06apXr55DNf7222+aMmWK5s6dq//+97/au3evnnjiiUr1dt9992nr1q1as2aNffymTZu0YcMGJSUlOVSfp0hOTtYLL7xgP6epqalq0KCBff2oUaM0ZMgQfffdd+revbuKiorUvn17ffzxx9qyZYseeugh9evXT6tXr5YkJSQkqKCgQBs2bJAkrVy5UvXq1dPKlSvt+0xPT1fnzp2rt9Ea6kznp0+fPpo/f779f0ok6f3331eDBg3s/27vvfderVq1SgsWLNCmTZt0xx136C9/+Yt27NjhrnbgAG7+B84jjRs31iuvvCKbzabmzZtr8+bNeuWVV/Tggw+eddt58+bJZrPp7bffVkBAgFq0aKF9+/aV2Xbv3r1q27at4uPjJUlNmzZ1uMYTJ07ozTff1CWXXCJJGjRokCZMmFCp3iIjI9W9e3fNnj1bl19+uSRp9uzZ6ty5s5o1a+Zwjee7goICTZ06VW+88Yb69+8vSbrkkkvUqVMnZWdnS5Iee+wx3XbbbWW2+2OAHjx4sJYuXaqFCxfqyiuvVFhYmNq0aaP09HS1b99e6enpevzxxzV+/HgVFBSosLBQ27dvV5cuXaqrzRqrovPz888/6/HHH1dmZqauueYaSVJqaqp69+4tLy8v7dy5U/Pnz9dPP/2kCy+8UNLJ87Z06VLNnj1bzz//vNv6wrlhxgw4j1x11VVl7ufq0KGDduzYoZKSkrNuu23bNrVq1UoBAQH2ZVdccUWZMY888ogWLFigNm3aaOTIkfryyy/t63r06KHg4GAFBwerZcuWZzxOUFCQPZRJUkREhA4dOiRJysjIsO8jODhY8+bNO+feHnzwQc2fP19FRUU6ceKE5s2bp/vuu++sfXui7777TsXFxeratesZx5wK36eUlJToueeeU6tWrVS3bl0FBwfrs88+0969e+1junTpovT0dFmWpYyMDN1yyy267LLLlJmZqRUrVqhBgwaKjY2tsr7OFxWdn/r166tbt272n43du3frq6++Up8+fSRJ69evl2VZiomJKfOztHLlSu3cubNa+4BzmDEDPITNZitz+UMqe1O3ZVnlbtL/8/gePXpoz549+uSTT/T555+ra9euevTRRzVlyhTNnDlTx44dkyT5+vqesY4/r/tjXfHx8fb72SSVubR2NjfddJP8/f21ePFi+fv7q7i4WLfffvs5b+9JAgMDzzqmVq1aZd6/9NJLeuWVV/Tqq68qLi5OtWrV0mOPPabjx4/bx3Tp0kUpKSnauHGjvLy81KJFC3Xu3FkrV67UL7/8wmXMc3S289OnTx8NHTpUr7/+ulJTU9WyZUu1bt1a0sl707y9vbVu3Tp5e3uX2S44OLjKaobrMGMGnEe+/vrrcu+jo6Pl7e2t+vXrl7nJfseOHfrtt9/s72NjY7Vp0yYVFxfbl61du7bcMerXr6+kpCS99957evXVVzVjxgxJUqNGjRQVFaWoqCg1adLEqfoDAwPt+4iKilJISMg59SZJPj4+6t+/v2bPnq3Zs2frrrvuUlBQkFN1nO+io6MVGBioL7744py3OTUD1rdvX7Vu3VrNmjUrd8/SqfvMXn31VXXu3Fk2m02dO3dWeno695c54GznJzExUUVFRVq6dKlSU1PVt29f+7q2bduqpKREhw4dKvOzFBUVpYYNG1ZXC6gEZsyA88iPP/6oYcOG6eGHH9b69ev1+uuv66WXXpIkXXfddXrjjTd01VVXqbS0VKNGjSoze9W7d2+NGTNGDz30kEaPHq29e/dqypQpkmSfSRs3bpzat2+vli1bqri4WB9//LEuvfRSt/d2ygMPPGCvZ9UqHup5JgEBARo1apRGjhwpPz8/dezYUT///LO+/fbbM17ejIqK0qJFi/Tll18qPDxcL7/8snJycsqc/1P3mb333nuaOnWqpJNh7Y477tCJEye4v+wcVXR+7r//ftWqVUu33HKLnnzySX333Xfq3bu3fduYmBj16dNH99xzj1566SW1bdtWhw8f1n/+8x/FxcWpZ8+ebuwM54JgBpwjR5/E7w733HOPjh07piuuuELe3t4aPHiwHnroIUknL0Xde++9SkhI0IUXXqipU6dq3bp19m1DQ0P173//W4888ojatGmjuLg4jRs3Tr1797bfd+bn56fk5GRlZ2crMDBQ11xzjRYsWOD23k6Jjo7W1VdfrSNHjujKK6+slrrOxPSn8T/55JPy8fHRuHHjtH//fkVERGjAgAEVjt+9e7e6d++uoKAgPfTQQ0pMTFReXl6Zcddee63Wr19vD2Hh4eFq0aKF9u/fX20h/lyY/jT+s52fPn366MYbb1RCQoIuuuiiMtvOnj1bzz77rIYPH659+/apbt266tChA6GshrBZf76JBAD+37x583TvvfcqLy/vnO5Lqirn+nVTlmUpNjZWDz/8sIYNG1Y9xQGACzFjBsBuzpw5atasmRo1aqSNGzdq1KhR6tWrl1tD2bk6dOiQ5s6dq3379unee+91dzkA4BSCGQC7nJwcjRs3Tjk5OYqIiNAdd9yh5557zt1lnZMGDRqoXr16mjFjhsLDw91dDgA4hUuZAAAAhuBxGQAAAIYgmAEAABiCYAYAAGAIghkAAIAhCGYAAACGIJgBAAAYgmAGAABgCIIZAACAIQhmAAAAhiCYAQAAGIJgBgAAYAiCGQAAgCEIZgAAAIYgmAEAABiCYAYAAGAIghkAAIAhCGYAAACGIJgBAAAYgmAGAABgCIIZAACAIQhmAAAAhiCYAQAAGIJgBgAAYAiCGQAAgCEIZgAAAIYgmAEAABiCYAYAAGAIghkAAIAhCGYAAACGIJgBAAAYgmAGAABgCIIZAACAIQhmAAAAhiCYAQAAGIJgBgAAYAiCGQAAgCEIZgAAAIYgmAEAABiCYAYAAGAIghkAAIAhCGYAAACGIJgBAAAYgmAGAABgCIIZAACAIQhmAAAAhiCYAQAAGIJgBgAAYAiCGQAAgCEIZgAAAIYgmAEAABiCYAYAAGAIghkAAIAhCGYAAACGIJgBAAAYgmAGAABgCIIZAACAIQhmAAAAhiCYAQAAGIJgBgAAYAiCGQAAgCEIZgAAAIYgmAEAABiCYAYAAGAIghkAAIAhCGYAAACGIJgBAAAYgmAGAABgCIIZAACAIQhmAAAAhiCYAQAAGIJgBgAAYAiCGQAAgCEIZgAAAIYgmAEAABiCYAYAAGAIghkAAIAhfNxdAM5daWmp9u/fr5CQENlsNneXAwAAzoFlWSooKNCFF14oL6+K58QIZjXI/v371bhxY3eXAQAAnPDjjz8qMjKywjEEsxokJCRE0skTGxoa6uZqAADAucjPz1fjxo3tv8crQjCrQU5dvgwNDSWYAQBQw5zLbUjc/A8AAGAIghkAAIAhCGYAAACGIJgBAAAYgmAGAABgCD6VWQMljJ0vb//As45b9+I91VANAABwFWbMAAAADEEwAwAAMATBDAAAwBAEMwAAAEMQzAAAAAxBMAMAADAEwQwAAMAQBDMAAABDEMwAAAAMQTADAAAwBMEMAADAEAQzAAAAQxDMAAAADEEwAwAAMATBDAAAwBAEMwAAAEMQzAAAAAxBMAMAADCEj7sLgOPmBU9VSIB3mWUXjdvspmoAAICrMGMGAABgCIIZAACAIQhmAAAAhiCYAQAAGIJg5qAuXbrosccec3cZAADgPEQwO4P09HTZbDbl5ua6uxQAAOAhCGYGOHHihLtLAAAABvDoYFZcXKwhQ4boggsuUEBAgDp16qQ1a9YoOztb1157rSQpPDxcNptNSUlJ9u1KS0s1cuRI1alTRw0bNtTTTz9dZr95eXl66KGHdMEFFyg0NFTXXXedNm7caF//9NNPq02bNpo1a5aaNWsmf39/WZZVHS0DAACDeXQwGzlypBYtWqR3331X69evV1RUlLp3766QkBAtWrRIkrRt2zYdOHBAU6dOtW/37rvvqlatWlq9erUmT56sCRMmaPny5ZIky7J04403KicnR2lpaVq3bp3atWunrl276ujRo/Z9/PDDD/rnP/+pRYsWKSsr67T1FRcXKz8/v8wLAACcvzw2mBUWFmr69Ol68cUX1aNHD7Vo0UJvv/22AgMDNWvWLNWpU0eSdMEFF6hhw4YKCwuzb9uqVSs99dRTio6O1j333KP4+Hh98cUXkqQVK1Zo8+bNWrhwoeLj4xUdHa0pU6aodu3a+uCDD+z7OH78uObOnau2bduqVatWstls5WqcOHGiwsLC7K/GjRtX8b8VAADgTh4bzHbu3KkTJ06oY8eO9mW+vr664oor9N1331W4batWrcq8j4iI0KFDhyRJ69at06+//qq6desqODjY/tq9e7d27txp36ZJkyaqX79+hcdJTk5WXl6e/fXjjz862iYAAKhBPPa7Mk/d0/XnmSrLsk47e/VHvr6+Zd7bbDaVlpZKOnn/WUREhNLT08ttV7t2bfufa9WqddYa/f395e/vf9ZxAADg/OCxM2ZRUVHy8/NTZmamfdmJEye0du1aXXrppfLz85MklZSUOLTfdu3aKScnRz4+PoqKiirzqlevnkt7AAAA5xePDWa1atXSI488ohEjRmjp0qXaunWrHnzwQf3222+6//771aRJE9lsNn388cf6+eef9euvv57Tfq+//np16NBBiYmJWrZsmbKzs/Xll19q7NixWrt2bRV3BQAAajKPDWaSNGnSJN1+++3q16+f2rVrpx9++EHLli1TeHi4GjVqpPHjx2v06NFq0KCBBg0adE77tNlsSktLU0JCgu677z7FxMTorrvuUnZ2tho0aFDFHQEAgJrMZvEArRojPz9fYWFh2pJ8qUICvMusu2jcZjdVBQAAKnLq93deXp5CQ0MrHOvRM2YAAAAmIZgBAAAYgmAGAABgCI99jllNdl/tEPkE/unUvd5Rqwavck9BAADAJZgxAwAAMATBDAAAwBAEMwAAAEMQzAAAAAxBMAMAADAEwQwAAMAQBDMAAABDEMwAAAAMQTADAAAwBMEMAADAEAQzAAAAQxDMAAAADEEwAwAAMATBDAAAwBAEMwAAAEMQzAAAAAxBMAMAADAEwQwAAMAQBDMAAABD+Li7ADhu+YDlCg0NdXcZAADAxZgxAwAAMATBDAAAwBAEMwAAAEMQzAAAAAxBMAMAADAEwQwAAMAQBDMAAABDEMwAAAAMQTADAAAwBMEMAADAEAQzAAAAQxDMAAAADEEwAwAAMISPuwuA4xLGzpe3f6DL9rfuxXtcti8AAOA8ZswAAAAMQTADAAAwBMEMAADAEAQzAAAAQxDMAAAADEEwAwAAMATBDAAAwBAEMwAAAEMQzAAAAAxBMAMAADAEwQwAAMAQBDMAAABDEMwAAAAMQTADAAAwBMEMAADAEAQzAAAAQxDMAAAADEEwAwAAMATBDAAAwBA+7i4AjpsXPFUhAd4u3OM9LtwXAABwFjNmAAAAhiCYAQAAGIJgBgAAYAiCGQAAgCEIZgAAAIYwMphNnDhRl19+uUJCQnTBBRcoMTFR27Ztc2gfSUlJSkxMLLc8KytLNptN2dnZZ9w2OztbNptNWVlZjhV+jpo3by4/Pz/t27evSvYPAABqJiOD2cqVK/Xoo4/q66+/1vLly/X777/rhhtuUGFhobtLq7TMzEwVFRXpjjvu0DvvvOPucgAAgEGMDGZLly5VUlKSWrZsqdatW2v27Nnau3ev1q1bJ0n6/vvvFRQUpNTUVPs2H374oQICArR582Z3lX1OUlJS1Lt3b/Xr10+zZs2SZVnuLgkAABjCyGD2Z3l5eZKkOnXqSJJiY2M1ZcoUDRw4UHv27NH+/fv14IMPatKkSYqLi6uSGjIyMhQcHFzh6/nnn69wHwUFBVq4cKH69u2rbt26qbCwUOnp6WccX1xcrPz8/DIvAABw/jL+yf+WZWnYsGHq1KmTLrvsMvvygQMHKi0tTf369ZOfn5/at2+voUOHuuSYXl5eZf4pSfHx8We95+xUcDyTBQsWKDo6Wi1btpQk3XXXXUpJSdG111572vETJ07U+PHjHagcAADUZMYHs0GDBmnTpk3KzMwst27WrFmKiYmRl5eXtmzZIpvN5pJjBgcHl/mnJAUGBioqKqpS+01JSVHfvn3t7/v27auEhATl5uaqdu3a5cYnJydr2LBh9vf5+flq3LhxpWoAAADmMvpS5uDBg7VkyRKtWLFCkZGR5dZv3LhRhYWFKiwsVE5OTpl1oaGh9kugf5SbmytJCgsLO+NxT637YzCr7KXMrVu3avXq1Ro5cqR8fHzk4+Ojq666SseOHdP8+fNPu42/v79CQ0PLvAAAwPnLyBkzy7I0ePBgLV68WOnp6br44ovLjTl69KiSkpI0ZswY5eTkqE+fPlq/fr0CAwMlnbwPbf78+SoqKlJAQIB9uzVr1qh+/foKDw8/4/G9vb3L3ZRf2UuZKSkpSkhI0LRp08osnzt3rlJSUvTII49UuG8AAHD+s1kGfixw4MCBSk1N1b/+9S81b97cvjwsLMwevHr16qW9e/cqMzNTx48fV7t27dS1a1d78MnLy1NsbKw6deqkUaNGKTw8XF999ZWGDBmi5ORkjRgx4ozH37dvn7p27aqPPvpIsbGxle7nxIkTatSokSZMmKABAwaUWbdjxw7FxMQoKytLrVu3rnA/+fn5CgsL05bkSxUS4F3puk65aJzZn2QFAKAmO/X7Oy8v76xXv4y8lDl9+nTl5eWpS5cuioiIsL/ef/99SdKcOXOUlpamuXPnysfHR0FBQZo3b55mzpyptLQ0SSdDXEZGhizLUmJiolq3bq3JkyfrmWee0fDhwys8/okTJ7Rt2zYVFRW5pJ8lS5boyJEjuvXWW8uti46OVlxcnFJSUlxyLAAAUHMZOWOG02PGDACAmqfGz5gBAAB4IoIZAACAIQhmAAAAhjDycRmo2H21Q+QT6MJT93pH1+3rD1YNXlUl+wUA4HzFjBkAAIAhCGYAAACGIJgBAAAYgmAGAABgCIIZAACAIQhmAAAAhiCYAQAAGIJgBgAAYAiCGQAAgCEIZgAAAIYgmAEAABiCYAYAAGAIghkAAIAhCGYAAACGIJgBAAAYgmAGAABgCIIZAACAIQhmAAAAhiCYAQAAGMLH3QXAccsHLFdoaKi7ywAAAC7GjBkAAIAhCGYAAACGIJgBAAAYgmAGAABgCIIZAACAIQhmAAAAhiCYAQAAGIJgBgAAYAinglmzZs105MiRcstzc3PVrFmzShcFAADgiZwKZtnZ2SopKSm3vLi4WPv27at0UQAAAJ7Ioa9kWrJkif3Py5YtU1hYmP19SUmJvvjiCzVt2tRlxQEAAHgSh4JZYmKiJMlms6l///5l1vn6+qpp06Z66aWXXFYcAACAJ3EomJWWlkqSLr74Yq1Zs0b16tWrkqIAAAA8kUPB7JTdu3e7ug44IGHsfHn7B7q7jGq17sV73F0CAABVzqlgJklffPGFvvjiCx06dMg+k3bKrFmzKl0YAACAp3EqmI0fP14TJkxQfHy8IiIiZLPZXF0XAACAx3EqmL355pt655131K9fP1fXAwAA4LGceo7Z8ePHdfXVV7u6FgAAAI/mVDB74IEHlJqa6upaAAAAPJpTlzKLioo0Y8YMff7552rVqpV8fX3LrH/55ZddUhwAAIAncSqYbdq0SW3atJEkbdmypcw6PggAAADgHKeC2YoVK1xdBwAAgMdz6h6zU3744QctW7ZMx44dkyRZluWSogAAADyRU8HsyJEj6tq1q2JiYtSzZ08dOHBA0skPBQwfPtylBQIAAHgKp4LZ448/Ll9fX+3du1dBQUH25XfeeaeWLl3qsuIAAAA8iVP3mH322WdatmyZIiMjyyyPjo7Wnj17XFIYAACAp3FqxqywsLDMTNkphw8flr+/f6WLAgAA8EROBbOEhATNmTPH/t5ms6m0tFQvvviirr32WpcVBwAA4EmcupT54osvqkuXLlq7dq2OHz+ukSNH6ttvv9XRo0e1atUqV9cIAADgEZyaMWvRooU2bdqkK664Qt26dVNhYaFuu+02bdiwQZdccomrawQAAPAITs2YSVLDhg01fvx4V9YCAADg0ZwOZkVFRdq0aZMOHTqk0tLSMutuvvnmShcGAADgaZwKZkuXLtU999yjw4cPl1tns9lUUlJS6cIAAAA8jc1y4nuUoqKi1L17d40bN04NGjSoirpwGvn5+QoLC9OW5EsVEuDt7nLc6qJxm91dAgAA5+TU7++8vDyFhoZWONapm/8PHTqkYcOGEcoAAABcyKlg9re//U3p6ekuLgUAAMCzOXWP2RtvvKE77rhDGRkZiouLk6+vb5n1Q4YMcUlxAAAAnsSpYJaamqply5YpMDBQ6enpstls9nU2m41gBgAA4ASngtnYsWM1YcIEjR49Wl5eTl0NBQAAwJ84laqOHz+uO++8s9KhbOLEibr88ssVEhKiCy64QImJidq2bZtD+0hKSlJiYmK55VlZWbLZbMrOzq5UjVXhoYcekre3txYsWODuUgAAgEGcSlb9+/fX+++/X+mDr1y5Uo8++qi+/vprLV++XL///rtuuOEGFRYWVnrfpvrtt9/0/vvva8SIEUpJSXF3OQAAwCBOXcosKSnR5MmTtWzZMrVq1arczf8vv/zyOe1n6dKlZd7Pnj1bF1xwgdatW6eEhAR9//33ateunWbOnKnevXtLkj788EP17t1ba9asUVxcnDPlu9XChQvVokULJScnKyIiQtnZ2WratKm7ywIAAAZwKpht3rxZbdu2lSRt2bKlzLo/fhDAUXl5eZKkOnXqSJJiY2M1ZcoUDRw4UB07dpSvr68efPBBTZo0ySWhrEePHsrIyKhwzK+//mr/c3BwcIVjr7nmGn366acVjklJSVHfvn0VFhamnj17avbs2Wf8ztHi4mIVFxfb3+fn51e4bwAAULM59eT/qmBZlm655Rb98ssv5cLSX//6V+Xn58vPz09eXl5atmyZPQAmJSUpNzdXH330UZltsrKy1LZtW+3evfuMM1L79u3TsWPHKqwrKirK/ucffvihwrGBgYFq1KjRGdfv2LFDLVu21P79+1WvXj199NFHGjJkiLKzs097v97TTz992tDGk/958j8AoOZw5Mn/Tn+JuXQyqOzcuVMJCQkKDAyUZVlOz5gNGjRImzZtUmZmZrl1s2bNUkxMjLy8vLRly5ZKzcr9UUUh6nT+GNKckZKSou7du6tevXqSpJ49e+r+++/X559/rhtuuKHc+OTkZA0bNsz+Pj8/X40bN65UDQAAwFxOBbMjR46oV69eWrFihWw2m3bs2KFmzZrpgQceUO3atfXSSy85tL/BgwdryZIl+u9//6vIyMhy6zdu3KjCwkJ5eXkpJydHF154oX1daGio9uzZU26b3NxcSVJYWNgZj1udlzJLSko0Z84c5eTkyMfHp8zylJSU0wYzf39/+fv7V3hMAABw/nAqmD3++OPy9fXV3r17demll9qX33nnnXr88cfPOZhZlqXBgwdr8eLFSk9P18UXX1xuzNGjR5WUlKQxY8YoJydHffr00fr16xUYGCjp5H1o8+fPV1FRkQICAuzbrVmzRvXr11d4ePgZjz9z5syzXsr8o6ysrArXn6rpdNLS0lRQUKANGzbI2/t/lyG///579enTR0eOHFHdunXPuRYAAHD+ceoes4YNG2rZsmVq3bq1QkJCtHHjRjVr1ky7d+9WXFxcmVmmigwcOFCpqan617/+pebNm9uXh4WF2UNOr169tHfvXmVmZur48eNq166dunbtqmnTpkk6+YGB2NhYderUSaNGjVJ4eLi++uorDRkyRMnJyRoxYoSj7VWJxMREBQQElHt2mWVZaty4sUaMGKGhQ4dWuI9T16i5x4x7zAAANYcj95g59RyzwsJCBQUFlVt++PBhhy69TZ8+XXl5eerSpYsiIiLsr1PPSJszZ47S0tI0d+5c+fj4KCgoSPPmzdPMmTOVlpYm6WSIy8jIkGVZSkxMVOvWrTV58mQ988wzGj58uDPtudzBgwf1ySef6Pbbby+3zmaz6bbbbuOZZgAAwLkZsxtvvFHt2rXTM888o5CQEG3atElNmjTRXXfdpdLSUn3wwQdVUavHY8bsf5gxAwDUFFX+qcwXX3xRXbp00dq1a3X8+HGNHDlS3377rY4ePapVq1Y5VTQAAICnc+pSZosWLbRp0yZdfvnl6tatmwoLC3Xbbbdpw4YNuuSSS1xdIwAAgEdw+jlmDRs21IQJE1xZCwAAgEdzOphlZGTorbfe0q5du7Rw4UI1atRIc+fO1cUXX6xOnTq5skb8yX21Q+QTWKlnA9cIqwZzWRwA4FmcupS5aNEide/eXYGBgVq/fr39+xwLCgr0/PPPu7RAAAAAT+FUMHv22Wf15ptv6u2335avr699+dVXX63169e7rDgAAABP4lQw27ZtmxISEsotDw0NtX8VEgAAABzjVDCLiIjQDz/8UG55ZmammjVrVumiAAAAPJFTwezhhx/W0KFDtXr1atlsNu3fv1/z5s3TE088oYEDB7q6RgAAAI/g1Ef7Ro4cqby8PF177bUqKipSQkKC/P399cQTT2jQoEGurhEAAMAjOBzMSkpKlJmZqeHDh2vMmDHaunWrSktL1aJFCwUHB1dFjQAAAB7B4WDm7e2t7t2767vvvlOdOnUUHx9fFXUBAAB4HKfuMYuLi9OuXbtcXQsAAIBHcyqYPffcc3riiSf08ccf68CBA8rPzy/zAgAAgOOcuvn/L3/5iyTp5ptvls1msy+3LEs2m00lJSWuqQ4AAMCDOBXMVqxY4eo6AAAAPJ5Twaxz586urgMAAMDjORXMNm3adNrlNptNAQEBuuiii+Tv71+pwgAAADyNU8GsTZs2Ze4t+zNfX1/deeedeuuttxQQEOB0cQAAAJ7EqU9lLl68WNHR0ZoxY4aysrK0YcMGzZgxQ82bN1dqaqpSUlL0n//8R2PHjnV1vQAAAOctp2bMnnvuOU2dOlXdu3e3L2vVqpUiIyP15JNP6ptvvlGtWrU0fPhwTZkyxWXFAgAAnM+cmjHbvHmzmjRpUm55kyZNtHnzZkknL3ceOHCgctUBAAB4EKdmzGJjYzVp0iTNmDFDfn5+kqQTJ05o0qRJio2NlSTt27dPDRo0cF2lsFs+YLlCQ0PdXQYAAHAxp4LZtGnTdPPNNysyMlKtWrWSzWbTpk2bVFJSoo8//liStGvXLg0cONClxQIAAJzPbJZlWc5s+Ouvv+q9997T9u3bZVmWYmNj1bt3b4WEhLi6Rvy//Px8hYWFKS8vjxkzAABqCEd+fzs1YyZJwcHBSkhIUNOmTXX8+HFJ//tGgJtvvtnZ3QIAAHgsp4LZrl27dOutt2rz5s2y2Wz278g8he/KBAAAcJxTn8ocOnSoLr74Yh08eFBBQUHasmWLVq5cqfj4eKWnp7u4RAAAAM/g1IzZV199pf/85z+qX7++vLy85O3trU6dOmnixIkaMmSINmzY4Oo6AQAAzntOzZiVlJQoODhYklSvXj3t379f0snnmG3bts111QEAAHgQp2bMLrvsMm3atEnNmjXTlVdeqcmTJ8vPz08zZsxQs2bNXF0jAACAR3AqmI0dO1aFhYWSpGeffVZ//etfdc0116hu3bp6//33XVogAACAp3D6OWZ/dvToUYWHh5f5dCZci+eYAQBQ81TLc8z+rE6dOq7aFQAAgEdy6uZ/AAAAuB7BDAAAwBAEMwAAAEO47B4zVJ+EsfPl7R/okn2te/Eel+wHAABUHjNmAAAAhiCYAQAAGIJgBgAAYAiCGQAAgCEIZgAAAIYgmAEAABiCYAYAAGAIghkAAIAhCGYAAACGIJgBAAAYgmAGAABgCIIZAACAIQhmAAAAhiCYAQAAGIJgBgAAYAiCGQAAgCEIZgAAAIYgmAEAABiCYAYAAGAIm2VZlruLwLnJz89XWFiYtiRfqpAAb3eXo4vGbXZ3CQAAGO/U7++8vDyFhoZWOJYZMwAAAEMQzAAAAAxBMAMAADAEwQwAAMAQBDMAAABDGBnMpk+frlatWik0NFShoaHq0KGDPv30U4f2kZSUpMTExHLLs7KyZLPZlJ2dfcZts7OzZbPZlJWV5Vjh56h58+by8/PTvn37qmT/AACgZjIymEVGRmrSpElau3at1q5dq+uuu0633HKLvv32W3eXVmmZmZkqKirSHXfcoXfeecfd5QAAAIMYGcxuuukm9ezZUzExMYqJidFzzz2n4OBgff3115Kk77//XkFBQUpNTbVv8+GHHyogIECbN5v9bK2UlBT17t1b/fr106xZs8Rj5AAAwClGBrM/Kikp0YIFC1RYWKgOHTpIkmJjYzVlyhQNHDhQe/bs0f79+/Xggw9q0qRJiouLq5I6MjIyFBwcXOHr+eefr3AfBQUFWrhwofr27atu3bqpsLBQ6enpZxxfXFys/Pz8Mi8AAHD+8nF3AWeyefNmdejQQUVFRQoODtbixYvVokUL+/qBAwcqLS1N/fr1k5+fn9q3b6+hQ4e65NheXl5l/ilJ8fHxZ73nrE6dOhWuX7BggaKjo9WyZUtJ0l133aWUlBRde+21px0/ceJEjR8/3oHKAQBATWZsMGvevLmysrKUm5urRYsWqX///lq5cmWZcDZr1izFxMTIy8tLW7Zskc1mc8mxg4ODy/xTkgIDAxUVFVWp/aakpKhv377293379lVCQoJyc3NVu3btcuOTk5M1bNgw+/v8/Hw1bty4UjUAAABzGXsp08/PT1FRUYqPj9fEiRPVunVrTZ06tcyYjRs3qrCwUIWFhcrJySmzLjQ0VHl5eeX2m5ubK0kKCws747FPrftjMKvspcytW7dq9erVGjlypHx8fOTj46OrrrpKx44d0/z580+7jb+/v/2TqadeAADg/GXsjNmfWZal4uJi+/ujR48qKSlJY8aMUU5Ojvr06aP169crMDBQ0sn70ObPn6+ioiIFBATYt1uzZo3q16+v8PDwMx7L29u73E35lb2UmZKSooSEBE2bNq3M8rlz5yolJUWPPPJIhfsGAADnP5tl4McC//73v6tHjx5q3LixCgoKtGDBAk2aNElLly5Vt27dJEm9evXS3r17lZmZqePHj6tdu3bq2rWrPfjk5eUpNjZWnTp10qhRoxQeHq6vvvpKQ4YMUXJyskaMGHHG4+/bt09du3bVRx99pNjY2Er3c+LECTVq1EgTJkzQgAEDyqzbsWOHYmJilJWVpdatW1e4n1PfTr8l+VKFBHhXuq7Kumic2Z+ABQDABKd+f+fl5Z316peRlzIPHjyofv36qXnz5uratatWr15dJpTNmTNHaWlpmjt3rnx8fBQUFKR58+Zp5syZSktLk3TycmRGRoYsy1JiYqJat26tyZMn65lnntHw4cMrPP6JEye0bds2FRUVuaSfJUuW6MiRI7r11lvLrYuOjlZcXJxSUlJcciwAAFBzGTljhtNjxgwAgJqnxs+YAQAAeCKCGQAAgCEIZgAAAIaoMY/LwP/cVztEPoEGnLrXO7pkN6sGr3LJfgAAqOmYMQMAADAEwQwAAMAQBDMAAABDEMwAAAAMQTADAAAwBMEMAADAEAQzAAAAQxDMAAAADEEwAwAAMATBDAAAwBAEMwAAAEMQzAAAAAxBMAMAADAEwQwAAMAQBDMAAABDEMwAAAAMQTADAAAwBMEMAADAED7uLgCOWz5guUJDQ91dBgAAcDFmzAAAAAxBMAMAADAEwQwAAMAQBDMAAABDEMwAAAAMQTADAAAwBMEMAADAEAQzAAAAQxDMAAAADEEwAwAAMATBDAAAwBAEMwAAAEMQzAAAAAxBMAMAADCEj7sLgOMSxs6Xt3+gu8sAAOC8su7Fe9xdAjNmAAAApiCYAQAAGIJgBgAAYAiCGQAAgCEIZgAAAIYgmAEAABiCYAYAAGAIghkAAIAhCGYAAACGIJgBAAAYgmAGAABgCIIZAACAIQhmAAAAhiCYAQAAGIJgBgAAYAiCGQAAgCEIZgAAAIYgmAEAABiCYAYAAGAIH3cXAMfNC56qkABvd5cBAICRLhq32d0lOI0ZMwAAAEMQzAAAAAxBMAMAADAEwQwAAMAQBDMAAABDuDWYTZ8+Xa1atVJoaKhCQ0PVoUMHffrppw7tIykpSYmJieWWZ2VlyWazKTs72zXFutBDDz0kb29vLViwwN2lAAAAg7g1mEVGRmrSpElau3at1q5dq+uuu0633HKLvv32W3eWVaV+++03vf/++xoxYoRSUlLcXQ4AADCIW4PZTTfdpJ49eyomJkYxMTF67rnnFBwcrK+//lqS9P333ysoKEipqan2bT788EMFBARo8+aa+YyShQsXqkWLFkpOTtaqVauMnNEDAADuYcw9ZiUlJVqwYIEKCwvVoUMHSVJsbKymTJmigQMHas+ePdq/f78efPBBTZo0SXFxcZU+Zo8ePRQcHFzh64/ONrZHjx5nPWZKSor69u2rsLAw9ezZU7Nnzz7j2OLiYuXn55d5AQCA85fbn/y/efNmdejQQUVFRQoODtbixYvVokUL+/qBAwcqLS1N/fr1k5+fn9q3b6+hQ4e65NgzZ87UsWPHznl8VlZWhesDAwMrXL9jxw59/fXX+vDDDyVJffv21ZAhQ/TUU0/Jy6t8Rp44caLGjx9/zvUBAICaze3BrHnz5srKylJubq4WLVqk/v37a+XKlWXC2axZsxQTEyMvLy9t2bJFNpvNJcdu1KiRQ+OjoqIqdbyUlBR1795d9erVkyT17NlT999/vz7//HPdcMMN5cYnJydr2LBh9vf5+flq3LhxpWoAAADmcnsw8/Pzswee+Ph4rVmzRlOnTtVbb71lH7Nx40YVFhbKy8tLOTk5uvDCC+3rQkNDtWfPnnL7zc3NlSSFhYWd8dg9evRQRkZGhfX9+uuv9j//+dLmn11zzTVn/FRpSUmJ5syZo5ycHPn4+JRZnpKSctpg5u/vL39//wqPCQAAzh9uD2Z/ZlmWiouL7e+PHj2qpKQkjRkzRjk5OerTp4/Wr19vv2wYGxur+fPnq6ioSAEBAfbt1qxZo/r16ys8PPyMx6rOS5lpaWkqKCjQhg0b5O39vy8g//7779WnTx8dOXJEdevWPedaAADA+cdmWZblroP//e9/V48ePdS4cWMVFBRowYIFmjRpkpYuXapu3bpJknr16qW9e/cqMzNTx48fV7t27dS1a1dNmzZNkpSXl6fY2Fh16tRJo0aNUnh4uL766isNGTJEycnJGjFihLvaKyMxMVEBAQHlnl1mWZYaN26sESNGnPXeufz8fIWFhWlL8qUKCfCucCwAAJ7qonFmPbnh1O/vvLw8hYaGVjjWrZ/KPHjwoPr166fmzZura9euWr16dZlQNmfOHKWlpWnu3Lny8fFRUFCQ5s2bp5kzZyotLU3SyUuVGRkZsixLiYmJat26tSZPnqxnnnlGw4cPd2d7dgcPHtQnn3yi22+/vdw6m82m2267jWeaAQAA986YwTHMmAEAcHbMmAEAAKDSCGYAAACGIJgBAAAYwrjHZeDs7qsdIp9ATh0AwCyrBq9ydwk1HjNmAAAAhiCYAQAAGIJgBgAAYAiCGQAAgCEIZgAAAIYgmAEAABiCYAYAAGAIghkAAIAhCGYAAACGIJgBAAAYgmAGAABgCIIZAACAIQhmAAAAhiCYAQAAGIJgBgAAYAiCGQAAgCEIZgAAAIYgmAEAABjCx90FwHHLByxXaGiou8sAAAAuxowZAACAIQhmAAAAhiCYAQAAGIJgBgAAYAiCGQAAgCEIZgAAAIbgcRk1iGVZkqT8/Hw3VwIAAM7Vqd/bp36PV4RgVoMcOXJEktS4cWM3VwIAABxVUFCgsLCwCscQzGqQOnXqSJL27t171hNb0+Xn56tx48b68ccfz+uH6XpKn5Ln9OopfUqe06un9Cl5Tq/V3adlWSooKNCFF1541rEEsxrEy+vkLYFhYWHn9Q/MH4WGhnpEr57Sp+Q5vXpKn5Ln9OopfUqe02t19nmuEyrc/A8AAGAIghkAAIAhCGY1iL+/v5566in5+/u7u5Qq5ym9ekqfkuf06il9Sp7Tq6f0KXlOryb3abPO5bObAAAAqHLMmAEAABiCYAYAAGAIghkAAIAhCGYAAACGIJgZ5h//+IcuvvhiBQQEqH379srIyKhw/MqVK9W+fXsFBASoWbNmevPNN6up0spzpNcDBw6od+/eat68uby8vPTYY49VX6GV5EifH374obp166b69esrNDRUHTp00LJly6qx2spxpNfMzEx17NhRdevWVWBgoGJjY/XKK69UY7XOc/Tn9JRVq1bJx8dHbdq0qdoCXciRXtPT02Wz2cq9vv/++2qs2DmOntPi4mKNGTNGTZo0kb+/vy655BLNmjWrmqqtHEd6TUpKOu05bdmyZTVW7BxHz+m8efPUunVrBQUFKSIiQvfee6/9qxCrlQVjLFiwwPL19bXefvtta+vWrdbQoUOtWrVqWXv27Dnt+F27dllBQUHW0KFDra1bt1pvv/225evra33wwQfVXLnjHO119+7d1pAhQ6x3333XatOmjTV06NDqLdhJjvY5dOhQ64UXXrC++eYba/v27VZycrLl6+trrV+/vpord5yjva5fv95KTU21tmzZYu3evduaO3euFRQUZL311lvVXLljHO3zlNzcXKtZs2bWDTfcYLVu3bp6iq0kR3tdsWKFJcnatm2bdeDAAfvr999/r+bKHePMOb355putK6+80lq+fLm1e/dua/Xq1daqVauqsWrnONprbm5umXP5448/WnXq1LGeeuqp6i3cQY72mZGRYXl5eVlTp061du3aZWVkZFgtW7a0EhMTq7lyyyKYGeSKK66wBgwYUGZZbGysNXr06NOOHzlypBUbG1tm2cMPP2xdddVVVVajqzja6x917ty5xgSzyvR5SosWLazx48e7ujSXc0Wvt956q9W3b19Xl+ZSzvZ55513WmPHjrWeeuqpGhPMHO31VDD75ZdfqqE613G0z08//dQKCwuzjhw5Uh3luVRlf04XL15s2Ww2Kzs7uyrKcxlH+3zxxRetZs2alVn22muvWZGRkVVW45lwKdMQx48f17p163TDDTeUWX7DDTfoyy+/PO02X331Vbnx3bt319q1a3XixIkqq7WynOm1JnJFn6WlpSooKLB/gb2pXNHrhg0b9OWXX6pz585VUaJLONvn7NmztXPnTj311FNVXaLLVOactm3bVhEREeratatWrFhRlWVWmjN9LlmyRPHx8Zo8ebIaNWqkmJgYPfHEEzp27Fh1lOw0V/ycpqSk6Prrr1eTJk2qokSXcKbPq6++Wj/99JPS0tJkWZYOHjyoDz74QDfeeGN1lFwGX2JuiMOHD6ukpEQNGjQos7xBgwbKyck57TY5OTmnHf/777/r8OHDioiIqLJ6K8OZXmsiV/T50ksvqbCwUL169aqKEl2mMr1GRkbq559/1u+//66nn35aDzzwQFWWWinO9Lljxw6NHj1aGRkZ8vGpOX/lOtNrRESEZsyYofbt26u4uFhz585V165dlZ6eroSEhOoo22HO9Llr1y5lZmYqICBAixcv1uHDhzVw4EAdPXrU6PvMKvt30oEDB/Tpp58qNTW1qkp0CWf6vPrqqzVv3jzdeeedKioq0u+//66bb75Zr7/+enWUXEbN+VvCQ9hstjLvLcsqt+xs40+33ESO9lpTOdvn/Pnz9fTTT+tf//qXLrjggqoqz6Wc6TUjI0O//vqrvv76a40ePVpRUVG6++67q7LMSjvXPktKStS7d2+NHz9eMTEx1VWeSzlyTps3b67mzZvb33fo0EE//vijpkyZYmwwO8WRPktLS2Wz2TRv3jyFhYVJkl5++WX97W9/07Rp0xQYGFjl9VaGs38nvfPOO6pdu7YSExOrqDLXcqTPrVu3asiQIRo3bpy6d++uAwcOaMSIERowYIBSUlKqo1w7gpkh6tWrJ29v73Jp/tChQ+VS/ykNGzY87XgfHx/VrVu3ymqtLGd6rYkq0+f777+v+++/XwsXLtT1119flWW6RGV6vfjiiyVJcXFxOnjwoJ5++mljg5mjfRYUFGjt2rXasGGDBg0aJOnkL3XLsuTj46PPPvtM1113XbXU7ihX/ZxeddVVeu+991xdnss402dERIQaNWpkD2WSdOmll8qyLP3000+Kjo6u0pqdVZlzalmWZs2apX79+snPz68qy6w0Z/qcOHGiOnbsqBEjRkiSWrVqpVq1aumaa67Rs88+W61XoLjHzBB+fn5q3769li9fXmb58uXLdfXVV592mw4dOpQb/9lnnyk+Pl6+vr5VVmtlOdNrTeRsn/Pnz1dSUpJSU1Pdcn+DM1x1Ti3LUnFxsavLcxlH+wwNDdXmzZuVlZVlfw0YMEDNmzdXVlaWrrzyyuoq3WGuOqcbNmww9rYKybk+O3bsqP379+vXX3+1L9u+fbu8vLwUGRlZpfVWRmXO6cqVK/XDDz/o/vvvr8oSXcKZPn/77Td5eZWNRN7e3pL+dyWq2lT7xw1wRqc+3puSkmJt3brVeuyxx6xatWrZP/0yevRoq1+/fvbxpx6X8fjjj1tbt261UlJSatzjMs61V8uyrA0bNlgbNmyw2rdvb/Xu3dvasGGD9e2337qj/HPmaJ+pqamWj4+PNW3atDIfUc/NzXVXC+fM0V7feOMNa8mSJdb27dut7du3W7NmzbJCQ0OtMWPGuKuFc+LMf7t/VJM+lelor6+88oq1ePFia/v27daWLVus0aNHW5KsRYsWuauFc+JonwUFBVZkZKT1t7/9zfr222+tlStXWtHR0dYDDzzgrhbOmbP//fbt29e68sorq7tcpzna5+zZsy0fHx/rH//4h7Vz504rMzPTio+Pt6644opqr51gZphp06ZZTZo0sfz8/Kx27dpZK1eutK/r37+/1blz5zLj09PTrbZt21p+fn5W06ZNrenTp1dzxc5ztFdJ5V5NmjSp3qKd4EifnTt3Pm2f/fv3r/7CneBIr6+99prVsmVLKygoyAoNDbXatm1r/eMf/7BKSkrcULljHP1v949qUjCzLMd6feGFF6xLLrnECggIsMLDw61OnTpZn3zyiRuqdpyj5/S7776zrr/+eiswMNCKjIy0hg0bZv3222/VXLVzHO01NzfXCgwMtGbMmFHNlVaOo32+9tprVosWLazAwEArIiLC6tOnj/XTTz9Vc9WWZbOs6p6jAwAAwOlwjxkAAIAhCGYAAACGIJgBAAAYgmAGAABgCIIZAACAIQhmAAAAhiCYAQAAGIJgBgAAYAiCGQBUoS5duuixxx6TJDVt2lSvvvqqW+sBYDYfdxcAAJ5izZo1qlWrlrvLAGAwghkAVJP69eu7uwQAhuNSJgBUkz9fyrTZbJo5c6ZuvfVWBQUFKTo6WkuWLCmzzdatW9WzZ08FBwerQYMG6tevnw4fPlzNlQOoLgQzAHCj8ePHq1evXtq0aZN69uypPn366OjRo5KkAwcOqHPnzmrTpo3Wrl2rpUuX6uDBg+rVq5ebqwZQVQhmAOBGSUlJuvvuuxUVFaXnn39ehYWF+uabbyRJ06dPV7t27fT8888rNjZWbdu21axZs7RixQpt377dzZUDqArcYwYAbtSqVSv7n2vVqqWQkBAdOnRIkrRu3TqtWLFCwcHB5bbbuXOnYmJiqq1OANWDYAYAbuTr61vmvc1mU2lpqSSptLRUN910k1544YVy20VERFRLfQCqF8EMAAzVrl07LVq0SE2bNpWPD39dA56Ae8wAwFCPPvqojh49qrvvvlvffPONdu3apc8++0z33XefSkpK3F0egCpAMAMAQ1144YVatWqVSkpK1L17d1122WUaOnSowsLC5OXFX9/A+chmWZbl7iIAAADAjBkAAIAxCGYAAACGIJgBAAAYgmAGAABgCIIZAACAIQhmAAAAhiCYAQAAGIJgBgAAYAiCGQAAgCEIZgAAAIYgmAEAABji/wCaaq7SyDBBIAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cuser_vs_auto['agreement'] = cuser_vs_auto['UvsA']\n",
    "\n",
    "cuser_vs_auto['sort'] = cuser_vs_auto['agreement'].map(lambda x: 0 if x[0]== \"o\" else int(x[0]))\n",
    "\n",
    "divide_sum = cuser_vs_auto.groupby(['ds']).sum('line')\n",
    "display(divide_sum)\n",
    "for ds in divide_sum.index:\n",
    "    cuser_vs_auto.loc[cuser_vs_auto['ds'] == ds, 'line'] /= divide_sum.loc[ds,'line']\n",
    "\n",
    "\n",
    "#cuser_vs_auto\n",
    "sns.barplot(cuser_vs_auto.sort_values('sort', ascending=True), y='agreement', x='line', hue='ds')\n",
    "\n",
    "\n",
    "legend = plt.legend(\n",
    "    bbox_to_anchor=(0.5, 1.20),\n",
    "    loc=\"lower center\",\n",
    "    borderaxespad=0,\n",
    "    frameon=False,\n",
    "    ncol=3,\n",
    ")\n",
    "export_legend(legend, \"legend_\" + 'automatic_vs_users.svg' )\n",
    "plt.savefig('automatic_vs_users.svg')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "6f83927d-6792-4c3b-9554-803333b9dd86",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "a = G.reset_index().groupby(['id', 'consensus']).count()\n",
    "b = G.reset_index().groupby(['id']).count()\n",
    "\n",
    "r = pd.DataFrame(a['index']/b['index']).reset_index()\n",
    "\n",
    "r['ds'] = r['id'].map(lambda x: x.split(\"_\")[0])\n",
    "#display(r.groupby(['ds','consensus']).median())\n",
    "\n",
    "short_names = {\"bug(fix)\":\"B\", \"bug(fix) + refactoring\":\"B+R\", \"documentation\":\"D\", \"refactoring\":\"R\", \"test\":\"T\", \"test + refactoring\":\"T+R\"}\n",
    "r['Consensus']= r['consensus'].map(short_names)\n",
    "sns.boxenplot(r, y='Consensus', x='index', hue='ds')\n",
    "#r.groupby(['ds', 'consensus']).agg(median_bugfix=('index', 'median'), mean_bugfix=('index', 'mean'), std_bugfix=('index', 'std'))\n",
    "\n",
    "plt.savefig(\"distribution_of_labels_in_fixes.svg\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "8ae3ac7d-cf4f-4b73-9bc3-f870074921bc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "bug(fix) bugs-in-py BootstrapResult(confidence_interval=ConfidenceInterval(low=0.21428571428571427, high=0.26969696969696966), bootstrap_distribution=array([0.22886905, 0.25      , 0.25      , ..., 0.25      , 0.21702899,\n",
      "       0.25      ]), standard_error=0.01136057953047476)\n",
      "bug(fix) crawl BootstrapResult(confidence_interval=ConfidenceInterval(low=0.20638568588469186, high=0.39803370786516856), bootstrap_distribution=array([0.36738956, 0.27443182, 0.26101514, ..., 0.31109403, 0.35369909,\n",
      "       0.30170316]), standard_error=0.041637756270771195)\n",
      "bug(fix) cve BootstrapResult(confidence_interval=ConfidenceInterval(low=0.3, high=0.5), bootstrap_distribution=array([0.34351679, 0.38863636, 0.4037037 , ..., 0.43910256, 0.38867925,\n",
      "       0.34409722]), standard_error=0.03768749059994093)\n",
      "bug(fix) + refactoring bugs-in-py BootstrapResult(confidence_interval=ConfidenceInterval(low=0.05714285714285714, high=0.14285714285714285), bootstrap_distribution=array([0.064     , 0.11320755, 0.125     , ..., 0.14285714, 0.11428571,\n",
      "       0.10447761]), standard_error=0.01747439785206581)\n",
      "bug(fix) + refactoring crawl BootstrapResult(confidence_interval=ConfidenceInterval(low=0.008403361344537815, high=0.06997084548104957), bootstrap_distribution=array([0.03125   , 0.03293413, 0.01666667, ..., 0.03389831, 0.00840336,\n",
      "       0.04188482]), standard_error=0.011429719170530725)\n",
      "bug(fix) + refactoring cve BootstrapResult(confidence_interval=ConfidenceInterval(low=0.014814814814814815, high=0.0718562874251497), bootstrap_distribution=array([0.03076923, 0.03076923, 0.02700546, ..., 0.02159969, 0.03076923,\n",
      "       0.02991453]), standard_error=0.008201251231421947)\n",
      "documentation bugs-in-py BootstrapResult(confidence_interval=ConfidenceInterval(low=0.1111111111111111, high=0.1468013468013468), bootstrap_distribution=array([0.125     , 0.13333333, 0.13771552, ..., 0.125     , 0.12973352,\n",
      "       0.13157895]), standard_error=0.00576847682007412)\n",
      "documentation crawl BootstrapResult(confidence_interval=ConfidenceInterval(low=0.16216216216216217, high=0.375609756097561), bootstrap_distribution=array([0.28169014, 0.28169014, 0.29752066, ..., 0.30612245, 0.29752066,\n",
      "       0.3       ]), standard_error=0.04633228205187024)\n",
      "documentation cve BootstrapResult(confidence_interval=ConfidenceInterval(low=0.13333333333333333, high=0.2222222222222222), bootstrap_distribution=array([0.15306122, 0.18181818, 0.14814815, ..., 0.15306122, 0.17857143,\n",
      "       0.19512195]), standard_error=0.0180167268944615)\n",
      "refactoring bugs-in-py BootstrapResult(confidence_interval=ConfidenceInterval(low=0.041666666666666664, high=0.14285714285714285), bootstrap_distribution=array([0.06410256, 0.0625    , 0.07738095, ..., 0.05718954, 0.05131579,\n",
      "       0.06330128]), standard_error=0.013188289389171331)\n",
      "refactoring crawl BootstrapResult(confidence_interval=ConfidenceInterval(low=0.008969541139240506, high=0.06740196078431374), bootstrap_distribution=array([0.03719008, 0.03773585, 0.03746297, ..., 0.03436426, 0.01452728,\n",
      "       0.03773585]), standard_error=0.01211168420068428)\n",
      "refactoring cve BootstrapResult(confidence_interval=ConfidenceInterval(low=0.01282051282051282, high=0.12453318385878136), bootstrap_distribution=array([0.04641034, 0.0456621 , 0.02632784, ..., 0.03238715, 0.0456621 ,\n",
      "       0.02884615]), standard_error=0.020819460446969917)\n",
      "test bugs-in-py BootstrapResult(confidence_interval=ConfidenceInterval(low=nan, high=nan), bootstrap_distribution=array([0.57142857, 0.6       , 0.6       , ..., 0.56156156, 0.58333333,\n",
      "       0.58333333]), standard_error=0.01911918682460763)\n",
      "test crawl BootstrapResult(confidence_interval=ConfidenceInterval(low=0.312280701754386, high=0.5085324232081911), bootstrap_distribution=array([0.40860215, 0.36619718, 0.52481305, ..., 0.35502959, 0.40721649,\n",
      "       0.4491018 ]), standard_error=0.036424937000445474)\n",
      "test cve BootstrapResult(confidence_interval=ConfidenceInterval(low=0.325, high=0.5543478260869565), bootstrap_distribution=array([0.428     , 0.46086957, 0.42401501, ..., 0.44444444, 0.46086957,\n",
      "       0.45360825]), standard_error=0.04959788630608252)\n",
      "test + refactoring bugs-in-py BootstrapResult(confidence_interval=ConfidenceInterval(low=0.0425531914893617, high=0.31746031746031744), bootstrap_distribution=array([0.05747126, 0.23529412, 0.08536585, ..., 0.09090909, 0.09090909,\n",
      "       0.10344828]), standard_error=0.03904468439167388)\n",
      "test + refactoring crawl BootstrapResult(confidence_interval=ConfidenceInterval(low=0.005917159763313609, high=0.0975609756097561), bootstrap_distribution=array([0.04201259, 0.03325751, 0.05313095, ..., 0.05313095, 0.05309735,\n",
      "       0.0222615 ]), standard_error=0.013820937631667028)\n",
      "test + refactoring cve BootstrapResult(confidence_interval=ConfidenceInterval(low=0.007032348804500703, high=0.1744186046511628), bootstrap_distribution=array([0.1       , 0.00703235, 0.0952381 , ..., 0.00950119, 0.13043478,\n",
      "       0.13043478]), standard_error=0.04259429697747309)\n",
      "other bugs-in-py BootstrapResult(confidence_interval=ConfidenceInterval(low=0.05223880597014925, high=0.4810126582278481), bootstrap_distribution=array([0.08      , 0.05223881, 0.08      , ..., 0.08      , 0.08      ,\n",
      "       0.48101266]), standard_error=0.18026096895052202)\n",
      "other crawl BootstrapResult(confidence_interval=ConfidenceInterval(low=0.006243645144890696, high=0.625), bootstrap_distribution=array([0.25      , 0.03935618, 0.03935618, ..., 0.06030151, 0.01841085,\n",
      "       0.06030151]), standard_error=0.08173629638938901)\n",
      "other cve BootstrapResult(confidence_interval=ConfidenceInterval(low=0.0091324200913242, high=0.28289473684210525), bootstrap_distribution=array([0.01157556, 0.01787891, 0.0876494 , ..., 0.04015152, 0.00360834,\n",
      "       0.02216086]), standard_error=0.054236331336933535)\n"
     ]
    }
   ],
   "source": [
    "\n",
    "types_of_lines_in_commit = list(r['consensus'].unique())\n",
    "DS = list(r['ds'].unique())\n",
    "\n",
    "for t in types_of_lines_in_commit:\n",
    "    for d in DS:\n",
    "        filter = (r['consensus']==t) & (r['ds']==d)\n",
    "        print(t, d, bootstrap( ( r[filter]['index'],), np.median, confidence_level=0.99))\n",
    "\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8a01aa26-af7d-4837-98e2-4e6665b6e034",
   "metadata": {},
   "source": [
    "# Table"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "9974e57a-877d-45fe-a314-79b73e8b67f6",
   "metadata": {},
   "outputs": [],
   "source": [
    "tmp = flow_auto.reset_index()\n",
    "tmp['user_x'].unique()\n",
    "#tmp['match'] = tmp['annotation_y'] == tmp['annotation_x']\n",
    "tmp['type'] = tmp['user_x'].map(lambda x: x[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "227e0709-54e9-41d6-b7c9-d8480777ea35",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_232579/2837343866.py:20: FutureWarning: The default value of numeric_only in DataFrameGroupBy.sum 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",
      "  diag_bugs = tmp.groupby(['ds', 'id', 'file', 'image', 'line']).count().reset_index().groupby(['ds']).sum().reset_index()\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>Dataset</th>\n",
       "      <th>Bugs</th>\n",
       "      <th>Files</th>\n",
       "      <th>Lines</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>bugs-in-py</td>\n",
       "      <td>496</td>\n",
       "      <td>1382</td>\n",
       "      <td>20221</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>crawl</td>\n",
       "      <td>145</td>\n",
       "      <td>759</td>\n",
       "      <td>26102</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>cve</td>\n",
       "      <td>152</td>\n",
       "      <td>601</td>\n",
       "      <td>21640</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      Dataset  Bugs  Files  Lines\n",
       "0  bugs-in-py   496   1382  20221\n",
       "1       crawl   145    759  26102\n",
       "2         cve   152    601  21640"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Sankey whole data\n",
    "diag_ret = []\n",
    "\n",
    "diag_bugs = tmp.groupby(['ds', 'id']).agg(Files = ('id','count')).reset_index().groupby(['ds']).count().reset_index()\n",
    "\n",
    "\n",
    "diag_bugs = diag_bugs[['ds', 'id']]\n",
    "diag_bugs.columns = ['Dataset', 'Bugs']\n",
    "\n",
    "diag_ret.append(diag_bugs)\n",
    "\n",
    "\n",
    "diag_bugs = tmp.groupby(['ds', 'id', 'file']).count().reset_index().groupby(['ds']).count().reset_index()\n",
    "diag_bugs = diag_bugs[['ds', 'file',]]\n",
    "diag_bugs.columns = ['ds', 'Files']\n",
    "\n",
    "diag_ret.append(diag_bugs)\n",
    "\n",
    "\n",
    "diag_bugs = tmp.groupby(['ds', 'id', 'file', 'image', 'line']).count().reset_index().groupby(['ds']).sum().reset_index()\n",
    "diag_bugs = diag_bugs[['ds', 'line',]]\n",
    "diag_bugs.columns = ['ds', 'Lines']\n",
    "\n",
    "#diag_ret.append(diag_bugs)\n",
    "\n",
    "df_inter_rater['ds'] = df_inter_rater['id'].map(lambda x: x.split(\"_\")[0])\n",
    "diag_bugs = df_inter_rater.groupby('ds').count()[['line']].reset_index()\n",
    "diag_bugs = diag_bugs[['ds', 'line',]]\n",
    "diag_bugs.columns = ['ds', 'Lines']\n",
    "\n",
    "\n",
    "diag_ret.append(diag_bugs)\n",
    "\n",
    "summary = pd.concat(diag_ret, axis=1)[['Dataset','Bugs', 'Files', 'Lines']]\n",
    "\n",
    "#diag_bugs\n",
    "summary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "b3926efc-ba42-4395-bedd-25f03d39608a",
   "metadata": {},
   "outputs": [],
   "source": [
    "stats = df[~df['auto']].groupby(['ds', 'annotation']).count()['id'].reset_index().pivot(index='ds', columns='annotation')\n",
    "stats_org = stats.copy()\n",
    "S = stats.sum(axis=1)\n",
    "for c in stats.columns:\n",
    "    stats[c] /= S \n",
    "    stats[c] *= 100\n",
    "    #stats['S']=S\n",
    "\n",
    "#stats_style = stats.style.format('{:.2f}\\%', decimal='.', precision=2)\n",
    "stats.columns = [c[1] for c in stats.columns]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "8cbeb1f2-9f58-4558-9e23-1a81885d3494",
   "metadata": {},
   "outputs": [],
   "source": [
    "# G['ds'] = G['id'].map(lambda x: x.split(\"_\")[0])\n",
    "# stats = G.groupby(['ds']).count()\n",
    "# #S = stats.groupby('ds').sum()\n",
    "# #S\n",
    "# G"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "dacb2492-905d-49ee-9b3f-c9b45e12e949",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "18dbbe96-b3ad-493c-81be-8ec605490739",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cf5f1d8e-df3f-476b-91a9-76151a52e657",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "c939dde4-6110-4ec9-98f6-bb148c6ed6d4",
   "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>2xU != A</th>\n",
       "      <th>2xU == A</th>\n",
       "      <th>3xU != A</th>\n",
       "      <th>3xU == A</th>\n",
       "      <th>other</th>\n",
       "      <th>3xU</th>\n",
       "      <th>2|3xU</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ds</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>bugs-in-py</th>\n",
       "      <td>706</td>\n",
       "      <td>2133</td>\n",
       "      <td>693</td>\n",
       "      <td>16331</td>\n",
       "      <td>358</td>\n",
       "      <td>84.189704</td>\n",
       "      <td>97.822282</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>crawl</th>\n",
       "      <td>911</td>\n",
       "      <td>2942</td>\n",
       "      <td>1523</td>\n",
       "      <td>20458</td>\n",
       "      <td>268</td>\n",
       "      <td>84.211938</td>\n",
       "      <td>98.654972</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cve</th>\n",
       "      <td>1578</td>\n",
       "      <td>2172</td>\n",
       "      <td>1943</td>\n",
       "      <td>15702</td>\n",
       "      <td>245</td>\n",
       "      <td>81.538817</td>\n",
       "      <td>98.496705</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            2xU != A  2xU == A  3xU != A  3xU == A  other        3xU  \\\n",
       "ds                                                                     \n",
       "bugs-in-py       706      2133       693     16331    358  84.189704   \n",
       "crawl            911      2942      1523     20458    268  84.211938   \n",
       "cve             1578      2172      1943     15702    245  81.538817   \n",
       "\n",
       "                2|3xU  \n",
       "ds                     \n",
       "bugs-in-py  97.822282  \n",
       "crawl       98.654972  \n",
       "cve         98.496705  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "filter = cuser_vs_auto['UvsA'].map(lambda x: x.split(\"x\")[0] == \"3\")\n",
    "cuser_vs_auto[\"UvsA\"] == '3xU == A'\n",
    "\n",
    "#display(cuser_vs_auto)\n",
    "B_copy = cuser_vs_auto[['ds','UvsA', 'annotation']].pivot(index='ds', columns='UvsA')\n",
    "\n",
    "B_copy.columns = [pair[1] for pair in B_copy.columns]\n",
    "B_copy['3xU'] = (B_copy['3xU == A'] + B_copy['3xU != A'])/B_copy.sum(axis=1) * 100\n",
    "#B_copy['2xU'] = (B_copy['2xU == A'] + B_copy['2xU != A'])/B_copy.sum(axis=1)  * 100\n",
    "B_copy['2|3xU'] = (B_copy['2xU == A'] + B_copy['2xU != A']\n",
    "                   + B_copy['3xU == A'] + B_copy['3xU != A'])/B_copy.sum(axis=1)  * 100\n",
    "\n",
    "display(B_copy)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "e685b508-80a1-431b-8e0e-b4a454887abf",
   "metadata": {},
   "outputs": [],
   "source": [
    "table_paper = summary.merge(stats, left_on='Dataset', right_index=True).merge(B_copy[['3xU', '2|3xU']], right_index=True, left_on='Dataset',)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "fefbde52-6cee-47da-ba59-d90f874357af",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\begin{tabular}{llrrrrrrrrrrrr}\n",
      "\\toprule\n",
      "{} &     Dataset &  Bugs &  Files &  Lines &  bug(fix) &  bug(fix) + refactoring &  documentation &  other &  refactoring &  test &  test + refactoring &   3xU &  2|3xU \\\\\n",
      "\\midrule\n",
      "0 &  bugs-in-py &   496 &   1382 &  20221 &     24.7\\% &                    2.6\\% &          13.2\\% &   0.5\\% &         1.9\\% & 55.2\\% &                1.9\\% & 84.2\\% &  97.8\\% \\\\\n",
      "1 &       crawl &   145 &    759 &  26102 &     26.1\\% &                    2.1\\% &          29.2\\% &   1.2\\% &         1.5\\% & 38.8\\% &                1.2\\% & 84.2\\% &  98.7\\% \\\\\n",
      "2 &         cve &   152 &    601 &  21640 &     29.6\\% &                    2.2\\% &          14.2\\% &   1.2\\% &         9.5\\% & 42.0\\% &                1.3\\% & 81.5\\% &  98.5\\% \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_232579/4270149668.py:1: 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",
      "  print(table_paper.to_latex(float_format=\"{:.1f}%\".format))\n"
     ]
    }
   ],
   "source": [
    "print(table_paper.to_latex(float_format=\"{:.1f}%\".format))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d9a7d7cd-12d7-4e35-84c5-bfd66c37a3f9",
   "metadata": {},
   "source": [
    "# Stat test"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "db1f2e64-311a-4149-8226-bfdbc68c3c7e",
   "metadata": {},
   "source": [
    "## Datasets are diffrent "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "1c97937e-d9df-4537-8a23-7dfd48ceaf12",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Chi2ContingencyResult(statistic=15266.30789785402, pvalue=0.0, dof=12, expected_freq=array([[16248.03362519,  1394.72823192, 11906.20210433,   597.27325436,\n",
       "         2526.2160106 , 27099.78668236,   866.76009124],\n",
       "       [20981.85196086,  1801.07833125, 15375.04018836,   771.28711647,\n",
       "         3262.2218527 , 34995.23237437,  1119.288176  ],\n",
       "       [17395.11441395,  1493.19343683, 12746.75770731,   639.43962917,\n",
       "         2704.56213671, 29012.98094327,   927.95173276]]))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "Chi2ContingencyResult(statistic=8187.467892758226, pvalue=0.0, dof=6, expected_freq=array([[17969.42231159,  1410.12190524, 14520.04021616,   769.65187885,\n",
       "         3334.70487202, 26124.19672406,   791.86209208],\n",
       "       [21674.57768841,  1700.87809476, 17513.95978384,   928.34812115,\n",
       "         4022.29512798, 31510.80327594,   955.13790792]]))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "Chi2ContingencyResult(statistic=6321.002413574595, pvalue=0.0, dof=6, expected_freq=array([[15452.48172298,  1410.96035842, 13457.14897981,   531.12859765,\n",
       "         1021.66971823, 27856.10199719,   909.50862572],\n",
       "       [19954.51827702,  1822.03964158, 17377.85102019,   685.87140235,\n",
       "         1319.33028177, 35971.89800281,  1174.49137428]]))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "Chi2ContingencyResult(statistic=4823.718012957028, pvalue=0.0, dof=6, expected_freq=array([[16516.48357346,  1465.27708886,  8300.50010752,   531.73041359,\n",
       "         3519.75590758, 29340.79729052,   964.45561847],\n",
       "       [17682.51642654,  1568.72291114,  8886.49989248,   569.26958641,\n",
       "         3768.24409242, 31412.20270948,  1032.54438153]]))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "S = stats_org.sum(axis=1)\n",
    "\n",
    "display(chi2_contingency(stats_org))\n",
    "display(chi2_contingency(stats_org.loc[['cve', 'crawl']]))\n",
    "display(chi2_contingency(stats_org.loc[['bugs-in-py', 'crawl']]))\n",
    "display(chi2_contingency(stats_org.loc[['bugs-in-py', 'cve']]))\n",
    "\n",
    "#display(chisquare(f_obs=stats_org.loc['cve'], f_exp=(stats_org.loc['crawl'] / (100 *  S['crawl'])).round(3)))\n",
    "\n",
    "#display(chisquare(f_obs=stats.loc['cve'] , f_exp=(stats.loc['bugs-in-py'] / (100 * S['bugs-in-py']).round())))\n",
    "\n",
    "#display(chisquare(f_obs=stats.loc['crawl'] , f_exp=(stats.loc['bugs-in-py'] / (100 * S['bugs-in-py']).round())))\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "29929992-8fc9-4182-9697-81823b829cc2",
   "metadata": {},
   "source": [
    "## Fleiss agremment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "dbe4839a-81a5-48a6-9901-8ac052679d1b",
   "metadata": {},
   "outputs": [],
   "source": [
    "def fleiss_data_row(row):\n",
    "    ret = {k:0 for k in possible_line_annotations}\n",
    "    val = None\n",
    "    for u in users_ids:\n",
    "        u_val = row['annotation_' + u]\n",
    "        if not pd.isnull(u_val):\n",
    "            ret[u_val] += 1\n",
    "            val = u_val\n",
    "            \n",
    "    if sum(ret.values()) < 3:\n",
    "        print(\"error\")\n",
    "\n",
    "    return ret"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e4872608-41c2-4f43-8c6d-9a77c49ed741",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "4d4233b3-616d-4bc4-90b0-44d67b862252",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_inter_rater_format = pd.DataFrame(df_inter_rater.apply(fleiss_data_row, axis=1).tolist())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "a340d2f7-d8a4-4c59-b2c8-2df91e1d33af",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "all datasets 0.8323339712332565\n",
      "CVE datasets 0.8203142050317874\n",
      "CRAWL datasets 0.8437777759543083\n",
      "BIP datasets 0.819130695165318\n"
     ]
    }
   ],
   "source": [
    "# all user agreement\n",
    "print(\"all datasets\", fleiss_kappa(df_inter_rater_format.to_numpy(), method='fleiss'))\n",
    "\n",
    "# CVE user agreement\n",
    "filter = (df_inter_rater['ds'] == 'cve')\n",
    "print(\"CVE datasets\", fleiss_kappa(df_inter_rater_format[filter].to_numpy(), method='fleiss'))\n",
    "\n",
    "# CRAWL user agreement\n",
    "filter = (df_inter_rater['ds'] == 'crawl')\n",
    "print(\"CRAWL datasets\", fleiss_kappa(df_inter_rater_format[filter].to_numpy(), method='fleiss'))\n",
    "\n",
    "# CRAWL user agreement\n",
    "filter = (df_inter_rater['ds'] == 'bugs-in-py')\n",
    "print(\"BIP datasets\", fleiss_kappa(df_inter_rater_format[filter].to_numpy(), method='fleiss'))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "94023408-9f4c-4fcb-ba7e-971ab981ab93",
   "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>id</th>\n",
       "      <th>file</th>\n",
       "      <th>fileid</th>\n",
       "      <th>image</th>\n",
       "      <th>line</th>\n",
       "      <th>annotation_U1</th>\n",
       "      <th>annotation_U2</th>\n",
       "      <th>annotation_U3</th>\n",
       "      <th>annotation_E1</th>\n",
       "      <th>annotation_E2</th>\n",
       "      <th>annotation_E3</th>\n",
       "      <th>ds</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>103</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>cve</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>137</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>cve</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>103</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>cve</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>cve_CVE-2020-10289</td>\n",
       "      <td>actionlib_tools/scripts/library.py</td>\n",
       "      <td>0</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>137</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>cve</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>cve_CVE-2016-10516</td>\n",
       "      <td>werkzeug/debug/tbtools.py</td>\n",
       "      <td>0</td>\n",
       "      <td>afterChange</td>\n",
       "      <td>353</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>bug(fix) + refactoring</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bug(fix)</td>\n",
       "      <td>NaN</td>\n",
       "      <td>cve</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67958</th>\n",
       "      <td>bugs-in-py_pandas-88</td>\n",
       "      <td>pandas/tests/reshape/test_pivot.py</td>\n",
       "      <td>0</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>900</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bugs-in-py</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67959</th>\n",
       "      <td>bugs-in-py_pandas-88</td>\n",
       "      <td>pandas/tests/reshape/test_pivot.py</td>\n",
       "      <td>0</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>901</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bugs-in-py</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67960</th>\n",
       "      <td>bugs-in-py_pandas-88</td>\n",
       "      <td>pandas/tests/reshape/test_pivot.py</td>\n",
       "      <td>0</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>902</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bugs-in-py</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67961</th>\n",
       "      <td>bugs-in-py_pandas-88</td>\n",
       "      <td>pandas/tests/reshape/test_pivot.py</td>\n",
       "      <td>0</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>903</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bugs-in-py</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67962</th>\n",
       "      <td>bugs-in-py_pandas-88</td>\n",
       "      <td>pandas/tests/reshape/test_pivot.py</td>\n",
       "      <td>0</td>\n",
       "      <td>beforeChange</td>\n",
       "      <td>904</td>\n",
       "      <td>NaN</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>test</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>bugs-in-py</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>67963 rows × 12 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                         id                                file  fileid  \\\n",
       "0        cve_CVE-2020-10289  actionlib_tools/scripts/library.py       0   \n",
       "1        cve_CVE-2020-10289  actionlib_tools/scripts/library.py       0   \n",
       "2        cve_CVE-2020-10289  actionlib_tools/scripts/library.py       0   \n",
       "3        cve_CVE-2020-10289  actionlib_tools/scripts/library.py       0   \n",
       "4        cve_CVE-2016-10516           werkzeug/debug/tbtools.py       0   \n",
       "...                     ...                                 ...     ...   \n",
       "67958  bugs-in-py_pandas-88  pandas/tests/reshape/test_pivot.py       0   \n",
       "67959  bugs-in-py_pandas-88  pandas/tests/reshape/test_pivot.py       0   \n",
       "67960  bugs-in-py_pandas-88  pandas/tests/reshape/test_pivot.py       0   \n",
       "67961  bugs-in-py_pandas-88  pandas/tests/reshape/test_pivot.py       0   \n",
       "67962  bugs-in-py_pandas-88  pandas/tests/reshape/test_pivot.py       0   \n",
       "\n",
       "              image  line annotation_U1           annotation_U2 annotation_U3  \\\n",
       "0       afterChange   103      bug(fix)                bug(fix)      bug(fix)   \n",
       "1       afterChange   137      bug(fix)                bug(fix)      bug(fix)   \n",
       "2      beforeChange   103      bug(fix)                bug(fix)      bug(fix)   \n",
       "3      beforeChange   137      bug(fix)                bug(fix)      bug(fix)   \n",
       "4       afterChange   353      bug(fix)  bug(fix) + refactoring           NaN   \n",
       "...             ...   ...           ...                     ...           ...   \n",
       "67958  beforeChange   900           NaN                    test          test   \n",
       "67959  beforeChange   901           NaN                    test          test   \n",
       "67960  beforeChange   902           NaN                    test          test   \n",
       "67961  beforeChange   903           NaN                    test          test   \n",
       "67962  beforeChange   904           NaN                    test          test   \n",
       "\n",
       "      annotation_E1 annotation_E2 annotation_E3          ds  \n",
       "0               NaN           NaN           NaN         cve  \n",
       "1               NaN           NaN           NaN         cve  \n",
       "2               NaN           NaN           NaN         cve  \n",
       "3               NaN           NaN           NaN         cve  \n",
       "4               NaN      bug(fix)           NaN         cve  \n",
       "...             ...           ...           ...         ...  \n",
       "67958          test           NaN           NaN  bugs-in-py  \n",
       "67959          test           NaN           NaN  bugs-in-py  \n",
       "67960          test           NaN           NaN  bugs-in-py  \n",
       "67961          test           NaN           NaN  bugs-in-py  \n",
       "67962          test           NaN           NaN  bugs-in-py  \n",
       "\n",
       "[67963 rows x 12 columns]"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_inter_rater"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "912245f3-f9d9-4fc0-aec5-a0a6b600c49e",
   "metadata": {},
   "outputs": [],
   "source": [
    "# consensus vs auto agreement\n",
    "tmp = pd.DataFrame(df_inter_rater.apply(consensus, axis=1).tolist())\n",
    "G = df_inter_rater.copy()\n",
    "G['consensus']= tmp\n",
    "\n",
    "Z = df[df['auto'] == True].groupby(cols).count().reset_index()\n",
    "R = Z.merge(G, left_on=colsj, right_on=colsj)\n",
    "#R = R[['consensus', 'annotation']]\n",
    "\n",
    "\n",
    "#df_inter_rater_format_for_auto = aggregate_raters(R.to_numpy())\n",
    "#fleiss_kappa(df_inter_rater_format_for_auto[0], method='fleiss')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "a1e3d706-65fd-4f13-8344-67d894e495a8",
   "metadata": {},
   "outputs": [],
   "source": [
    "R['A=C'] = R['consensus'] == R['annotation']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "178c8b8d-391b-449b-8f69-0d080d0e8d6a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "8c2da240-3c25-4c73-b781-16115f6492b3",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "matching_fix = R.groupby(['id']).aggregate(C=('A=C', 'count'), S=('A=C', 'sum')).reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "4057fdbd-173c-4fc1-9222-89af1edd7636",
   "metadata": {},
   "outputs": [],
   "source": [
    "matching_fix['%'] = matching_fix['S']/matching_fix['C']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "8c31a153-9d95-4d9f-8ed1-58ae308b9805",
   "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>id</th>\n",
       "      <th>C</th>\n",
       "      <th>S</th>\n",
       "      <th>%</th>\n",
       "      <th>ds</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>bugs-in-py_PySnooper-1</td>\n",
       "      <td>60</td>\n",
       "      <td>56</td>\n",
       "      <td>0.933333</td>\n",
       "      <td>bugs-in-py</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>bugs-in-py_PySnooper-2</td>\n",
       "      <td>27</td>\n",
       "      <td>27</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>bugs-in-py</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>bugs-in-py_PySnooper-3</td>\n",
       "      <td>30</td>\n",
       "      <td>30</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>bugs-in-py</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>bugs-in-py_ansible-1</td>\n",
       "      <td>39</td>\n",
       "      <td>39</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>bugs-in-py</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>bugs-in-py_ansible-10</td>\n",
       "      <td>24</td>\n",
       "      <td>22</td>\n",
       "      <td>0.916667</td>\n",
       "      <td>bugs-in-py</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>788</th>\n",
       "      <td>cve_CVE-2022-29217</td>\n",
       "      <td>214</td>\n",
       "      <td>211</td>\n",
       "      <td>0.985981</td>\n",
       "      <td>cve</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>789</th>\n",
       "      <td>cve_CVE-2022-29255</td>\n",
       "      <td>65</td>\n",
       "      <td>57</td>\n",
       "      <td>0.876923</td>\n",
       "      <td>cve</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>790</th>\n",
       "      <td>cve_CVE-2022-31015</td>\n",
       "      <td>27</td>\n",
       "      <td>26</td>\n",
       "      <td>0.962963</td>\n",
       "      <td>cve</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>791</th>\n",
       "      <td>cve_CVE-2022-31116</td>\n",
       "      <td>95</td>\n",
       "      <td>78</td>\n",
       "      <td>0.821053</td>\n",
       "      <td>cve</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>792</th>\n",
       "      <td>cve_CVE-2022-31124</td>\n",
       "      <td>422</td>\n",
       "      <td>410</td>\n",
       "      <td>0.971564</td>\n",
       "      <td>cve</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>793 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                         id    C    S         %          ds\n",
       "0    bugs-in-py_PySnooper-1   60   56  0.933333  bugs-in-py\n",
       "1    bugs-in-py_PySnooper-2   27   27  1.000000  bugs-in-py\n",
       "2    bugs-in-py_PySnooper-3   30   30  1.000000  bugs-in-py\n",
       "3      bugs-in-py_ansible-1   39   39  1.000000  bugs-in-py\n",
       "4     bugs-in-py_ansible-10   24   22  0.916667  bugs-in-py\n",
       "..                      ...  ...  ...       ...         ...\n",
       "788      cve_CVE-2022-29217  214  211  0.985981         cve\n",
       "789      cve_CVE-2022-29255   65   57  0.876923         cve\n",
       "790      cve_CVE-2022-31015   27   26  0.962963         cve\n",
       "791      cve_CVE-2022-31116   95   78  0.821053         cve\n",
       "792      cve_CVE-2022-31124  422  410  0.971564         cve\n",
       "\n",
       "[793 rows x 5 columns]"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "matching_fix['ds'] = matching_fix['id'].map(lambda x: x.split(\"_\")[0])\n",
    "matching_fix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "3eb8f9d5-7ca9-42e4-a2e5-55737ea49e5e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['bugs-in-py', 'crawl', 'cve']"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "bugs-in-py BootstrapResult(confidence_interval=ConfidenceInterval(low=1.0, high=1.0), bootstrap_distribution=array([1., 1., 1., ..., 1., 1., 1.]), standard_error=0.0)\n",
      "0.18461538461538463\n",
      "crawl BootstrapResult(confidence_interval=ConfidenceInterval(low=0.9523241092421776, high=0.9804518055320828), bootstrap_distribution=array([0.95294118, 0.96428571, 0.97849462, ..., 0.94954128, 0.96907216,\n",
      "       0.97849462]), standard_error=0.00993706493674338)\n",
      "0.0\n",
      "cve BootstrapResult(confidence_interval=ConfidenceInterval(low=0.9446079911433172, high=0.9762801769380716), bootstrap_distribution=array([0.95023697, 0.96296296, 0.96138939, ..., 0.97222222, 0.97186701,\n",
      "       0.97391304]), standard_error=0.010946455124172345)\n",
      "0.21568627450980393\n"
     ]
    }
   ],
   "source": [
    "display(DS)\n",
    "for d in DS:\n",
    "    filter = (matching_fix['ds']==d)\n",
    "    #display(matching_fix[filter]['%'],)\n",
    "    #print(d)\n",
    "    print(d, bootstrap( (matching_fix[filter]['%'],), np.median, confidence_level=0.90, method='basic'))\n",
    "    print(matching_fix[filter]['%'].min())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f34dbf4d-7893-482a-8b0f-adb38e4a855c",
   "metadata": {},
   "outputs": [],
   "source": [
    "matching_fix['ds']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0b800a5c-9292-4820-bb4b-7ed97f1b6e9d",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Fleiss kappa for annotatons when consensus != automation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "49f0b3c6-4652-4f18-85f6-4fa1dd535ccb",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "R_inter_rater_format = pd.DataFrame(R[R['annotation'] != R['consensus']].apply(fleiss_data_row, axis=1).tolist())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "78e3c0cd-8893-47be-800c-2b2bc3dae8af",
   "metadata": {},
   "outputs": [],
   "source": [
    "print(\"all datasets\", fleiss_kappa(R_inter_rater_format.to_numpy(), method='fleiss'))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4e496ade-3731-4196-82bb-478cd6eb8d44",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.11.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}