RQ3_ablation.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "\n",
    "df = pd.read_csv('labelling/text_quality_labelling.csv')\n",
    "# delete the last row\n",
    "df = df[:-1]\n",
    "df['dataset'] = df['state_filename'].apply(lambda x: 'Internal' if 'state_2022' in x else 'OpenSource')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>index</th>\n",
       "      <th>app_name</th>\n",
       "      <th>state_filename</th>\n",
       "      <th>textfield_id</th>\n",
       "      <th>category_matching</th>\n",
       "      <th>llm_base</th>\n",
       "      <th>droidfiller-reasoning</th>\n",
       "      <th>droidfiller-functioncall</th>\n",
       "      <th>droidfiller</th>\n",
       "      <th>dataset</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>org.koitharu.kotatsu</td>\n",
       "      <td>state_-1399196902</td>\n",
       "      <td>searchView_198_119</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>OpenSource</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2.0</td>\n",
       "      <td>AOL_News_Mail_Video_v6.47.2</td>\n",
       "      <td>state_-2077864795</td>\n",
       "      <td>search_edit_text_182_66</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>OpenSource</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3.0</td>\n",
       "      <td>de.kromke.andreas.opus1musicplayer</td>\n",
       "      <td>state_1643267838</td>\n",
       "      <td>search_src_text_111_71</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>OpenSource</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4.0</td>\n",
       "      <td>Petal_Maps‚ÄìGPS_Navigation_v2.5.0.303(002)</td>\n",
       "      <td>state_-1677063389</td>\n",
       "      <td>search_src_text_174_1600</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>OpenSource</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5.0</td>\n",
       "      <td>AutoScout24_Buy_sell_cars</td>\n",
       "      <td>state_1694299268</td>\n",
       "      <td>textinput_filter_edittext_188_301</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>OpenSource</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>115</th>\n",
       "      <td>127.0</td>\n",
       "      <td>Clock</td>\n",
       "      <td>state_2022-11-01_184421020236</td>\n",
       "      <td>worldclock_search_map_txt_find_158_140</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>Internal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>116</th>\n",
       "      <td>128.0</td>\n",
       "      <td>Clock</td>\n",
       "      <td>state_2022-11-01_184455299444</td>\n",
       "      <td>hour_48_1787</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>Internal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>117</th>\n",
       "      <td>129.0</td>\n",
       "      <td>Clock</td>\n",
       "      <td>state_2022-11-01_184455299444</td>\n",
       "      <td>minute_201_1787</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>Internal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>118</th>\n",
       "      <td>130.0</td>\n",
       "      <td>Clock</td>\n",
       "      <td>state_2022-11-01_184455299444</td>\n",
       "      <td>second_354_1787</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>Internal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>119</th>\n",
       "      <td>131.0</td>\n",
       "      <td>Clock</td>\n",
       "      <td>state_2022-11-01_184455299444</td>\n",
       "      <td>preset_name_48_1944</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>Internal</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>120 rows × 10 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     index                                     app_name  \\\n",
       "0      1.0                         org.koitharu.kotatsu   \n",
       "1      2.0                  AOL_News_Mail_Video_v6.47.2   \n",
       "2      3.0           de.kromke.andreas.opus1musicplayer   \n",
       "3      4.0  Petal_Maps‚ÄìGPS_Navigation_v2.5.0.303(002)   \n",
       "4      5.0                    AutoScout24_Buy_sell_cars   \n",
       "..     ...                                          ...   \n",
       "115  127.0                                        Clock   \n",
       "116  128.0                                        Clock   \n",
       "117  129.0                                        Clock   \n",
       "118  130.0                                        Clock   \n",
       "119  131.0                                        Clock   \n",
       "\n",
       "                    state_filename                            textfield_id  \\\n",
       "0                state_-1399196902                      searchView_198_119   \n",
       "1                state_-2077864795                 search_edit_text_182_66   \n",
       "2                 state_1643267838                  search_src_text_111_71   \n",
       "3                state_-1677063389                search_src_text_174_1600   \n",
       "4                 state_1694299268       textinput_filter_edittext_188_301   \n",
       "..                             ...                                     ...   \n",
       "115  state_2022-11-01_184421020236  worldclock_search_map_txt_find_158_140   \n",
       "116  state_2022-11-01_184455299444                            hour_48_1787   \n",
       "117  state_2022-11-01_184455299444                         minute_201_1787   \n",
       "118  state_2022-11-01_184455299444                         second_354_1787   \n",
       "119  state_2022-11-01_184455299444                     preset_name_48_1944   \n",
       "\n",
       "     category_matching  llm_base  droidfiller-reasoning  \\\n",
       "0                 True      True                   True   \n",
       "1                 True      True                  False   \n",
       "2                 True      True                   True   \n",
       "3                 True      True                   True   \n",
       "4                 True      True                   True   \n",
       "..                 ...       ...                    ...   \n",
       "115               True      True                   True   \n",
       "116              False      True                   True   \n",
       "117              False      True                   True   \n",
       "118              False      True                   True   \n",
       "119              False      True                   True   \n",
       "\n",
       "     droidfiller-functioncall  droidfiller     dataset  \n",
       "0                        True         True  OpenSource  \n",
       "1                        True         True  OpenSource  \n",
       "2                        True         True  OpenSource  \n",
       "3                        True         True  OpenSource  \n",
       "4                        True         True  OpenSource  \n",
       "..                        ...          ...         ...  \n",
       "115                      True         True    Internal  \n",
       "116                      True         True    Internal  \n",
       "117                      True         True    Internal  \n",
       "118                      True         True    Internal  \n",
       "119                      True         True    Internal  \n",
       "\n",
       "[120 rows x 10 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# convert 'O' -> True, 'X' -> False\n",
    "df = df.replace({'O': True, 'X': False})\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Technique</th>\n",
       "      <th>num_true</th>\n",
       "      <th>dataset</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>DroidFilller</td>\n",
       "      <td>118</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-FunctionCall</td>\n",
       "      <td>116</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-Reasoning</td>\n",
       "      <td>112</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-Reasoning-FunctionCall</td>\n",
       "      <td>111</td>\n",
       "      <td>Internal</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 Technique  num_true   dataset\n",
       "0             DroidFilller       118       NaN\n",
       "1            -FunctionCall       116       NaN\n",
       "2               -Reasoning       112       NaN\n",
       "3  -Reasoning-FunctionCall       111  Internal"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results = [] \n",
    "\n",
    "\n",
    "results.append({\n",
    "    'Technique': 'DroidFilller',\n",
    "    'num_true': df[(df['droidfiller'] == True)].shape[0],\n",
    "})\n",
    "\n",
    "results.append({\n",
    "    'Technique': '-FunctionCall',\n",
    "    'num_true': df[(df['droidfiller-functioncall'] == True)].shape[0],\n",
    "})\n",
    "\n",
    "results.append({\n",
    "    'Technique': '-Reasoning',\n",
    "    'num_true': df[(df['droidfiller-reasoning'] == True)].shape[0],\n",
    "})\n",
    "\n",
    "results.append({\n",
    "    'Technique': '-Reasoning-FunctionCall',\n",
    "    'dataset': 'Internal',\n",
    "    'num_true': df[(df['llm_base'] == True)].shape[0]\n",
    "})\n",
    "\n",
    "eval_df = pd.DataFrame(results)\n",
    "eval_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 500x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.ticker import MaxNLocator\n",
    "\n",
    "plt.figure(figsize=(5, 3))\n",
    "\n",
    "# bigger font \n",
    "sns.set(font_scale=2)\n",
    "\n",
    "sns.set_theme(style=\"whitegrid\")\n",
    "\n",
    "# pretty palette\n",
    "ax = sns.barplot(x=\"num_true\", y=\"Technique\", data=eval_df, palette=\"Set2\")\n",
    "# ax.set_title('Text Quality Comparison on Different Ablation Settings', fontsize=15)\n",
    "ax.set_ylabel('')\n",
    "ax.set_xlabel('# of correct text generations')\n",
    "# xtick should be integer\n",
    "ax.set_xlim(100, 120)\n",
    "ax.xaxis.set_major_locator(MaxNLocator(integer=True))\n",
    "\n",
    "\n",
    "# save figure\n",
    "plt.savefig('/Users/greenmon/Dropbox/COINSE/DroidFiller/figures/RQ3_ablation.pdf', bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "sr-stem",
   "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": "3.9.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}