plot.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4b6284c0",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "e62fa581",
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "ed670494",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "8707ea6d",
   "metadata": {},
   "source": [
    "# Table"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "8d3fcde5",
   "metadata": {},
   "outputs": [],
   "source": [
    "data_path = \"data.05_21_16_53\"\n",
    "\n",
    "\n",
    "def read_json_lines_to_df(file_path):\n",
    "    df = pd.read_json(open(file_path, \"r\", encoding=\"utf8\"), lines=True)\n",
    "    return df\n",
    "\n",
    "\n",
    "def get_sample_df(data_path=data_path, task=\"text_summarization\"):\n",
    "    df1 = read_json_lines_to_df(f\"{data_path}/{task}.txt\")\n",
    "    df2 = read_json_lines_to_df(f\"{data_path}/{task}_result.txt\")\n",
    "    df3 = read_json_lines_to_df(f\"{data_path}/{task}_ppl.txt\")\n",
    "    df4 = read_json_lines_to_df(f\"{data_path}/{task}_score.txt\")\n",
    "    # merge by id, watermark_process column\n",
    "    df = pd.merge(df1, df2, on=[\"id\", \"watermark_processor\"])\n",
    "    df = pd.merge(df, df3, on=[\"id\", \"watermark_processor\"])\n",
    "    # df4 misses some rows, so we use left join\n",
    "    df = pd.merge(df, df4, on=[\"id\", \"watermark_processor\"], how=\"left\")\n",
    "    return df\n",
    "\n",
    "\n",
    "def get_bootstrap_df(data_path=data_path, task=\"machine_translation\"):\n",
    "    df1 = read_json_lines_to_df(f\"{data_path}/{task}_bleu.txt\")\n",
    "    return df1\n",
    "\n",
    "\n",
    "def extract_watermark_info(df, return_wp_list=False):\n",
    "    show_wp = [\n",
    "        \"No Watermark\",\n",
    "        \"$\\delta$-reweight\",\n",
    "        \"$\\gamma$-reweight\",\n",
    "        \"$\\delta$-reweight (woh)\",\n",
    "        \"$\\gamma$-reweight (woh)\",\n",
    "    ]\n",
    "    john_wps_set = set()\n",
    "\n",
    "    def map_wp_str(wp_str):\n",
    "        if \"Delta\" in wp_str or \"Gamma\" in wp_str:\n",
    "            woh = \", True)\" in wp_str\n",
    "            if \"Delta\" in wp_str and not woh:\n",
    "                return show_wp[1]\n",
    "            elif \"Delta\" in wp_str and woh:\n",
    "                return show_wp[3]\n",
    "            elif \"Gamma\" in wp_str and not woh:\n",
    "                return show_wp[2]\n",
    "            elif \"Gamma\" in wp_str and woh:\n",
    "                return show_wp[4]\n",
    "        elif \"John\" in wp_str:\n",
    "            import re\n",
    "\n",
    "            delta = re.findall(r\"delta=(\\d+\\.?\\d*)\", wp_str)[0]\n",
    "            n = \"Soft\" + f\"($\\delta$={delta})\"\n",
    "            john_wps_set.add(n)\n",
    "            return n\n",
    "        if wp_str == \"None\":\n",
    "            return show_wp[0]\n",
    "        else:\n",
    "            raise ValueError(\"Unknown watermark: {}\".format(wp_str))\n",
    "\n",
    "    df = df.assign(show_wp_name=df[\"watermark_processor\"].apply(map_wp_str))\n",
    "    john_wps = sorted(list(john_wps_set))\n",
    "    show_wp = show_wp + john_wps\n",
    "    if return_wp_list:\n",
    "        return df, show_wp\n",
    "    else:\n",
    "        return df\n",
    "\n",
    "\n",
    "def sample_df_2_stat(df, bootstrap=False, show_wp=None):\n",
    "    sdf = df.melt(\n",
    "        id_vars=[\"show_wp_name\"],\n",
    "        value_vars=[c for c in df.columns if df[c].dtype == np.float64],\n",
    "        var_name=\"score\",\n",
    "        value_name=\"value\",\n",
    "    )\n",
    "    sdf = sdf.groupby([\"show_wp_name\", \"score\"]).agg([\"mean\", \"std\", \"count\"])\n",
    "\n",
    "    def format_fn(x):\n",
    "        mean = x[\"mean\"]\n",
    "        if not bootstrap:\n",
    "            std = x[\"std\"] / np.sqrt(x[\"count\"])\n",
    "        else:\n",
    "            std = x[\"std\"]\n",
    "        if not np.isfinite(std):\n",
    "            return f\"{mean:.2f}±{std:.2f}\"\n",
    "        useful_digits = np.max(-int(np.floor(np.log10(std / 3))), 0)\n",
    "        fmt_str = f\"{{:.{useful_digits}f}}±{{:.{useful_digits}f}}\"\n",
    "        return fmt_str.format(mean, std)\n",
    "\n",
    "    sdf = sdf[\"value\"].apply(format_fn, axis=1).unstack()\n",
    "    if show_wp:\n",
    "        sdf = sdf.loc[show_wp]\n",
    "    return sdf\n",
    "\n",
    "\n",
    "def merge_stat_df(df1, df2):\n",
    "    df = pd.merge(df1, df2, left_index=True, right_index=True)\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "2ce5adad",
   "metadata": {},
   "outputs": [],
   "source": [
    "tsdf, show_wp = extract_watermark_info(get_sample_df(), return_wp_list=True)\n",
    "mtdf = extract_watermark_info(get_sample_df(task=\"machine_translation\"))\n",
    "mtbsdf = extract_watermark_info(get_bootstrap_df())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "1578e7ab",
   "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>score</th>\n",
       "      <th>bertscore.f1</th>\n",
       "      <th>bertscore.precision</th>\n",
       "      <th>bertscore.recall</th>\n",
       "      <th>best_dist_q</th>\n",
       "      <th>best_sum_score</th>\n",
       "      <th>lens</th>\n",
       "      <th>ppl</th>\n",
       "      <th>rouge1</th>\n",
       "      <th>rouge2</th>\n",
       "      <th>rougeL</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>show_wp_name</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>No Watermark</th>\n",
       "      <td>0.3270±0.0008</td>\n",
       "      <td>0.3180±0.0009</td>\n",
       "      <td>0.3361±0.0010</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>5.024±0.018</td>\n",
       "      <td>0.3856±0.0009</td>\n",
       "      <td>0.1388±0.0008</td>\n",
       "      <td>0.2445±0.0008</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>$\\delta$-reweight</th>\n",
       "      <td>0.3271±0.0008</td>\n",
       "      <td>0.3180±0.0009</td>\n",
       "      <td>0.3365±0.0010</td>\n",
       "      <td>0.0527±0.0005</td>\n",
       "      <td>67.46±0.23</td>\n",
       "      <td>76.80±0.16</td>\n",
       "      <td>5.022±0.018</td>\n",
       "      <td>0.3857±0.0009</td>\n",
       "      <td>0.1392±0.0008</td>\n",
       "      <td>0.2451±0.0008</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>$\\gamma$-reweight</th>\n",
       "      <td>0.3269±0.0008</td>\n",
       "      <td>0.3180±0.0009</td>\n",
       "      <td>0.3360±0.0010</td>\n",
       "      <td>0.0289±0.0004</td>\n",
       "      <td>16.87±0.06</td>\n",
       "      <td>76.42±0.16</td>\n",
       "      <td>5.019±0.018</td>\n",
       "      <td>0.3860±0.0009</td>\n",
       "      <td>0.1397±0.0008</td>\n",
       "      <td>0.2451±0.0008</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>$\\delta$-reweight (woh)</th>\n",
       "      <td>0.3276±0.0008</td>\n",
       "      <td>0.3185±0.0009</td>\n",
       "      <td>0.3370±0.0010</td>\n",
       "      <td>0.0100±0.0003</td>\n",
       "      <td>71.86±0.24</td>\n",
       "      <td>76.94±0.16</td>\n",
       "      <td>5.000±0.018</td>\n",
       "      <td>0.3857±0.0009</td>\n",
       "      <td>0.1398±0.0008</td>\n",
       "      <td>0.2455±0.0008</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>$\\gamma$-reweight (woh)</th>\n",
       "      <td>0.3269±0.0008</td>\n",
       "      <td>0.3178±0.0009</td>\n",
       "      <td>0.3361±0.0010</td>\n",
       "      <td>0.00684±0.00024</td>\n",
       "      <td>18.46±0.06</td>\n",
       "      <td>76.64±0.16</td>\n",
       "      <td>5.010±0.018</td>\n",
       "      <td>0.3856±0.0009</td>\n",
       "      <td>0.1393±0.0008</td>\n",
       "      <td>0.2447±0.0008</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Soft($\\delta$=0.0)</th>\n",
       "      <td>0.3270±0.0008</td>\n",
       "      <td>0.3180±0.0009</td>\n",
       "      <td>0.3361±0.0010</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>5.024±0.018</td>\n",
       "      <td>0.3856±0.0009</td>\n",
       "      <td>0.1388±0.0008</td>\n",
       "      <td>0.2445±0.0008</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Soft($\\delta$=1.0)</th>\n",
       "      <td>0.3235±0.0008</td>\n",
       "      <td>0.3092±0.0009</td>\n",
       "      <td>0.3382±0.0009</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>5.313±0.018</td>\n",
       "      <td>0.3820±0.0009</td>\n",
       "      <td>0.1344±0.0007</td>\n",
       "      <td>0.2400±0.0007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Soft($\\delta$=2.0)</th>\n",
       "      <td>0.3121±0.0008</td>\n",
       "      <td>0.2908±0.0008</td>\n",
       "      <td>0.3339±0.0009</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>6.253±0.022</td>\n",
       "      <td>0.3717±0.0008</td>\n",
       "      <td>0.1238±0.0007</td>\n",
       "      <td>0.2293±0.0007</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "score                     bertscore.f1 bertscore.precision bertscore.recall  \\\n",
       "show_wp_name                                                                  \n",
       "No Watermark             0.3270±0.0008       0.3180±0.0009    0.3361±0.0010   \n",
       "$\\delta$-reweight        0.3271±0.0008       0.3180±0.0009    0.3365±0.0010   \n",
       "$\\gamma$-reweight        0.3269±0.0008       0.3180±0.0009    0.3360±0.0010   \n",
       "$\\delta$-reweight (woh)  0.3276±0.0008       0.3185±0.0009    0.3370±0.0010   \n",
       "$\\gamma$-reweight (woh)  0.3269±0.0008       0.3178±0.0009    0.3361±0.0010   \n",
       "Soft($\\delta$=0.0)       0.3270±0.0008       0.3180±0.0009    0.3361±0.0010   \n",
       "Soft($\\delta$=1.0)       0.3235±0.0008       0.3092±0.0009    0.3382±0.0009   \n",
       "Soft($\\delta$=2.0)       0.3121±0.0008       0.2908±0.0008    0.3339±0.0009   \n",
       "\n",
       "score                        best_dist_q best_sum_score        lens  \\\n",
       "show_wp_name                                                          \n",
       "No Watermark                     nan±nan        nan±nan     nan±nan   \n",
       "$\\delta$-reweight          0.0527±0.0005     67.46±0.23  76.80±0.16   \n",
       "$\\gamma$-reweight          0.0289±0.0004     16.87±0.06  76.42±0.16   \n",
       "$\\delta$-reweight (woh)    0.0100±0.0003     71.86±0.24  76.94±0.16   \n",
       "$\\gamma$-reweight (woh)  0.00684±0.00024     18.46±0.06  76.64±0.16   \n",
       "Soft($\\delta$=0.0)               nan±nan        nan±nan     nan±nan   \n",
       "Soft($\\delta$=1.0)               nan±nan        nan±nan     nan±nan   \n",
       "Soft($\\delta$=2.0)               nan±nan        nan±nan     nan±nan   \n",
       "\n",
       "score                            ppl         rouge1         rouge2  \\\n",
       "show_wp_name                                                         \n",
       "No Watermark             5.024±0.018  0.3856±0.0009  0.1388±0.0008   \n",
       "$\\delta$-reweight        5.022±0.018  0.3857±0.0009  0.1392±0.0008   \n",
       "$\\gamma$-reweight        5.019±0.018  0.3860±0.0009  0.1397±0.0008   \n",
       "$\\delta$-reweight (woh)  5.000±0.018  0.3857±0.0009  0.1398±0.0008   \n",
       "$\\gamma$-reweight (woh)  5.010±0.018  0.3856±0.0009  0.1393±0.0008   \n",
       "Soft($\\delta$=0.0)       5.024±0.018  0.3856±0.0009  0.1388±0.0008   \n",
       "Soft($\\delta$=1.0)       5.313±0.018  0.3820±0.0009  0.1344±0.0007   \n",
       "Soft($\\delta$=2.0)       6.253±0.022  0.3717±0.0008  0.1238±0.0007   \n",
       "\n",
       "score                           rougeL  \n",
       "show_wp_name                            \n",
       "No Watermark             0.2445±0.0008  \n",
       "$\\delta$-reweight        0.2451±0.0008  \n",
       "$\\gamma$-reweight        0.2451±0.0008  \n",
       "$\\delta$-reweight (woh)  0.2455±0.0008  \n",
       "$\\gamma$-reweight (woh)  0.2447±0.0008  \n",
       "Soft($\\delta$=0.0)       0.2445±0.0008  \n",
       "Soft($\\delta$=1.0)       0.2400±0.0007  \n",
       "Soft($\\delta$=2.0)       0.2293±0.0007  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample_df_2_stat(tsdf, show_wp=show_wp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "203eddb6",
   "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>score</th>\n",
       "      <th>bertscore.f1</th>\n",
       "      <th>bertscore.precision</th>\n",
       "      <th>bertscore.recall</th>\n",
       "      <th>best_dist_q</th>\n",
       "      <th>best_sum_score</th>\n",
       "      <th>lens</th>\n",
       "      <th>ppl</th>\n",
       "      <th>bleu</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>show_wp_name</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>No Watermark</th>\n",
       "      <td>0.559±0.003</td>\n",
       "      <td>0.545±0.004</td>\n",
       "      <td>0.574±0.003</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>8±6</td>\n",
       "      <td>21.8±0.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>$\\delta$-reweight</th>\n",
       "      <td>0.563±0.003</td>\n",
       "      <td>0.549±0.004</td>\n",
       "      <td>0.577±0.003</td>\n",
       "      <td>0.083±0.004</td>\n",
       "      <td>14.61±0.28</td>\n",
       "      <td>34.9±0.4</td>\n",
       "      <td>8±6</td>\n",
       "      <td>21.7±0.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>$\\gamma$-reweight</th>\n",
       "      <td>0.562±0.003</td>\n",
       "      <td>0.548±0.004</td>\n",
       "      <td>0.576±0.003</td>\n",
       "      <td>0.0570±0.0026</td>\n",
       "      <td>3.68±0.07</td>\n",
       "      <td>34.9±0.4</td>\n",
       "      <td>8±6</td>\n",
       "      <td>21.8±0.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>$\\delta$-reweight (woh)</th>\n",
       "      <td>0.569±0.003</td>\n",
       "      <td>0.555±0.003</td>\n",
       "      <td>0.583±0.003</td>\n",
       "      <td>0.0381±0.0029</td>\n",
       "      <td>15.78±0.29</td>\n",
       "      <td>34.9±0.4</td>\n",
       "      <td>2.114±0.020</td>\n",
       "      <td>21.9±0.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>$\\gamma$-reweight (woh)</th>\n",
       "      <td>0.563±0.003</td>\n",
       "      <td>0.549±0.003</td>\n",
       "      <td>0.577±0.003</td>\n",
       "      <td>0.0280±0.0019</td>\n",
       "      <td>4.17±0.07</td>\n",
       "      <td>34.9±0.4</td>\n",
       "      <td>2.24±0.04</td>\n",
       "      <td>21.8±0.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Soft($\\delta$=0.0)</th>\n",
       "      <td>0.559±0.003</td>\n",
       "      <td>0.545±0.004</td>\n",
       "      <td>0.574±0.003</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>8±6</td>\n",
       "      <td>21.8±0.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Soft($\\delta$=1.0)</th>\n",
       "      <td>0.551±0.003</td>\n",
       "      <td>0.536±0.004</td>\n",
       "      <td>0.567±0.004</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>8±6</td>\n",
       "      <td>21.0±0.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Soft($\\delta$=2.0)</th>\n",
       "      <td>0.538±0.003</td>\n",
       "      <td>0.522±0.004</td>\n",
       "      <td>0.554±0.003</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>nan±nan</td>\n",
       "      <td>9±6</td>\n",
       "      <td>19.5±0.3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "score                   bertscore.f1 bertscore.precision bertscore.recall  \\\n",
       "show_wp_name                                                                \n",
       "No Watermark             0.559±0.003         0.545±0.004      0.574±0.003   \n",
       "$\\delta$-reweight        0.563±0.003         0.549±0.004      0.577±0.003   \n",
       "$\\gamma$-reweight        0.562±0.003         0.548±0.004      0.576±0.003   \n",
       "$\\delta$-reweight (woh)  0.569±0.003         0.555±0.003      0.583±0.003   \n",
       "$\\gamma$-reweight (woh)  0.563±0.003         0.549±0.003      0.577±0.003   \n",
       "Soft($\\delta$=0.0)       0.559±0.003         0.545±0.004      0.574±0.003   \n",
       "Soft($\\delta$=1.0)       0.551±0.003         0.536±0.004      0.567±0.004   \n",
       "Soft($\\delta$=2.0)       0.538±0.003         0.522±0.004      0.554±0.003   \n",
       "\n",
       "score                      best_dist_q best_sum_score      lens          ppl  \\\n",
       "show_wp_name                                                                   \n",
       "No Watermark                   nan±nan        nan±nan   nan±nan          8±6   \n",
       "$\\delta$-reweight          0.083±0.004     14.61±0.28  34.9±0.4          8±6   \n",
       "$\\gamma$-reweight        0.0570±0.0026      3.68±0.07  34.9±0.4          8±6   \n",
       "$\\delta$-reweight (woh)  0.0381±0.0029     15.78±0.29  34.9±0.4  2.114±0.020   \n",
       "$\\gamma$-reweight (woh)  0.0280±0.0019      4.17±0.07  34.9±0.4    2.24±0.04   \n",
       "Soft($\\delta$=0.0)             nan±nan        nan±nan   nan±nan          8±6   \n",
       "Soft($\\delta$=1.0)             nan±nan        nan±nan   nan±nan          8±6   \n",
       "Soft($\\delta$=2.0)             nan±nan        nan±nan   nan±nan          9±6   \n",
       "\n",
       "score                        bleu  \n",
       "show_wp_name                       \n",
       "No Watermark             21.8±0.3  \n",
       "$\\delta$-reweight        21.7±0.3  \n",
       "$\\gamma$-reweight        21.8±0.3  \n",
       "$\\delta$-reweight (woh)  21.9±0.3  \n",
       "$\\gamma$-reweight (woh)  21.8±0.3  \n",
       "Soft($\\delta$=0.0)       21.8±0.3  \n",
       "Soft($\\delta$=1.0)       21.0±0.3  \n",
       "Soft($\\delta$=2.0)       19.5±0.3  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "merge_stat_df(\n",
    "    sample_df_2_stat(mtdf, show_wp=show_wp),\n",
    "    sample_df_2_stat(mtbsdf, bootstrap=True, show_wp=show_wp),\n",
    ")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "c866a604",
   "metadata": {},
   "source": [
    "# violin diagram"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "2d61ff9f",
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "def plot_sample_df(df, score:str, print_name:str, show_wp=None, show_mean=False, violin_kwargs={}):\n",
    "    sns.violinplot(data=df, x=\"show_wp_name\", y=score, order=show_wp, **violin_kwargs)\n",
    "    plt.ylabel(print_name)\n",
    "    plt.xlabel(\"\")\n",
    "    if show_mean:\n",
    "        plt.axhline(\n",
    "            df[df.show_wp_name == \"No Watermark\"][score].mean(),\n",
    "            color=\"red\",\n",
    "            linestyle=\"--\",\n",
    "            label=\"Mean performace w/o watermark\",\n",
    "        )\n",
    "        plt.legend(loc=\"lower left\")\n",
    "    plt.xticks(rotation=15, ha='center')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "e8fad046",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot_sample_df(tsdf, \"bertscore.f1\", \"F1\", show_wp=show_wp, show_mean=True)\n",
    "# plt.ylim(0.1, 0.6)\n",
    "# plot_sample_df(tsdf, \"bertscore.precision\", \"Precision\", show_wp=show_wp, show_mean=True)\n",
    "# plt.ylim(0.1, 0.6)\n",
    "# plot_sample_df(tsdf, \"bertscore.recall\", \"Recall\", show_wp=show_wp, show_mean=True)\n",
    "# plt.ylim(0.1, 0.6)\n",
    "# plot_sample_df(tsdf, \"rouge1\", \"ROUGE-1\", show_wp=show_wp, show_mean=True)\n",
    "# plt.ylim(0.15, 0.65)\n",
    "# plot_sample_df(tsdf, \"rouge2\", \"ROUGE-2\", show_wp=show_wp, show_mean=True)\n",
    "# plt.ylim(0.0, 0.4)\n",
    "plot_sample_df(tsdf, \"ppl\", \"Perplexity\", show_wp=show_wp, show_mean=True)\n",
    "plt.ylim(1, 10)\n",
    "plt.show()\n",
    "# plt.savefig(\"figures/Violin_TS.pdf\", format=\"pdf\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "e1204bf3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot_sample_df(mtdf, \"bertscore.f1\", \"F1\", show_wp=show_wp, show_mean=True)\n",
    "# plt.ylim(0.2, 1.0)\n",
    "# plot_sample_df(mtdf, \"bertscore.precision\", \"Precision\", show_wp=show_wp, show_mean=True)\n",
    "# plt.ylim(0.1, 1.0)\n",
    "# plot_sample_df(mtdf, \"bertscore.recall\", \"Recall\", show_wp=show_wp, show_mean=True)\n",
    "# plt.ylim(0.1, 1.0)\n",
    "plot_sample_df(mtbsdf, \"bleu\", \"BLEU\", show_wp=show_wp, show_mean=True)\n",
    "plt.show()\n",
    "# plt.savefig(\"figures/Violin_MT.pdf\", format=\"pdf\")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "4475d4ae",
   "metadata": {},
   "source": [
    "# distribution of score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "4df5587c",
   "metadata": {},
   "outputs": [],
   "source": [
    "def extract_len_info(df):\n",
    "    df = df.assign(\n",
    "        output_words=df[\"display_output\"].apply(lambda x: len(x.split(\" \"))),\n",
    "        output_chars=df[\"display_output\"].apply(lambda x: len(x)),\n",
    "    )\n",
    "    return df\n",
    "\n",
    "\n",
    "def filter_wh_score(df):\n",
    "    df = df[df[\"best_sum_score\"].notna()]\n",
    "    df = df[~df.show_wp_name.str.contains(\"woh\")]\n",
    "    return df\n",
    "\n",
    "\n",
    "def merge_tasks(dfs: dict):\n",
    "    # add new column called task, and use key as task name\n",
    "    for k, v in dfs.items():\n",
    "        v[\"task\"] = k\n",
    "    df = pd.concat(dfs.values())\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "427e47c8",
   "metadata": {},
   "outputs": [],
   "source": [
    "merged_df = merge_tasks(\n",
    "    {\n",
    "        \"Text summarization\": extract_len_info(filter_wh_score(tsdf)),\n",
    "        \"Machine translation\": extract_len_info(filter_wh_score(mtdf)),\n",
    "    }\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "26c20810",
   "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>variable</th>\n",
       "      <th>best_score</th>\n",
       "      <th>entropy</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>task</th>\n",
       "      <th>show_wp_name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">Text summarization</th>\n",
       "      <th>$\\delta$-reweight</th>\n",
       "      <td>0.8784±0.0015</td>\n",
       "      <td>0.9694±0.0009</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>$\\gamma$-reweight</th>\n",
       "      <td>0.2207±0.0004</td>\n",
       "      <td>0.9695±0.0009</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">Machine translation</th>\n",
       "      <th>$\\delta$-reweight</th>\n",
       "      <td>0.4192±0.0043</td>\n",
       "      <td>0.5260±0.0024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>$\\gamma$-reweight</th>\n",
       "      <td>0.1056±0.0011</td>\n",
       "      <td>0.5271±0.0024</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "variable                                  best_score        entropy\n",
       "task                show_wp_name                                   \n",
       "Text summarization  $\\delta$-reweight  0.8784±0.0015  0.9694±0.0009\n",
       "                    $\\gamma$-reweight  0.2207±0.0004  0.9695±0.0009\n",
       "Machine translation $\\delta$-reweight  0.4192±0.0043  0.5260±0.0024\n",
       "                    $\\gamma$-reweight  0.1056±0.0011  0.5271±0.0024"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "merged_df[[\"task\", \"show_wp_name\", \"best_score\", \"entropy\"]].explode(\n",
    "    [\"best_score\", \"entropy\"]\n",
    ").astype({\"best_score\": float, \"entropy\": float}).melt(\n",
    "    id_vars=[\"task\", \"show_wp_name\"], value_vars=[\"best_score\", \"entropy\"]\n",
    ").groupby([\"task\", \"show_wp_name\", \"variable\"]).agg(\n",
    "    [\"mean\", \"std\", \"count\"]\n",
    ")['value'].apply(\n",
    "    lambda x: f\"{x['mean']:.4f}±{x['std']/np.sqrt(x['count']):.4f}\", axis=1\n",
    ").unstack().loc[[\"Text summarization\", \"Machine translation\"]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "2f37b47f",
   "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>task</th>\n",
       "      <th>Text summarization</th>\n",
       "      <th>Machine translation</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>show_wp_name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>$\\delta$-reweight</th>\n",
       "      <td>0.8784±0.0015</td>\n",
       "      <td>0.4192±0.0043</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>$\\gamma$-reweight</th>\n",
       "      <td>0.2207±0.0004</td>\n",
       "      <td>0.1056±0.0011</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "task              Text summarization Machine translation\n",
       "show_wp_name                                            \n",
       "$\\delta$-reweight      0.8784±0.0015       0.4192±0.0043\n",
       "$\\gamma$-reweight      0.2207±0.0004       0.1056±0.0011"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\begin{tabular}{lll}\n",
      "\\toprule\n",
      "task & Text summarization & Machine translation \\\\\n",
      "show\\_wp\\_name      &                    &                     \\\\\n",
      "\\midrule\n",
      "\\$\\textbackslash delta\\$-reweight &      0.8784±0.0015 &       0.4192±0.0043 \\\\\n",
      "\\$\\textbackslash gamma\\$-reweight &      0.2207±0.0004 &       0.1056±0.0011 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_2287530/4219267337.py:11: 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(tempdf.to_latex())\n"
     ]
    }
   ],
   "source": [
    "tempdf=merged_df[[\"task\", \"show_wp_name\", \"best_score\", \"entropy\"]].explode(\n",
    "    [\"best_score\", \"entropy\"]\n",
    ").astype({\"best_score\": float, \"entropy\": float}).melt(\n",
    "    id_vars=[\"task\", \"show_wp_name\"], value_vars=[\"best_score\", \"entropy\"]\n",
    ").groupby([\"task\", \"show_wp_name\", \"variable\"]).agg(\n",
    "    [\"mean\", \"std\", \"count\"]\n",
    ")['value'].apply(\n",
    "    lambda x: f\"{x['mean']:.4f}±{x['std']/np.sqrt(x['count']):.4f}\", axis=1\n",
    ").unstack()['best_score'].unstack().loc[[\"Text summarization\", \"Machine translation\"]].transpose()\n",
    "display(tempdf)\n",
    "print(tempdf.to_latex())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "4622d0b0",
   "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>task</th>\n",
       "      <th>Text summarization</th>\n",
       "      <th>Machine translation</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>show_wp_name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>$\\delta$-reweight</th>\n",
       "      <td>0.8784±1.4354</td>\n",
       "      <td>0.4192±1.1361</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>$\\gamma$-reweight</th>\n",
       "      <td>0.2207±0.3678</td>\n",
       "      <td>0.1056±0.2916</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "task              Text summarization Machine translation\n",
       "show_wp_name                                            \n",
       "$\\delta$-reweight      0.8784±1.4354       0.4192±1.1361\n",
       "$\\gamma$-reweight      0.2207±0.3678       0.1056±0.2916"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\begin{tabular}{lll}\n",
      "\\toprule\n",
      "task & Text summarization & Machine translation \\\\\n",
      "show\\_wp\\_name      &                    &                     \\\\\n",
      "\\midrule\n",
      "\\$\\textbackslash delta\\$-reweight &      0.8784±1.4354 &       0.4192±1.1361 \\\\\n",
      "\\$\\textbackslash gamma\\$-reweight &      0.2207±0.3678 &       0.1056±0.2916 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_2287530/4135557175.py:11: 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(tempdf.to_latex())\n"
     ]
    }
   ],
   "source": [
    "tempdf=merged_df[[\"task\", \"show_wp_name\", \"best_score\", \"entropy\"]].explode(\n",
    "    [\"best_score\", \"entropy\"]\n",
    ").astype({\"best_score\": float, \"entropy\": float}).melt(\n",
    "    id_vars=[\"task\", \"show_wp_name\"], value_vars=[\"best_score\", \"entropy\"]\n",
    ").groupby([\"task\", \"show_wp_name\", \"variable\"]).agg(\n",
    "    [\"mean\", \"std\", \"count\"]\n",
    ")['value'].apply(\n",
    "    lambda x: f\"{x['mean']:.4f}±{x['std']:.4f}\", axis=1\n",
    ").unstack()['best_score'].unstack().loc[[\"Text summarization\", \"Machine translation\"]].transpose()\n",
    "display(tempdf)\n",
    "print(tempdf.to_latex())"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "8b6905e4",
   "metadata": {},
   "source": [
    "# latex output"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "7604012d",
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "matplotlib.rcParams.update(\n",
    "    {\n",
    "        \"font.family\": \"serif\",\n",
    "        \"text.usetex\": True,\n",
    "    }\n",
    ")\n",
    "plt.rc(\"text.latex\", preamble=r\"\\usepackage{amsmath}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "d55fdf62",
   "metadata": {},
   "outputs": [],
   "source": [
    "metric_show_name = {\n",
    "    \"bertscore.f1\": \"BERTScore F1\",\n",
    "    \"bertscore.precision\": \"BERTScore Precision\",\n",
    "    \"bertscore.recall\": \"BERTScore Recall\",\n",
    "    \"rouge1\": \"ROUGE-1\",\n",
    "    \"rouge2\": \"ROUGE-2\",\n",
    "    \"rougeL\": \"ROUGE-L\",\n",
    "    \"bleu\": \"BLEU\",\n",
    "    \"ppl\": \"Perplexity\",\n",
    "}\n",
    "\n",
    "scaled_columns = [\n",
    "    \"bertscore.f1\",\n",
    "    \"bertscore.precision\",\n",
    "    \"bertscore.recall\",\n",
    "    \"rouge1\",\n",
    "    \"rouge2\",\n",
    "    \"rougeL\",\n",
    "]\n",
    "\n",
    "\n",
    "def scale_df(df, scaled_columns=scaled_columns, scale=100):\n",
    "    df_scaled = df.copy()\n",
    "    for c in scaled_columns:\n",
    "        if c in df_scaled.columns:\n",
    "            df_scaled[c] = df_scaled[c] * scale\n",
    "    return df_scaled"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "627069c4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\begin{tabular}{llllll}\n",
      "\\toprule\n",
      "score & bertscore.f1\\_x &     ROUGE-1 &   Perplexity & bertscore.f1\\_y &      BLEU \\\\\n",
      "show\\_wp\\_name            &                &             &              &                &           \\\\\n",
      "\\midrule\n",
      "No Watermark            &     32.70±0.08 &  38.56±0.09 &  5.024±0.018 &       55.9±0.3 &  21.8±0.3 \\\\\n",
      "\\$\\textbackslash delta\\$-reweight       &     32.71±0.08 &  38.57±0.09 &  5.022±0.018 &       56.3±0.3 &  21.7±0.3 \\\\\n",
      "\\$\\textbackslash gamma\\$-reweight       &     32.69±0.08 &  38.60±0.09 &  5.019±0.018 &       56.2±0.3 &  21.8±0.3 \\\\\n",
      "\\$\\textbackslash delta\\$-reweight (woh) &     32.76±0.08 &  38.57±0.09 &  5.000±0.018 &       56.9±0.3 &  21.9±0.3 \\\\\n",
      "\\$\\textbackslash gamma\\$-reweight (woh) &     32.69±0.08 &  38.56±0.09 &  5.010±0.018 &       56.3±0.3 &  21.8±0.3 \\\\\n",
      "Soft(\\$\\textbackslash delta\\$=0.0)      &     32.70±0.08 &  38.56±0.09 &  5.024±0.018 &       55.9±0.3 &  21.8±0.3 \\\\\n",
      "Soft(\\$\\textbackslash delta\\$=1.0)      &     32.35±0.08 &  38.20±0.09 &  5.313±0.018 &       55.1±0.3 &  21.0±0.3 \\\\\n",
      "Soft(\\$\\textbackslash delta\\$=2.0)      &     31.21±0.08 &  37.17±0.08 &  6.253±0.022 &       53.8±0.3 &  19.5±0.3 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_2287530/1244251916.py:8: 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(merge_stat_df(df1, df2).rename(columns=metric_show_name).to_latex())\n"
     ]
    }
   ],
   "source": [
    "df1 = sample_df_2_stat(scale_df(tsdf), show_wp=show_wp)[\n",
    "    [\"bertscore.f1\", \"rouge1\", \"ppl\"]\n",
    "]\n",
    "df2 = merge_stat_df(\n",
    "    sample_df_2_stat(scale_df(mtdf), show_wp=show_wp),\n",
    "    sample_df_2_stat(mtbsdf, bootstrap=True, show_wp=show_wp),\n",
    ")[[\"bertscore.f1\", \"bleu\"]]\n",
    "print(merge_stat_df(df1, df2).rename(columns=metric_show_name).to_latex())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "49b0b852",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig_scale = 1\n",
    "plt.figure(figsize=(4 * fig_scale, 3 * fig_scale))\n",
    "plot_sample_df(\n",
    "    tsdf[~tsdf.show_wp_name.str.contains(\"woh\")],\n",
    "    \"ppl\",\n",
    "    \"Perplexity\",\n",
    "    show_wp=filter(lambda x: \"woh\" not in x, show_wp),\n",
    "    #show_mean=True,\n",
    ")\n",
    "plt.ylim(1, 10)\n",
    "plt.tight_layout()\n",
    "plt.savefig(\n",
    "    \"figures/Violin_TS_PPL.pdf\",\n",
    "    format=\"pdf\",\n",
    "    bbox_inches=\"tight\",\n",
    "    pad_inches=0.0,\n",
    ")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "e2a89df4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEiCAYAAAAWOs4eAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/NK7nSAAAACXBIWXMAAA9hAAAPYQGoP6dpAABtBElEQVR4nO2dd3gU1frHP7MlvSckkJAQQug9VGmiFEHkotJsKF4V9V69XrzegCgKNorY/amAXnuhWEBAgSiI0gQC0muAJARIQnrbbJnfH5sdsqmbZLO7CefzPPtk2s68JzN7vnPOed/3SLIsywgEAoFAAKicbYBAIBAIXAchCgKBQCBQEKIgEAgEAgUhCgKBQCBQEKIgEAgEAgUhCgKBQCBQEKIgEAgEAgUhCgKBQCBQEKIgEAgEAgWNMy6amJhIQkICAHv27GH58uUEBAQAKNtzcnLYs2cPU6dOJS4uzhlmCgQCwTWHU1oKCQkJxMfHEx8fT79+/RgxYoSyb/LkyQQFBTFp0iTatWvH5MmTnWGiQCAQXJM4XBQSExNZsGCBsj5p0iQSExNJSkoCYNWqVVYtA0sLQiAQCASNj8NFIS4ujuXLlyvrOTk5AAQFBQEwcuRIZd+qVat4+OGHHWqfQCAQXMtIzs6SOmvWLBITE9m8ebOyLTExkRUrVhAcHEx8fHyV39PpdOh0OmXdZDKRlZVFcHAwkiQ1ut0CgUDgCsiyTH5+PuHh4ahUdnjPl51Idna2HBMTI2dnZ1e5b8aMGfKqVauq/O7zzz8vA+IjPuIjPuIDckpKil3qZae2FB5++GFmzZpFTExMlfsTEhIYNWoU2dnZlcYWKrYUcnNziYqKIiUlBT8/v8Y0mzFjxijLP//8c6Ney1Hcd999XL58GWg+ZXr99dfZtGkT0HzK9Ouvv7J48WKg+ZRJ0DDy8vKIjIwkJycHf3//Bp/PKS6pAIsXL1YEwTKusHfvXiZPnkx2djaAIhZJSUmV3FLd3d1xd3evdF4/P79GFwWN5uq/rbGv5Sjc3NyUcjWXMnl4eDS7Mnl7eytl8vX1FV2lAgV7PQtOcUldvXo1cXFxiiCsXLmSgIAAgoKCrAaaExMTCQgIEHEKDsCJDUZBHSh/n4xGoxMtETRXHN5SSEpKqhR7EBAQwIwZM4iLi2Pq1KksW7YMgM2bN7Nv3z5Hmyhohsiy3OzeqktLS61arQKBPXD4ExUTE1PjW+mkSZOU5RkzZjjCJME1gE6nw8PDw9lmNJjywlZaWoqXl5cTrRE0R0TuI8E1QXFxsbNNsAt6vV5ZLu9oIRDYCyEKgmZLeSEoLCx0oiX2o3yZhCgIGgMhCoJmiSzLilcbQEFBgfOMsSPlRaGoqMiJlgiaK0IUBM2SkpIS9u/fr6zn5+c70Rr7Ub4czaVMAtdCiEIdEW6ATZPmUIHKsqzE8IA5aEkgsDdCFOrIlStXrNZLS0udZImgLjQHUSgpKbGKYhaiIGgMhCjUkfL91FWtC1wDg8FgtZ6bm+skSxoP8ewJGgMhCnWkfPO9qvWmiCzLmEwmZ5thV7KysqzWMzMznWRJ41GxjAKBPRCiUEcyMjKs1i1J5JoyJSUlpKenK+vNwdXxwoULVuspKSlOssR+VBTu5vDsCVwPIQp1pGLl0hwqm4o0h7dqy0x+Fk6fPt3k8zuVF25ons+ewPkIUagjZ86csVo/efKkkyyxHxX93S9evOgkS+zH8ePHlWUZmdzcXNLS0pxoUcOp+OxduHCh2QTlCVwHIQp1wGQycfToUatthw8fbvJvoKmpqVbrycnJTrLEPsiyzOHDh69uCDT/OXjwoHMMshPlyxRQxTaBwB4IUagDJ0+etIqMlVGRkZHB+fPnnWhVwzl79qzVesU30qZGcnKy1SCs3MIs2omJic4yyS6UD8ZrU/Z3z549zjFG0GwRolAHdu7cabVu9AsDYNeuXc4wx25UFIHTp087yRL7UPF+yKFmUfjzzz+brJdVenq6lXi3K/vb1J89geshRKEO7Nixw2rd6B8JwPbt251hjt04deqU1fqZM2cq+fk3JbZt22a9IRhkrTka+NChQ84xqoFUfPZiMP94k5OTxYCzwK4IUbCR9PR0Tpw4QfnRA2OAWRQOHTrUZIOjDAaDVcvAU22itLS0UpdSUyEtLa1yxa8COcJ85zZv3uwEqxpOxRcPTyC6bLmiYAgEDUGIgo1YfpQm7xbKNtndB6NXECaTqcn+MM+cOWOVqqOtnzm305EjR5xlUoP46aefgKvjCBbkNub1X375hZKSEofb1RAKCwurHA/pVPb3999/d6xBgmaNEAUb2bp1KwDGgCir7cZA85Dfb7/95miT7MJff/1ltR7rb6hye1PAYDCwYcMGAOToCh5hLUD2liksLGTLli1OsK7+bN++Hb1eT3CF7Z3L/h46dKjJx5bk5OSItB0ughAFG7h8+TIHDhwAwBAUbbXPEBQDwO7du5vkQ717926r9c6B5pm99u7d2+Qywu7atYuMjAxkNxm5VQVRkEBua962du1aJ1hXfyxdXl0qbA9AIhKzC25CQoLD7bIXer2eadOmMW3atCY9ltVcEKJgA2vXrkWWZYy+rZDdfaz2yV6BGL1bYDQaWbdunZMsrB/5+flWbo4A7fyM+GhN5ObmNrnWwpo1a4CyVoK68n65rQySuWus4uC6q3LhwgX+/PNPAHpUsb932d81a9Y0Wc+qvLw8cnNzyc3NFRMHuQBCFGqhsLCQH374AQB9y4rvamYMYebtq1evblJ5g7Zs2YLBYCDC++rbmUoFfVqYWwtNaVA2JSVFafXI7aoJJvQAU2tzxfntt986yrQG8fXXXyPLMu2BoCr29wA8MItHJa+rJkhTDwRtDghRqIWVK1eSn5+PycNfGT+oiCE4BpObD1lZWXz//fcOtrD+WFo214VZzwkxtJV5/ddffmkyaRRWrFgBYO428qn+ODn2qheSq/fDp6WlKWMkw6o5xh2JAWXLH3/8cZPr8hO4HkIUaiAzM5Ovv/4agNLWcSBV8+9SqdFHmBvyn332WZMYWzh8+DDHjx9HI8kMbmUtCh0DDIR7GykuKWH9+vVOstB20tLSFK8jU8daulCCQQ6W0ev1fP755w6wrv4sXboUg8FAOyAaqdrjBmN2UT179qwiIgJBfRGiUAPvvfceJSUlGH1CMZYNKFeHoUV7jF5BFBQUsHz5cgdZWH+++OILAAa1LMXXzbrJLklwU6TZbXPFihUuPbucLMu8/fbb6PV6c+RySC1fkMDUzSwca9ascdmEhvv27WPLli1IwE21HOuJxPCy5WVLlzbZmBmBayBEoRp27dqleHSURg8y15Q1IanMxwE//vij4q3kihw9epQdO3YgIXNLdNU++0NblRLobiIjI4Mff/zRwRbazrp168wxIiow9TJRwwv1VUJBbm2eWOjll1+muLi40e2sCyUlJSxZsgSA/kArGwo1AAgFcvPy+L//+79Gta8xkWr7nQkaHSEKVZCXl8err74KgL5lN0zetb1+mjH5tkTfoiMACxcudElPClmWWbZsGQBDWpUS7l11d4ubGia0NVeWn332mUuW5ciRI7z55psAmLqawN/275riTMjuMmfPnmXhwoUu5bmzdOlSLly4gB8w0sbvqJGYgFkTf/755yabekUMNDsfIQoVkGWZV199lYyMDEwefpS27lOn75dGDcDk5kNaWhpvvPFGI1lZf3bv3k1iYiIaSWZiTM2RvcPDSwnzNJKdnc0333zjIAttIy0tjaefftrcbRQuI3esY2XiDqZB5pbFli1b+OijjxrH0Dqyc+dOxTNqAuBhU9PHTBQSg8qWFy5c6PID6QLXRIhCBdauXctvv/2GLKnQtbsB1Nq6nUDjhi52OCCxceNGNm3a1Bhm1guDwcD7778PwOgoHSGeNb8da1QwJdbcWvjmm28qTUXqLHJzc/nvf/9LTk4OcoCMaYCN3UYVCQFTX/P/4PPPP3d6nElmZiavvPIKAAOBDvUo1EigJeb/0UsvveRSLSBbEC0F5yNEoRwpKSm8++67AJRG9sXk06KWb1SNybclpWXeSG+88QaXLl2ym40N4bvvvuPs2bP4aE1MqGYsoSL9Q/W09zdQUlLiEn3Ver2eZ599lpSUFGQvGdMQE2jqfz45WsbUxVxxLlmyhL1799rJ0jraIcssWrSI3NxcWgKj63keDRJTAC3m+SNWrVplPyMbCSEEroUQhTJMJhMLFy5Ep9Nh9AvH0LJ7g86nj+iF0SeUwsJCXn31Vac/+OfPn1e8oqa0K8Zba5s9kgT3dixCJcn8+uuvSg4oZyDLMq+//ro50lqDWRA87XDeLjKmKBMmk4nnnnvOKTPPbdiwgd27d6MBJgPaejV9zLRA4uay5WXLlrl8au3yvw2R5sL5CFEoY8OGDRw6dAhZpUEXM6x2b6PakFTo2l2PLKnZs2cPv/76q30MrQeFhYU888wz6HQ6ugXpuSGibi6mbf2M3NLG3LJYuHAh586dawQra2fFihXmuAkJjAONdRpYrhEJ5L4ycrBMQUEBs2fPdqhbZ05OjtKtNwIIbYAgWOgDxGJuWb3++utOfympifIBdyL4zvkIUQCys7OVH2Vp6z6V8hvVF9nDH31ELwDeeecd8vPz7XLeulBaWsqzzz5LcnIyge4mHu1aWC+9uz2mhE4BeoqKioiPj3f4IObWrVuVe2TqYYJWdr6A2jzwLHvJpKamKiLqCN577z3y8vIIA66z0zklJMZj7lnbt2+fS6cs0ev1yrIQBedzzYuCxdsoPz8fo1cwhpZd7Xp+fasemDz8ycrKUtwnHYXRaOSll15i3759eKhlnuxZgL97/d4YNSr4V49CWnoZuXTpEv/5z3/Iy8uzs8VVc+jQIV566SVkWcbUzoTcvoYymIDCso8Fy3ptY64e5i4pWStz8OBBXnnllUYfqP3jjz/4+eefkTB7G6nt0EqwEITE9WXLb731Funp6XY7tz0pLwSi+8j5XPOisGrVKv744w9kSUVpzLDqU1nUF5UaXcz1gMTmzZsdlrZZlmVee+01tm7dikaS+XePAmUCnfri5yYT37uAADcTZ8+eJT4+vtEDvy5dusScOXMoLS1FbiUj95Jr9jQqBvUGNepNV9OkqjepUW9Qgy2m+pe5qqrMrqoff/xxg8tQHWlpaSxcuBCAQUCkHQXBwhAgAnNG3Pnz57tkdHp5IRCi4HyuaVHYsWMH7733HlAWX+BdcRqTcsgmJF0+kq5A2STpCpB0+SDX/DZp8g1V4h3eeOMN9u3b13Dja+GTTz5h3bp1SMj8s3sh3YLt82ML9TQxOy4fH62Jo0ePMn/+/EZr8hsMBp5//nlyc3PNrqcDTY55YkPB1Md8Tz/99FMldbU9yc/PZ/bs2eTl5RFBzUFqRmSykckpty0HyEbGSM0tPw0SkzFnUj106BBLlixxufGF8kJQvitJ4BycIgqJiYksXryYxYsXM3nyZKsEcjXtsydHjx5l3rx5mEwm9C06Kumvq0MqLcTrwAq8Dl1Nuex16Fu8DqxAKq09k6g+vCeG4HYYjUaeffZZzpw50+AyVMe2bduUN9zpnYroF2rfH1prHxNP9ixAq5LZsWMHH374oV3Pb+G7777j2LFjyFoZ0+CGuZ7WFTlaxhRrFoZXX33VruMLpaWlzJkzh3PnzuEH3IW58q6OPOB14J1y294p22ZLB15wmZuqCnO0s6sE6lko33oRouB8nCIKCQkJxMfHEx8fT79+/RgxYoRN++zFxYsXefrppykpKcHg35rS6MEN9zaqDUlCFzMUo29LCgsLG22w9uLFiyxYsACAMZEljGjdON0FHQKMPNTFLIZffvllpRncGkpJSYmStE/uIYOXXU9vE3J3GdlT5vLly3YLbJNlmYULF/LXX3/hDkwD/Bqh26gi7csGnsGctsSVsqmW74J0tTxU1yIOF4XExESl0gKYNGkSiYmJJCUl1bjPXuh0Op599lmys7MxegWha3+jeWYZR6DSUNJhFCYPfzIyMpg7d65d+1AtsRaFhYW09zdwZ/uaf2BGE2QUq8gsvlr+zGIVGcUqjDaMrw5qqWdUa7Or6oIFC+zqXZWQkGCOWPaSK8+37Cg0IHcyX/vbb7+1y6DzN998Q0JCAirgTqClAwTBQt9yA89Llizh2LFjDrt2TRQUFFS5LHAODheFuLg4q9TSlu6hoKCgGvfZi2XLlnHq1ClkjQe6DjeB2s1u57YJjTslHW9CVrtx5MgRuw5kbtiwgf379+Ouknm0ayHqWu5ulk7FzO3+zN511eF/9i5/Zm73J0tn26NxZ/tiWnkZycrK4oMPPmiI+QqyLPPdd9+Zl9vJTh35kqNlZI3ZTbWhYwsnTpxQkhHeDLRzoCBYuBHzXM8Gg4F58+a5xJt5+S7ipjAXSXPHKT+3SZMmKcsrVqxg5MiRBAQE1LqvPDqdjry8PKtPbSQlJSnJxnTtrkd2925YQeqJ7OGHru1QwDzdYmpqaoPPmZubq1TKE9sVE+rlmJw3bmp4sKwb6ccff+TIkSMNPue+ffs4ffo0slpGjnHyoKimbG5naFBSQMu8D0ajkS6YU2I7AxUSt2KO+7t48aIyiZQzKR8omJ2d7URLBOBk76OcnBxWr15dZX6WmvaBubvC399f+URGRtZ6vS+++AKTyYQhMBpjQO3HNybGoGgM/hEYDAa7/DA/+ugj8vLyaO1t5KZIx84T3THAyNBW5mu++eabDe5m+eSTT4CyytjBDbmqkDuYWyuJiYkcOnSoXuewfFeLuZUgOaGVYMETiTFly6tWrXJqWnRZlq3G1hwV+yKoHqeKwqxZs9i8eXOVLYGa9gE8/fTT5ObmKp/a8rvk5uYqeXssUcZORZLQh/cCzPMFN+SHefr0aSX+4b5ORbV2GzUGd8QW46mWOXHiRIMGMQ8ePMjBgwdBdbU/3+l4gamNWeg+++yzep3Ccn96A/5OFAQLXYBgzClQnJmCpaSkxGoQv6nMCd6ccZooLF68mFmzZhETE0NOTo5VX2JN+yy4u7vj5+dn9amJHTt2YDAYMHoF2TxpTmNj8m2Jyd2PkpKSenvvyLLMO++8g8lkon9oKZ0DnRP84+8uc1uMuX96+fLl9f5xr1mzBiirhO2Q7M5eWARq9+7dXLx4sU7f1el07Ny5E4A4u1tWP1RI9C5bdmaSw4o4IxWMwBqniMLq1auJi4tTKv2VK1cqLYKa9jWEXbt2AWAMbNPgc9kNSVLsqa8o/PHHH+zfvx+tSq7V26ixGR2po6WXeVKeL7/8ss7fNxgMyoxhln58l8EH5BZmm+o6q9nBgwcpKSnBDwhvBNPqS+eyvwf273dYnqfaEKLgfBwuCklJSUyePJlRo0YhSRKBgYHMmjWr1n0NQZZlc7plwOgX0eDz2ROjv7maqM+czgaDgaVLlwIwJqqEFrVMmtPYaFQowrRy5co659o5e/YsRUVFyFoZ7OdwZjfkMLMo1HVcwXJv2+HcsYSKtAD8gFK9nqNHjzrbHACXmcjpWsbhohATE4Msy1Yfi8dBTfsaQmZmJllZWchImHxco+vIgtEnFDDnwanrW9LmzZtJTk7GR2tivI2T5jQ2cSF6OgboKS0t5fPPP6/Td8+ePWte8Kd+M6k1MnKAWRTqmjr8xIkTALS2t0ENREJSbLLY6GgqusRevnyZkhLXeJavVa6J3EcXLlwAQHb3BZUDcyXYgsYdk9bceZ6Wlmbz14xGozLoOT66BC8XKZYkweR25h/1+vXr69RasDgLyH726ToaP348X3zxBePHj0eSJNsS4tVE2bBVampqnfI9WZ6/+s3j17hYbKrLs2dPTp06pSy7a7yQZZmTJ086xRaBmWtCFCytDdnNCbkSbEDWmu3Kysqy+Tvbt2/nwoUL+GhNjIhwjf5gC50CDXQM0GMwGJS4EFs4f/68ecE+01kwdepUoqKimDp1qjkJXEM9L71AVsno9fo6TbFqcZTwbeDlGwOLTXV59uyJpVsXINjH3JW6f/9+p9giMHNNiILF3VNWa+1yvopvoFJpA2ubMrvqEl36448/AnBDhA4PF2kllGdslFmofvrpJ5tTeVjeVmVf+7QUVqxYQXJyMitWrDC3FBr6TiChCJbl7b82ZFlWnj/3Bl4eKj97DR2WtdjkrFiF8hmDQ32jABolK63Adq4JUVAqJTvNlVDxDbR8Ou36IKvMuf9tzRBZXFys/JiGtnK9/PgAvUL0+GhN5OTk2JxjRwlcskftCaxbt4577rmHdevWmVsK9nBxLbPN1iAro9GopKpW13KsLVR89nIaeD6LTc7ITlrx2QjzjwbgyJEjDp0OVWDNNSEKygMv2eNnWfkNtMHTd5aJla1v1KdPn8ZgMBDobqKVg9JZ1BWNCjoFmMtjqyhotWUtOTtNz2CpjO06f0CZbYqttVC+srVHg67isxfQwPNZfhHOmNzGErthwdvNF3/PEEwmk+JCLnA814QoKIOCdkqPXfENtMFjFWWiYOvg5ZUrVwBo4Wm0W8bvit0SObqGn9jiImurm2F0dDQA0hUXdD0CMIDl1bxt27Y2faV8yg97lKris9fQcQpLBeCMuZH/+OOPStvCA2Kr3SdwDNeEKEhKzWmfN0a7v4GWzdymqmMKb3tWnRW7JTJLGv5o1FWwBg4caP7eWan2+ZSdgHReQjJJhIeH25RrC8o/e/Z5+uz97Fn+zXV99hpKcXFxlWMHEYHtAfhz95/CNdVJXBOiYImIlkqdnya4KiS92S5fX9ve+ywpPfJK7Xf7KnZLhHg0vFbOKzVXiP7+/rUcaWbkyJH4+/sjFUpIZ1ystaAH6ajZpttvv92qsq8JT09PpavJFWcKsNhUW5oYe7Nr1y50Oh1ebtbXDfQKw8vNj+KSYtGF5CSuCVFo164dAKrCDDC52MTgxlJURebuoNjYWJu+0qpVKwDSi1UY7PRGXbFbIsC94W+iFwvNPdbh4bYld/D09OTBBx8EQHVIRYNda+yItF9CKpFo3bo1t956q+3fkyTl+TvXOKY1iDInYJufPXuxefNmAFoHdLDaLkkSkUEdrY4ROJZrQhTatm1LaGgoksmAJuNU7V9wIJr0E0iyTFRUFBERtqXgaNmyJT4+PhhlieQC+wye27tbQm9Csa0uFc748ePp3bs3GEG1W+US3UhSsoTqvAqVSsXTTz+Nm1vd8nlfd911APwJyHbqwrQH+chYkltYbHQEWVlZyiBzZEinSvvbBHcFzAPRYn4Fx3NNiIJKpWLq1KkAaC8kgsE1+iql0iLcLhwA4I477rD9e5JE9+7dATia5YJBCsCpHA16k0RAQIDN/e9gvldz5szB19cXKVtCOuzkbqRCkBLNNtxzzz3K/70u3Hbbbbi7u3MB+KvWox3HRsxj5127dqVbt24Ou+6aNWswGo0Ee7fC3yO40v4ArxYEebfEYDAoKccFjuOaEAWACRMm0KZNG1T6YtzP7QB7uinWB1nG7ewfSEYdHTp0YMyYMbV/pxyWQdnd6S4wC00V7Ek396MPHDjQ5v53C2FhYcyePRsA1QkVOCtHmgyqP1VIeomuXbsyffr0ep0mICCAe++9F4ANQK4LtBaOIfMX5heMxx57rM73qL4UFhYqU622D+tT7XGWfatXr3bqJEDXIteMKLi5ufH000+jUqnQXElCk37cqfZoLh1Gk5OMVqtl9uzZaDR1e+MfPnw4Wq2Ws3kaTubYpwvJXhTqJX6/aI7yGjlyZL3OMXToUMaNGweAaq/KbrELdUFKkpAyJTw8PJg7d26d71F57rzzTjp16kQxsBIwOFEYspH5rmz5jjvuoGvXrg679ueff05ubi6+HkG0Lhs7qIrIwI74egSSm5tbrzTsgvpzzYgCQJcuXXj44YcBcDu/E1W+7flr7Ikq9wLuyWZ3vMcee6xeg3yBgYGMHj0agC9OemFy/sunwsozHpQYJWJiYujXr1+9z/PPf/6TkJAQpAIJ6ZiDu5GKQTpkvuaMGTNsHiyvDo1Gw7x58/Dx9iYZc4vBGeMLpch8CZRg/j1YBvYdQVJSEitWrACgR+thqGrIMKBSqenR+nrAPI+5kkFX0OhcU6IA5jej4cOHI8kmPE4mIJU4dk5YqSgbj1O/ADJjxoypkydLRR588EG8vb1JytPwfZKH3WxsCPsztPySarbliSeeaFC3hI+PD0888QQAquMqGpzTwVZkUCWau406d+7MbbfdZpfThoeHM/e555AkiT3Azlq/YV9MyKwELgNBQUG88MILNkdmNxSDwcArr7yC0WgkPCBWiUeoifCAWFr5t8NgMLBgwQKnRF1fi1xzoiBJErNnz6ZDhw5IhhI8TmwEvWMGnqXSIjxObEQyltKtWzf+85//NKjSDA4OZubMmQD8cNaDvemO+YFXR1qhivePeAMwceJEsxdRAxk2bBjDhg1T+vdr7UbyBOPNRoyjrx5oHG3EeLPR5txH0nkJKU1Co9EQHx+PWm2/7rnrrruORx55BICfgaMOai3IyGwATmBO0fHSSy8RGhrqkGuDudvo5MmTuKk96NNmlE3fkSSJPtGj0KrdOX78uOhGchDXnCgAeHl5sXDhQsLCwlCV5OJxKgFMNdc2sps3Rb2mUtR9orKtqPtEinpNRXbzrv2iJgPuJzejKi0gIiKCV155BXf3hmd+Gz16NLfddhsyEu8f9iYpzznjC7mlEksO+FBkkOjWrRv/+Mc/7HJeSZJ48sknCQwMRMqVkP6qRURVgHfZx4Jl3ZanPQ9U+80H/v3vf1diDOzJHXfcwYQJE5CBVUBqLcLgBzwJPF5u2+Nl22wNOdsF7Mb8/5w7d65DvY2OHj3Kp59+CkBcm5F4utmeK8zLzZe4qBEAfPLJJxw/7tyxwGuBa1IUAEJCQli8eDHe3t6o8y/hdr6W6ElJhezua5X8Tnb3MU/cU1v21TJPI3VhBn5+/ixevNgu805bePzxxxkwYAA6k8RrB3zIKHbsbdUZ4fUDPqQXq2nVqhWvvPKKXbslgoKCmDNnDpIkoTqjQjrfSOMLelDtUIEB+vTpw5133tkol5EkiSeeeIKBAwdiAL4GCmoQBjUSgVgnvwsAApFQ25Ds5AwyP5UtP/LIIwwfPry+pteZ4uJiXn75ZUwmE5FBnYgK7lz7lyoQFdyF1oEdMRqNvPTSSyL9RSNzzYoCmIPanivr49WmH0N9pXEGszSZp9BmnkalUjF//rw6+e3bdP6yQczY2FhyS1W8esCHQr1jBmZNMrx/2JszeRp8fX159dVX7Sp4FgYMGMB9990HgGqfCuwd0ySDao8KKV+iRYsWPPfcc3btNqqIRqPh+eefJyoqijzgOxpn4LkQmdWY8y6NGTOmTvEw9uC9994jJSUFT60PcW3q54lm6Uby0HqTnJzM+++/b2crBeW5pkUBzH28d999NwDuZ/9o+IQ5FZB0BbidMw8pPvDAA/TpU71vdkPw9vZm0aJFtGjRgrRCNa//5U1pLf3vQe4m3hicy8KBV3PXLxyYyxuDcwlyrz2UWJbhi5Oe7M1wQ6vV8sorrxAVFdXQolTL9OnTzfEZRlDtVIEdp5KQTkpIFyS0Wi0vvvgigYGB9jt5NXh7e/Piiy/i5ubGKWCPnc8vI7MWc36j6OjoBo9h1ZWdO3eyZs0aAPq3HYu7pv4TWrhrPOnfdiwA33//vciL1Ihc86IA5r7j9u3bIxl1uJ3bYb8TWwLUTHq6devGXXfdZb9zV0GLFi2ULrETOVo+OOJdo6uqWmVObx3ieVUAQjxNtPA0obbhyVh/3p1NKWZPozlz5tCzZ8+GFqFGVCoVc+fOpWXLlkiFkjl+wR4v15lluZaAf/3rX3Tp0sUOJ7WNtm3bMmPGDAA2Yd/AtiPAUUCtVjN37ly7jGHZSk5ODosWLQLMgWiWCXQaQkv/trQPjQNg4cKFYiKeRkKIAuam/OzZs1Gr1Wiyz6HOOmeX86qvnEGTm4pWq2XWrFmN2h1hoV27drz88stotVr+THfjy5P2mG6sMjsuafnmtHkeiccee4wRI0Y0ynUq4uvrywsvvIBGo0G6ICElNfDNV1+WY0k2B9r97W9/s4+hdWDixIl069YNHbAW+3QjFSOzvmz57rvvpn372l1A7YUsy7z22mtkZWXh5xFMj9bD7Hbu7pHD8PMIJisri9dff91u5xVcRYhCGe3bt1cGFt3O7QCDrmEn1Bfjft7cbTR9+nTatGnTUBNtJi4ujjlz5gCwMcWDzSn2fUM8maNmWZnr6ZQpU5gyZYpdz18bnTp1Ut6uVX81LJuqtF9CKpJo1aqVw7tXLKjVauLj49FoNJwEDtnhnD9j7jaKiopSUmw4is2bN/Pbb78hSSoGxIxDrbJffi6NSkv/mJuRJBVbtmwhISHBbucWmBGiUI777ruPyMhIVPoi3FIa1sPrfn4XkkFHu3btGs2LpSZGjBihVJxfnPTkdK59Wil5pRLvHPLBIEsMGzbMbq6ndWXKlCnm8RljWfxCfbKpXkDJfvrss8/i7W2Da3EjER0dzbRp0wBztHNRA1oLScgkli3Hx8fXOatrQ7h8+TJvvPEGAF1aXUegd5jdrxHk3ZIurcy5v15//XUuX75s92tcywhRKIe7uzv//e9/AdCmH0eVd7Fe51HnpKC5cgaVSqW8ATqDu+++m+uvvx6jLPHBkdoHnm3hk+NeZOtUREVFMWfOHIfP2GVBpVIxe/ZsvL29kbIkpBN1fMMvKfNiwpyXqD7ZT+3NPffcQ9u2bSkExYW0ruiRWVO2fOutt9KjRw87WVc7BoOBefPmUVhYSLB3KzqHD6z2WJNsolCXS6HuakaBQl0ehbpcTHLtCt+51UCCvFtRUFDACy+8IKKd7YgQhQr06tWL8ePHA2ZvpNqC2iphNOB2bjsAkyZNonPnuvtl2wtJkoiPjyckJIRLRWrWnGtYKoz9GeZxCpVKxXPPPYeXVwPnpm4gYWFh/Otf/wJAdVQFdchYIu2XkHTm/Ez3339/I1lYN7RaLfHx8UiSxAHMb/x1ZSuQhdnpwJLny1G8++67HDlyBK3anQExt9SY26i4NJ/1B5ex8cjHyraNRz5m/cFlFJfW3h+oUqkZGDMOrdqNQ4cO8d5779mlDAIhClXyyCOPEBQUhKokF+2lw3X6rjbtACpdAaGhofz9739vJAttx9fXV8kftOG8B+n1DGzTm+DzskHrqVOn0qFDh1q+4RjGjBnDoEGDwITt3khpoEqt/6Q5jUnXrl2ZMGECAD9St2yqGchsL1t+4oknHNod9uOPPyopsfu3vRkfj4BGv6aPRyD9294MmFNsr1u3rtGveS0gRKEKfH19efTRRwHQXtiPVFpo0/ekkny0Fw8CZtdGZ79JWxg2bBh9+vRBb5L45lT9vJE2pbiTXqwmKChICSJzBSxpMDw8PJCuSEjJtXQjmUB1wPzYT506lY4dq0/f7CweeughAgICyMScmsJWfsKcGmrgwIEMHTq0cYyrggMHDiieQN0ihhAR6LipPSMC29M1fDBgHl84ePCgw67dXBGiUA2jR4+ma9euSCYD2tR9Nn3HLWUPkmyiT58+Dv1R1oZlIhWVSsWf6W4cz67bGEduqcQPSWYxmTFjhsuInYXQ0FDFw0Y6ItU46CydlZAKJZcTt/L4+voqTgJbsW3Q+TQypzC7Vz/++OMO86K6cuUK8+bNw2g0EhnUic6tqh9HaCy6hF9H68AOGAwGnn/+ebKyshxuQ3NCiEI1WCpSAE3GKaSimvMqqAoz0WQlIUkS//znP53i2lgT7dq145ZbbgHM3UB1mX/h2zOeFBslOnbsWOcZ4hzFxIkTCQgIQCqUkFKr+d/LKAPS06ZNczlxK8/YsWNp164dJcDvtRxrQsYyxf1tt91m9zQq1SHLMgsXLiQrKwt/zxD6RY9xynMvSRL9247FzyOYK1eusGjRIrvNNX4tIkShBrp27crQoUORkHG7sL/GYy2tiVGjRtVr0hxHYJl/4Xy+hj8u2taPnlKgYssF87GW1oYr4unpqcx7IJ2upmK6BFKhhJ+fnzKrm6uiVquV1sJuak6YdxxIw/w/sLi1OoKtW7eye/duVJKa69r9DY3aeanbNWo3rov9GypJxc6dO9m2bZvTbGnquOYv3IWweKaos5KQSqoOq5eKstDkpKBSqVy2SwKs5wpedcbTJhfVb055ISNx/fXXN3oai4Yyfvx4VCoV0hWpSk8k1Vnz4z5mzBg8PFxjUqKaGDhwIJ06dUJP9RPyyMhYqr9JkyY1SjLCqjAajXz00UcAdGrVHz/PYIdctyb8PUPo2HIAAB999BEmU32CVwRCFGohNjbWPPk8oL18tMpjtJeOAHD99dc7rOleX26//XbCwsLI1qlISK050vlEjpq/rmhRqVTKW6srExISwoAB5kqhUhdSKUgXzdtuvvlmR5tWLyRJUpI17gX0VRyTClzA7M46adIkh9m2Z88ekpOT0ard6diy/lOu2puOLfuiVbtz7tw59u7d62xzmiRCFGzg9ttvB0CTebpy3IJRj+ZKEoDdpm1sTNzd3ZXWzLpzHpTUEPOz+ox5cHncuHEuL3YWRo0yz+pVURSkC+YB6JiYGGJiYpxhWr0YPHgwLVq0oAg4VcV+S+TyjTfe6JDMrhZ++skcXhcd3BWt2nGJ9mrDTeNBm2BzQsMNGzY42ZqmiVNEITExkcWLF7N48WImT55MTk5Opf2NlWK6PvTr14/g4GAkgw51XprVPnVOMpJJT3h4uMt3r1gYM2YMERER5Omrby0cy9ZwLFuLVqt1eO6chjBo0CC0Wi1SYQVRSDOvO3KCGXug0Wi46aabgMo5kQzIWKJoxo4d6zCbioqK2L7dHBERHeK4GdxsJTq4KwDbt2+nqMi+qfCvBZwiCgkJCcTHxxMfH0+/fv2sMmyuXr0aMAuDq6BWq5XKRJ193mqfpmxinhEjRricx1F1aDQapaLfkOyBvoqxhTVnzX3u48aNIyzM/vlrGgsvLy/i4uKsNxoxz1aP+c27qWH5fZypsP0sUIJ5ZjpHvpD8/vvvlJaW4usRSICX4+Z5tpVA75b4uAeg0+kU8RLYjsNFITExkQULFijrkyZNIjExkaSkJGW90o/aBbj++usB0OSkXN1oMqDOTQVwqbgEWxg1ahQtW7Ykr1TFzsvWnkjn89UczjKPJTgjmV9D6d+/v/WGLJCM5tgEV/UMq4mYmBgiIiKoqN3Hyv4OHTrUIWnZLfz2228ARAZ1cskXIUmSiAzqBFy1VWA7DheFuLg4li9frqxbuo6CgoIcbUqd6NatG35+fkjGq9N9qfMvIZkMtGjRwiUjY2tCo9EwceJEALZU6EKydCldf/31tGrVyuG2NZS+fftarUsZ5oorLi7OJSux2pAkqVILR8bsigowZMgQh9liMpnYt8/sfh0RYJ85GsaPH88XX3zB+PHjkSSJ4tKCBp8zIsAs/vv27RMxC3XEKd1H5b0kVqxYwciRIx3mSldfNBpNpTdQde4FwOw62BQrm7Fjx6LVakkpvBrhbDDBrkvmloMlB09TIzo6Gj8/P2VdyrwqCk2VQYMGWa2nY55GwsPDg169ejnMjoyMDIqLi1FJKvy9WtjlnFOnTiUqKoqpU6ciyzJFpXXIbFgN/l6hSJKKwsJCMjIy7GDltYPNovDhhx/a/eI5OTmsXr2aVatW1fm7Op2OvLw8q09jU0kUylJrV+quaCL4+fmZ5zwux9ErWoqN5snrHVnZ2BNJkujWrdwAaFkwutW2Jka3bt2sptM8W/a3Z8+eDp1m0/LyIwMS9nkRWrFiBcnJyaxYsQJJkvBy86v9S7UgSZJ5EnFw2YBLV8XmJDgLFy4kJyfH6o1elmVyc3OJi4vjxhtvrPPFZ82axebNm+vVSliwYAHz58+v8/caQsVKUlUWzNZUvI6qYujQofz++9VECgezzI/E4MGDm/SPqWPHjuzYYZ5vWzJJeHt7ExUV5WSr6o+bmxudO3fmwIEDAFhGthz97AUFBeHt7U1hYSGZBam08G24q/K6dev48ccfkSQJWZbxdPNp8Dkz81ORkfHx8XH5XghXw2ZRiIuL46mnnqpy39mzZ1myZAkzZsywarbXxOLFi5k1axYxMTHKuEJdbt7TTz/Nk08+qazn5eU1ui99WFgYgYGBZGdfzYMUERHRpB+6iv3vx7PNqQr69XOdgKT6UDEWoV27dk1a5MA8DalFFCyO0V27dnWoDRqNhuuvv54NGzZw4tJeu4iCpc/fnn3/Jy6ZZ0684YYbnDbJVVPF5l/JokWLqt3Xtm1bnnrqKZYtW2bTuVavXk1cXJwiCCtXrqyyYq0Yv1Aed3d3/Pz8rD6NjSRJVVY2TZmQkBCrweRLxWYvFkfO2NUYVGwVOHKO7Mai/LNm6Sx1hjfVXXfdhUqlIi3nNJn5qQ6/fm1k5KdyMTcJlUrFHXfc4Wxzmhw2i0Lbtm1rPcaWN+akpCQmT57MqFGjkCSJwMBAZs2apexPSEhQ1hcsWKDELbgKFVsjzaGyqeg51apVK/z9/Z1kjX1o0cJ6EDQiIsJJltiPis9eUFAQvr6+DrcjKipKSRVy6MIfDr9+bRxKNWeDuuWWW5pMJL4rYXO76sCBA7UOPJ45UzG8pjIxMTE1NhNHjhzJyJEja2yZOJPw8HCr9eZQ2VQUNlteAFydin77LVu2dJIl9qOie7Azn73p06fz008/kZGfQnbhZQK9XSPAMavwEpkFF9BoNEyfPt3Z5jRJbG4prFixgvz8/EoeP+fOnePXX3/l0UcfbfJdKbZQsXJpDpVNxcqldevWTrLEfnh4eFh5G4WEhDjRGvtQMbOrM5+90NBQbrjhBgCSMv5ymh0VsdgyfPjwZnHPnUGdxhQCAgIIDAy0+sTExDB58mT69OnDgw8+2Ji2ugShodZh/c1BFJpjmSRJsgqIbMrOANXh7PtkmZMiOesY+nJBnc6i1KAj+Yo5pM8yoZSg7tgsCjNmzOD06dNkZWVZfbKzs7ly5co1IQhQ+S26Yt91U6TiG1VzecPy8bnq2uiMvnd74+HhYRW9XFHMHU3v3r2JiopCbyzldHrNk1A5gtPp+zGYSomOjqZ3797ONqfJYrMoPPzww7Rt2xZ/f/9Kn2uJipWLVuu82absRcW+6uYiCuVx5ak3bUWSJKvkhM6+TyqVinvuuQeAY2k7KSrNd5othbo8jl/cBcA999zTJDMMuAo2i0J2djYffvghH374oVX08C+//MLy5ct55JFHeO211xrFSFeiOT5sWq3WStxdPQ9VfXBzs236UVenfDeYK3SJjR49mq5du2Iw6fkzaQMm2fGznZlkE3+e3YDBpKd79+7KnBrOZtWqVSxYsKDJzQBnsyhMnjyZvn378uCDD1rFBIwYMYKHHnqIhQsX8sorrzSKkYLGp/wgpitUNvamuYh5+S4xR8Tm1IZKpWLOnDl4eHiQnp/M0bQdDrfhyIXtZOSn4OnpyZw5c1zmXr/zzjv89NNPSgLBpoLNovDQQw8pLqnnzp2z+oC5InnooYcaw0aBAyj/Q/L09HSiJYKaKN8N5ir3KTIyUsl2cDRtJ2k5px127QvZpzlW1m303//+1yVdxPPzndetVh9sFoXg4KsTc2dnZzN58uRKgWXXgkvqtYCrvGkJKlM+/sKRifBqY/To0cp0tLuTNpBfkl3j8Z5uvozrMYObut6vbLup6/2M6zEDTzfbnALyS7L48+x6ACZOnMjIkSPrab2gPDaLQvkuhd69ezNlyhSeeuopoqOjle2iMhEIHIerjZM89thjdO/eHb1Rx47TP2Aw6qs9ViWp8Hb3x9v9aheYt7sf3u7+qKTaqyWDsZTtp9egN5bSo0cP/vnPf9qlDI1BU6sXbRaFpKQkq+A1SZIqBbPZEtEsEAjqT/kKxtU837RaLfPnzycoKIjc4kz2nd/UKBPcyLLM3nObyCvOJCgoiHnz5omkd3ak3sFr8fHxVusBAQEsXry4MW0VCATlcMWsryEhITz//POoVCrOXznaKNHOZzIOkJx1DJVKxfz5853umtvcqHfwWnZ2tlUQ2+nTp8VAs0DQyDSFqSV79+7NjBkzANif/CtXCi7a7dxXCtI4kPwrYI6dctW5TMq7oTaFe1YeuwSv+fv7ExMTw8MPP9yYtgoEgibCnXfeydChQzHJRnacXkOJvrDB5yzRF7Lj9BpMsonrr7/epdNiGwwGZ5tQb2wWBVvCxkVouUDQuDSVQUtJkpgzZw5RUVEU6/PZdWZdgwLbTLKJnWd+pFhfQJs2bXj66add+n9RXhRc2c6qaHCn5Lfffsurr77K6NGjGTNmjD1sEggEzQBvb29eeukluwS2lQ9Qe+mll1w+bUl5UWi23UfVMXHiRP773/+ycuVKTp92XNCKQHAt0tQqmOjoaOLj4wFzYFtm/oU6nyMjP1UJUIuPj28SE1td06JgISAggEmTJtnrdAKBoJkwcuRIbrrpJgD2nPsZo8lo83eNJgN7z/0MwNixYxkxYkSj2Ghv9PqrMRpGo+3ldQVsFoXvvvuu1mOcMV+sQFAdTa0v1xaaapn+9a9/ERQURH5JFqcu254L6OTlfeSXZBMUFMRjjz3WiBbal/KiUFrq/Lkm6oLNorB58+YqZ14TwWsCV6WpNdttoamWydfXV/FOPHZxJzp9Ua3fKdEXcSzN3G306KOPNqk5McoLQbMVhaVLl1Y585oIXhO4KpbpOEVwk2tw00030b59e/TGUmWMoCaOXdyFwVRKx44dXSYdtq2UlJQoy8XFxU60pO7YHBs+Y8YMZs2aVW2u/StXrghRELgUY8aMIScnh759+zrbFLvRVLuPwByB/fDDD/PUU09xOv0AHcL64qbxqPLYQl0uZ9IPAOYYKVeM3q6JgoKCKpebAjaLgiV4rTr8/f1F8FoTxpUybtoLjUajzAzWXGiq3UcW+vXrR69evThw4ABHL+6kV+QNVR53NG0nJtlI7969m6So5+TkKMu5ubnOM6QeiOC1emCZp7ljx45OtsR+PPLIIwBMmzbNyZYIaqIp9atXhSRJPPDAAwCcyzxMUWnlt+giXR7nrhwBaLJzv1+5cqXK5aZAndpkv/76K48++ihbtmwBzJPtjB49muDgYP7xj380ioGuyLPPPkvv3r0V/+vmwODBg1m9enWT/RFeK/Tv35/rr79eEfGmSM+ePenZsycm2VRlwrzT6fuRZRO9e/eme/fuTrCw4aSnpyvLzVYUvv32W0aOHMmePXuYNGkSBw4cYMaMGQQEBDB79mxOnz7No48+2pi2ugxdunThrbfeon379s42xa6EhoY26T7rawG1Ws2LL77IXXfd5WxTGoQlpul85hGr7SbZxNnMw1bHNEXKC0FGRoYTLak7No8prFy5kuzsbPz9/cnJyVEGni3BJP/973+ZMmVKoxkqEAiaD4MHD8bPz4+8vDyr7Rl5yegMRfj7+3Pdddc5ybqGk5mZqSxnZ2djMBiazJwPNrcU+vbti7+/P2COXh41alSl6MJ+/frZ1zqBQNAs0Wg0DBkypNL2tBxzrNOQIUOaTCVaFRVbB02pC8lmUQgMDLRaj4mJqfUYgUAgqI4BAwZU2nYp71y1+5oKsiyTlZVlta18y8HVsVkUzpw5Q35+vrJelZuViGgWCAS2UtFbsVhfSH5JFpIkERcX5ySrGk5OTo5VmguAy5cvO8maulPn6TjVajVqtZpJkyYpy2q1GpVKJYLXBAKBzQQEBFhlPL1SkAZAu3bt8PPzc5ZZDSY1NbXStpSUFCdYUj/qNB1n+ek3K07HmZWVJabjFAgEdcKSigTgSqFZFJqqG6qF48ePV9p28uRJJ1hSP+oU0WwZaK7pGIFAILCVzp07s379egCyCy4B0LVrV2ea1GD27aucBXb//v1NxgNJRDQLBAKnUT4rQHaROeCrU6dOzjKnwRQWFrJnzx6rbRrJnP9o//79TrKqbjStLFMCgaBZERERoSzLmPDw8FDSyDRFtmzZgl6vx0t9tWoN9TLnFdu4caOzzKoTQhQEAoHTUKvVVutRUVFNLiOqBVmW+f777wEI97ma/TXc2xMwC0Z2drZTbKsLTungSkxMJCEhAYA9e/awfPlyAgICAEhKSmL16tXExMSQlJSkpNIQCATNn/Ith6bGvn37OHXqFCoJwr08OZ1rnkjIT6vB301Dbqmeb7/91uXzizlFkhMSEoiPjyc+Pp5+/fpZRUZPnjyZ+Ph4Jk2axKRJk4RHk0DQjPHw8OD2229X1sPCwpxoTf2RZZn//e9/ALT29kRbrvtIkiTa+nkBsHr1apdPpe1wUUhMTGTBggXK+qRJk0hMTCQpKYmkpCSrY2NiYpQWhUAgaH5IkkRoaKiyXt0kXq7Ojh07OHz4MCoJRQDKE+rpjq9WQ1FREZ9//rkTLLQdh4tCXFwcy5cvV9Ytk1EEBQWRkJBQ6aEICgoiMTHRkSYKBAIHUt7VvTa3d1fEYDDw/vvvA9DG1wsPjbrSMZIk0T7AG4Dvv/+etLQ0h9pYF5zSfVQ+Je6KFSsYOXIkAQEBVrMVladiHhEAnU5HXl6e1UcgEDQ9vL29lWUfHx8nWlI/1q5dS3JyMlqVVGUrwUKIhxtB7lr0ej1Lly51oIV1w6nD/Dk5OaxevZpVq1bVelxFFixYgL+/v/KJjIxsJCsFAkFj4unpqSw3tWlhs7Ky+OijjwCI9fdGW4PnlCRJdAw0i96WLVvYu3evQ2ysK04VhVmzZrF582bFuyggIKBSqyArK6tK76Onn36a3Nxc5dOUcosIBIKraLVaZdnNzc2JltQNWZZ5/fXXyc/Px1erobWPZ63f8XPTElV23OLFiyksLGxsM+uM00Rh8eLFzJo1i5iYGHJycsjJyWHkyJFVHlvVxN3u7u74+flZfQQCQdOjvCiUX3Z11q5dy7Zt25CAbsG+qGyctbB9gDeeahWXLl1iyZIlyLLcuIbWEaeIwurVq4mLi1MEYeXKlQQEBFSaoyEpKYm+ffuKOAWBoBlTPh9QU8gNBHDs2DHefvttADoE+ODnZruYaVQqeoT4IwG//PIL3333XSNZWT8cfgeSkpKYPHmy1baAgABmzJgBwKpVq5g1axb9+vVjz549tY43CASCpk35LqOm0FLIyclh7ty56PV6Qj3daONbe7dRRQLctXQI8OFETgHvvvsuHTp0cJnssA4XhZiYmBqbSzExMSxatAho2hN3CwQC2yjfOnB1UTAajbzwwgukp6fjpVHTLdgPycZuo4q08fUkt1TPpSIdzz33HB9++CHBwcF2trjuNM0kIwKBoNlQPv9RxVxIrsbHH3/M3r17UUsSvUL8a/Q2qg1Jkuga5Iu3Rs2VK1d44YUXMBgMdrS2fghREAgETqV8AjxXFoU//viDzz77DIAuQb74ujW8o0WjUtGrhT9qSWL//v0uEb8gREEgEDiV8kJQ366Yxub8+fO89NJLAET5eBLu7VHLN2zHR6uhW7AvYA7mdXZqHyEKAoHAqZQXAldMm11UVMQzzzxDUVERge5aJQDNnrT08lCioRctWsSZM2fsfg1bcb07IBAIrinKC4ErisJrr71GcnIy7moVPUP8bY5HqCvt/b0J9nBDpzMPPBcVFTXKdWrD9e6AQCC4ZnE1Ufjll1/YvHkzEtAz2A93dc32mWSZYoORYoNR2WZZN9USpCZJEj2C/fBQq0hJSVGS7Dka17oDAoFA4CLk5OTw5ptvAuZ02IEetafg0BlNbEu7wo5LV9P17LiUxba0K+iMplq/76ZW0S3YnJ1hzZo1HDhwoF62NwQhCgKBwKmUH1NwpYHm//u//yM3NxcfrZp2/t61f8FOBHu40bpsOs9XX30VnU7nsGuDEAWBQOBkygezukoeoD///JONGzcC0DXIr9HGEaqjQ4AP7mXdSJ9++qlDry1EQSAQOBVXG2guKChQsipE+XgS4O74KGutSkXnMi+nr776imPHjjns2s6/AwKB4JomMDCQQYMGMXjwYHx9fZ1tDu+88w4ZGRl4atS0D3DepD9hXh609HLHZDLx8ssvO6wbqWmkJBQIBM0WSZJYuHChs80AYOfOnfz0008AdA/2RaNy7hhH50Bfskr0JCcn87///Y9HH3200a8pWgoCgUCAOUhtyZIlAET7ehLo7vwJf9zUKroGXY12PnnyZKNfU4iCQCAQAF9++aW520itItbfdeaKDvVyV7qR3n777UYfjBeiIBAIrnny8vJYvXo1AB0DfVA7uduoIh0DfFBJcPDgQfbt29eo1xKiIBAIrnkSEhIoLi7GR6sm1NPd2eZUwkOjprW3eTKfH3/8sVGvJURBIBBc8/zxxx8ARHh7ulQAXXkiygLaduzY0ajzLghREAgE1zSyLCtxAEEerjvzm69Wg0YlodPpSEpKarTrCFEQCATXNLm5uRQWFgLmuQ3sxfjx4/niiy8YP348kiShMxpr/1INSJKEj8Y890RaWpo9TKwSIQoCgeCaJi8vDwCNJNk1ncXUqVOJiopi6tSpyLJMsaH2hHi1oS3L0mqxuTEQoiAQCK5pLGMI9nb0XLFiBcnJyaxYsQJJkvDU2KG6dUBqKBHRLBAIrmmCg4MBMMoyOqMRdzvNE71u3Tp+/PFHJElClmW7nLeobJ6GFi1aNPhc1SFaCgKB4JrGy8uLDh06AJBeXGq381qCzOwVbFaoN1BoMKJSqejWrZtdzlkVQhQEAsE1z4gRIwA4n1fkMum7K3Iu3zw954ABAxo1caAQBYFAcM0zfvx4fLy9KTQYuVTk2EltbKHYYORCQQkAd955Z6NeS4iCQCC45vHx8WHqHXcAcCa30OVaC0l5hchAXFwcvXr1atRrCVEQCAQCYNKkSfj4+FBoMHK52HVaCyXlWgn3339/o19PiIJAIBAA3t7eTJo0CYCkXNcZWzibX4QM9OzZk549ezb69YQoCAQCQRkTJ07E09OTfL3BJVoLxQYjqQXFAEybNs0h1xSiIBAIBGX4+/szdepUAE5kF2AwObe1cCK7AJMMvXr1ol+/fg65phAFgUAgKMedd95Jy5YtKTGaOJGd7zQ70gpLuFysQ6VS8cQTTzgse6sQBYFAICiHp6cns2bNQpIkUgtLlO4bR5JXqudollmQ7r33Xtq1a+ewawtREAgEggr06dNH8fQ5mpVPho3jC+5qFcPCgxnUMkjZNqhlEMPCg3FX21bdFhmMJGbkYpRl+vbty3333Vf3AjQAIQoCgUBQBffddx+jR49GBg5k5nKlpPYUGCpJwlOjxlNzNc+RZd2WDKzFBiN7L2ejM5po27Yt8+fPR22nXEy2IkRBIBAIqkCSJGbPns2QIUMwyZCYkUuWDcJQX0oMRvam51BsNBEREcFrr73WqOksqsMpopCYmEifPn0qbU9KSmLWrFksW7aMWbNmkZOT43jjBAKBoAyNRsO8efMYOHAgJlkmMSOXvFK93a+jN5nYl5FDkcFIq1atePPNNwkJCbH7dWzB4aKwevVqwCwMFRk1ahRPP/00M2bMYOrUqcyaNcvR5gkEAoEVbm5uvPjii/Tu3RtjmTA0dBa18siyzF+ZuRTojQQHB/Pmm28SFhZmt/PXFYeLwqRJk4iLi6u0PSEhAYCAgADAnONj2bJljjRNIBAIqsTd3Z2XX36Z6OhodEYTBzPz7BbxfCa3kCslejw8PHj11Vdp1aqVXc5bX1xmTKG6rqKqWhQCgUDgaHx8fHjxxRfx8PAgS6fnfH7DXVVzdHrO5JlTYj/11FPExsY2+JwNxWVEIS4ujqSkJGXdIgZZWVlVHq/T6cjLy7P6CAQCQWPSpk0bHn/8cQBO5RZQoDfU+1xGk8zhK+Z6a+TIkYwePdouNjYUlxGFmJgYFi1axLJly8jJyVEEIigoqMrjFyxYgL+/v/KJjIx0pLkCgeAa5ZZbbmHAgAGYZDh0JQ9TPbuRTuUWUGgwEhISwsyZM+1sZf2RZCelArTMW1qRpKQkcnJyiImJITAwkOzsbGWcoTw6nQ6d7mpASV5eHpGRkeTm5uLn59eYpgsEgmucjIwMpk+fTn5+Pu38vYn197babzDJ/JKaAcCI1i3QqKxjFK6UlLI3PQeAxYsXM3DgwHrbkpeXh7+/v93qPpdpKYBZEGJiYpSupLi4uCoFAcwDP35+flYfgUAgcAQtWrRQ3u6Tcgvr5KZqMJmUbqPx48c3SBAaA6eKQsXB5T59+ijbli5dyqJFixxvlEAgENjAiBEjuOGGG5CBw1fybe5GOplTSInRRKtWrfjnP//ZuEbWA42jL5iQkMDmzZsB87hAv379lIktFi1aREJCAllZWUyePJmRI0c62jyBQCCwCUmS+Pe//82+ffvIy8sjpaCYNr5eNX4nr1RPSlmCvfj4eLy8aj7eGThtTMHe2LtfTSAQCGxh7dq1LFmyBI1KYlh4MFqVqtoxhT2Xs8nS6Rk5ciTPPfecXa7frMcUBAKBoKkxbtw42rRpg8Ekk1xD7EK2rpQsnR6NRsOMGTMcaGHdEKIgEAgEDUCtVitTZSbnF1c7tnCuLEhtzJgxtGzZ0mH21RUhCgKBQNBAbrzxRlq0aEGpycTFwpJK+4sMRtKLzRlWp0yZ4mjz6oQQBYFAIGggGo2GW2+9FYDkKmZqSynrVurXrx/R0dEOtKzuCFEQCAQCOzB+/Hi0Wi15pQaruAWTLHOh0CwKt912m7PMsxkhCgKBQGAHAgICGD58OIDVvM7pxTr0JpmwsDCuu+46J1lnO0IUBAKBwE6MHz8egMtFV1PwpJUJxM033+zwqTXrgxAFgUAgsBM9evSgZcuWmMptyy41Z1K96aabnGNUHRGiIBAIBHZCpVIxZMiQSttjY2MJDw93gkV1R4iCQCAQ2JGq5p+varZJV0WIgkAgENiRTp06VdrWuXNnJ1hSP4QoCAQCgR0JCgrCx8fHapurxyaUR4iCQCAQ2BFJkiqNHzSV8QQQoiAQCAR2JywsTFn28/PD09PTidbUDSEKAoFAYGdCQ0OV5fIC0RQQoiAQCAR2pkWLFlUuNwWEKAgEAoGdadWqlbIsWgoCgUBwjRMcHKwsBwUFOdGSuiNEQSAQCOyMr6+vstzUpgcWoiAQCAR2pry3UVPyPAIhCgKBQGB33NzclGWtVutES+qOEAWBQCCwM+WFQKVqWtVs07JWIBAImgDl502QZdmJltQdIQoCgUBgZ8q3FIQoCAQCwTWORqNRliVJcqIldUeIgkAgENiZ8kIgREEgEAgECqL7SCAQCAQKoqUgEAgEgiaLEAWBQCBoRJpa95Gm9kOaBpZ/fF5enpMtEQgEAjAYDAAUFRU1ar1kObe9xEeSm5qMVUNqaiqRkZHONkMgEAicQkpKCq1bt27weZqNKJhMJtLS0vD19W30gZ28vDwiIyNJSUlpchkQq0OUqWkgytR0cFS5ZFkmPz+f8PBwu6TUaDbdRyqVyi4qWRf8/Pya1UMMokxNBVGmpoMjyuXv72+3c4mBZoFAIBAoCFEQCAQCgYIQhXrg7u7O888/j7u7u7NNsRuiTE0DUaamQ1MtV7MZaBYIBAJBwxEtBYFAIBAoCFEQCAQCgYIQhSaOXq/n8uXLSvSkQCAQNAQhCk2YnJwc7r33XqZPn8758+cBcxCfyWRysmWC5sSXX37JlStXnG2GXWmOZQL7lEsMNDdxkpOTeeutt5gwYQLDhg1ztjm1YjQareavbSqYTKYao0VPnTpF27ZtrWbcamoUFxeTkpLCwYMH6d69OzExMWi1WlJTU0lISMDDw4M77rjD2WbWieZYJmjccomWQhNFlmVkWSYqKgoPDw9SU1MxGAyUlpby1VdfsXLlSkpKShxqU15eHp999hmrV6+u9pgNGzZw6dIlB1rVME6ePMnf/vY33nnnnWoTjuXm5rJp0yZyc3MdbJ19+eGHH0hKSiI6Oppff/2VefPmAdC6dWsGDhxIUlISJ0+edK6RdaQ5lgkat1xCFJookiQpOZ5iYmI4f/48paWlrFu3jrfffpuQkBCHv5FnZ2ezZcsWDh48SGZmJoDVWMeZM2fYt28f2dnZ6PX6JjEOYvmR5efnc/nyZQCle85oNAKwa9cutm7dytq1a/njjz8oLCx0mr22UJW47dmzhy+//JIxY8bQt29fRo4cyccff6zs79SpE5MmTapRHF2N5lCmqrqCG7tcQhSaIKmpqZw5c4ZHH32UZcuWMXToUDIzM8nNzSUqKooPP/yQG2+8Ea1W6zCbTCYTbdq0YcaMGYA5YyNYT2B++PBhfvjhB9566y3Wr1/v9MpTp9PVOP5iNBrx8vJixIgR6HQ6Lly4AJjzbBUVFaFWqzEYDKSkpHD27FlOnjxJdHQ0Xl5ejipCrej1enbt2sX69ev5/PPPSU9PrzJhZEREBJcvX6a0tBSA9u3b07JlS7788kulYunQoQN5eXku0dI7ePAg77zzDnBVnCvS1MpUXFzM9u3befvtt5k7dy4ZGRmoVKpKFXtjl0uIQhPAcvMLCwuZO3cuo0eP5rfffuPvf/87ERERhIWFIcsy77//PjNnzlTEwF5vP9999x3//ve/yc7OrrTP0o1l6W9v3bo1kiRx/vx5rly5wpo1a5g9ezYAYWFhdO/enU6dOnHrrbfaNYlXXZk7dy7/+te/KgmTpTyA0tJq06YNarWa06dPc/jwYT744APuueceiouL0Wg0TJ8+nQkTJtC2bVul/K7Crl27uHjxIl26dKF3795MnTq1yhZacnIyvr6+HD58WNnWv39/EhISrLrFBg4cyG+//eYQ26ujuLiY1NRUnnrqKYBqW8RNqUwmk4ndu3dz8eJFhgwZgiRJ9O3bl0uXLlV6nhq7XEIUXJT8/HwSExN56623WLhwIYWFhXh7e/Pkk09yww03cPToUfr168e4cePw9/cnICCA0tJS5s+fz8yZM1m9enWDKyfLW3Tr1q25+eabCQwMrHRM+W6slJQUIiMj8fPz49ixYxw7dozOnTtz8OBBVq1axcCBA7nxxhvJzc0lLS2tQbbVF0uZHnroIf7zn//g6+trtb98eXJycti0aRMtW7YkMDCQPXv2YDKZeOSRR/Dw8FDGbTQaDREREZw7d44zZ84A5laIKzBz5kyuu+462rZtS7du3SgoKOCrr75S9lv+H25ubnh5ebFz505l35gxY9i7dy8XL15Uto0aNYrvv//ecQWogoSEBAYNGoS/v79S6ZVv8VlEvSmVSaVSsW3bNkaOHElcXBxz5sxBpVLxzTffANYvK41dLiEKLkL5StJgMPDvf/8bnU7Hgw8+SHp6OkuXLqWwsJDAwEBiYmIICgqycj3r1q0bwcHBDB48mP/85z8sXLiQXr16NegNyPL2369fP0aMGFFpf2ZmJnv37uWDDz7gp59+Yv369ZSWlhIbG8v27dvx9PSkQ4cOjBo1ir1791JaWkq7du3Iysri8uXLyLJMUlJSve2rD5YKPyoqig4dOljtMxgM7Nixg9WrV/PEE0+QkJCg7IuNjSUpKUlptQ0ZMoTDhw8rrafOnTuTm5uLTqejqKiIQ4cOOahEVWOpJDt16sTmzZuV7XfccQerV6+u1OqLiooiKiqK3bt3K9uGDRtGcnKy1Zt4bGwsFy9edOp40Lhx4wgICGDs2LGsWLECsBaF8ve4qZTJYDCQkJCglMfDw4O4uDgOHDgAmMvnqHIJUWhEZFmuts/64sWL6PV6jh8/zhtvvMH999+v9P9pNBpSU1Np2bIl3t7e+Pr6cvz4ceUtoH379phMJoqLi5XzGQwGli9fzvz58+ncuTOPP/44zz77LIMGDaqX3eWRJAm1Wk1RUZHSf3vo0CGWL1+OyWRix44d/P777zzyyCO4ubkRGRlJp06dyMrKAsyikpKSQnZ2Np07d6a4uJikpCTOnz9Penp6ne1rCOVbT1lZWezfv1/5P27cuJGtW7cyYcIESktL+e233xg9ejRgbi21b9+ejIwMwNxcT09PR6/XA+bKNyIigo8++og1a9YQHh5uN5szMjL48MMPOXHihE3HW7rzrly5gre3t1U3w9ixYzl+/DjHjx8Hrgp/SEgIvXv35tChQ8p9CwoKIjw8XBlgt5R15MiRyv/BUWWqiilTprBhwwaAKt2FHV0mqN+9AvNv/ssvv+TGG28EzK1UtVrNwIEDAesussYulxCFRsAiBJIkKQ9reno6Z86c4cCBA3zzzTfcf//9fPDBB3h7ezNz5kz8/f2VpiLATz/9xLlz59i6dSujRo1CrVZz+PBhMjMz6dGjB9u3b2fx4sUcPXoUMPfXv/LKK8ycOZPw8HDuu+8+Jk2aZNNgs9FotBICS8VpNBqVN+Mnn3ySu+++m2PHjgHw/fffs3fvXvr378+UKVPYunWrUu42bdrg5eWlBNQNGTIEo9HIpk2bUKlUzJgxg23btvHLL7/QuXPnBv2v64LJZOLEiRPKj+SZZ55h7ty5Spm2b9/Oli1b0Gq1TJgwQdkO0LZtW/z9/RXh7tOnD6mpqfzf//0fZ86cITAwkKeeeooxY8YwbNiweotCaWmp1cDpiRMneOONN/j111/Jz8+v9nvluxcs98/Hx4eWLVty5MgR5bguXbqgVqv566+/Kon/nXfeSVRUFH/88Yey7fbbb1eeYa1Wy5UrVwgPD6+TE4M9ylQeiz1jx44lJSWFoqKiamNIGqtM9ipX+ZeUyMhI2rdvD0BAQAAHDx5k+vTpDi9X0420cRFkWa7Ud69SqTAYDFy6dIlly5bRrVs31qxZw5UrV/jXv/7FHXfcQV5eHhs2bGD48OFERkYyfvx4NmzYgF6vR6vVsn//fj744ANWrFhBeno63333HWvXrsVgMODv709YWBh9+/YlIiICgOuvv75ONptMJuXto+JA3f79+zl69Cjp6el4eHgwceJEHnzwQb755hvOnTtHt27dGDJkCJ9++ilnz56lqKiI4cOHKw9kcHAwkZGRHD58WKlYz5w5w6pVq+jduzd9+/alT58+dhuQrSmw7MSJE4SGhvLnn3+SkpKCLMvk5OQwYsQIFi9ezDvvvENKSgpxcXFMnjyZzZs3891335GZmcm9996rnMfPzw+VSsXKlStp164dHh4ehISE4Ovrq4xLuLm5MWrUqHqXIyUlhePHjzNs2DDlnsiyzOjRoxk+fLjVsRUrlfL/y4sXL3LkyBFuvPFGOnfuzKZNmzh69ChdunQBzC7M586dQ5IkpYIyGo34+Phw++23s2PHDtRqNVeuXKF169YMHjxYec7VajV79uzhwQcfdHiZjh07xnXXXYenpycGgwGNRkNsbCzr1q1jypQpVoGRJpOp0crUmOWylOGjjz5i4cKFeHh4KA4NWq220csFQhTqREpKCmFhYbi5uSnbyt9gg8GALMs89thjZGVlce+99xIQEIC/vz8vvvgi8+fPV5p3/fr1Y8uWLcoDM2rUKJYtW6as5+bmkpqayp49eygsLOQf//gH+/fvZ/DgwbRq1arOlY/RaESlUikDqeWFYPPmzXz66ac8//zzhISE8PXXX/Pyyy9TVFTEhx9+SHx8PJ988gleXl5cvHgRk8nEjTfeyN///nf+/PNPbrjhBqZMmaIImtFoZOfOnWzfvp1Ro0YxaNAgtmzZYjWoa6mQ7CEM5QUhOTmZzMxMevXqxfz581mxYgXLly/ns88+48svvyQ7O5sPP/yQF198ke+//x61Wq105fXu3Zvp06fTrl07br/9doqLizl9+jTR0dEYDAYyMjIICwsjJCSEjh07Kv2/tlBRiKsiPz+frKwsq/z7/v7+3HvvvaxatYo2bdoo28v/3/Ly8jh48CAHDx7kwoULeHp60qlTJ1QqFV27diUoKIiNGzcqojB27Fh++eUXwPycrVmzhoiICEaOHMl9991HVlYWf/75JxMmTFA8xCzXM5lMSsVjWXdUmSz2l+eWW27hm2++YcqUKYodBQUFfPvtt/Uqk+W5dFa51Go12dnZlJSUMHr0aAwGAxcuXCA2Npa8vDy+//77epfLVq55UcjLy2PTpk0UFRVxzz33VPvGWVJSwk8//cS4ceOUt3OAI0eOcOTIEdLS0vDx8eHWW2/lnnvuYdGiRaSkpPDkk09iMBgoLi6mXbt2ildKr1690Gg0iki0bNmSCxcuMGPGDObOncuNN95IdnY2/v7+9O7dG41GU+WPojoqPgiWB1yn05GZmcmSJUsAeP3119Fqteh0OmJjY9m/fz9r165lypQpFBcX0759e8LCwgCzf/SlS5fIzMwkNDSUtm3bYjAY+OSTT9i5cydBQUHceuutjBw5krfeequSZ09F91V7CEJRUREbNmzAzc2NkJAQnnjiCUJDQ1m/fj3PPPMM/fv3x2g0cvToUTZs2ICfnx833HADf/vb34CrA8iXL1+mdevWtGjRgv/9738cPnyY4uJi2rZtS79+/Zg2bRqvvfZapeuXF9ua+Ouvv0hJSeGPP/5g0aJFle6PyWRi8+bNTJs2Tdmm1+tp1aoVd999N3feeSeLFi1i6NChACQlJfHLL7+QlpaGwWAgNjaWoUOH0rFjR1QqFSdPnmT//v307t2b++67j2XLljFz5kzA3BV2yy23AODr68uePXto1aqVct2goCDGjBlTZTm+/fZbhg4diiRJHDhwwKFlOnXqFEePHrVqZT7yyCP079+fK1eukJGRQadOnfDw8Kh3mZxxr8qXC+DKlStcd911eHl58eOPP7J8+XLWrl2Ll5dXg8plK9ecKBQVFZGamqqkh9i8eTMff/wx8fHxqFQqpSvC8iBY1g8cOMDWrVvx9vYmNjaWzp07YzQa+frrr5k1axYGg4H//e9/zJkzh2XLlrF27VoApZnr6+tLVFQUGRkZ5OXl4efnhyzLvPjii2i1Wnx9ffnb3/5G3759lZs+ceJEwLZ4g+qaqJZybN26VXFPfOKJJwgODubPP/8kLy+P4cOHExYWhiRJ7Nixg65du6LT6ZSH2mg0otPp6NmzJz///DP+/v7MmDGDbt26MWvWLPr168esWbMoLCzkq6++orS0lIkTJ1Z64yrv7lkbFbuEyr/FpaWlcejQIcaMGcO6det45JFHmD59Oi+99BKffPIJ8fHxJCcnk5+fz5AhQ/jss8/o0qULkZGRdO/eHTAPCF66dIkWLVqwePFiSkpKmDt3LrGxsWzcuJFXX32VuLg4kpOT+eCDD1i/fj333nsvRqPRaqzIlqjx9PR0FixYwAcffKA4C1y+fJng4GClS0ClUnHkyBHS09MJCgrCZDKh1WrZtm0bvr6+dOzYkWXLlhEeHk67du3Izs6mffv2TJ06FT8/P1JTU9m/fz/r168nLy+PsLAwBg8eDJgHZLdv3877779Py5YtOXjwIM8884xi/7vvvlvl81TVvXrooYcU+51ZJrVaTWpqKjt27CA3N5cFCxYwbtw4YmNj0Wg09SqTLMtkZGQ4tVxGo5Gnn36ao0ePcvbsWdzd3Rk1ahQFBQX4+PjUu1yipVAOy484JyeHWbNm8dFHH/Hggw+ybNky9Ho9EydOpHfv3spgTPm3WMsDYDKZSEpK4tixY4SEhBAXF4e3tzdHjhxh9erVTJkyhfz8fNq2bYuPjw8A4eHhSJKE0WhUono1Gg1fffUVWq2WW265hfbt2xMXF0eXLl0IDQ1V3hQsdu/fv5+FCxfSpUsXXnjhhUplK3+zy990vV7P77//TmlpKb/++iuPPvooMTExyttHUFAQM2fO5JZbbiEhIYHIyEjF7l69erFnzx727dvH4MGDKSkp4ejRo4SFhXH8+HHOnTvHDTfcgE6nIy0tjWnTpnHXXXcp1y4oKKBFixaKTfVNtVFeEI4fP06nTp1ITk5mxYoV9OvXj3379nH+/Hnuvfde2rRpw9KlSzlw4AAdO3ZEkiQ2bNhAYGAgAQEB/O1vf+Ovv/7iiy++YNGiRZw+fZq0tDTi4uLIz89n+PDh3HLLLRiNRg4ePMhdd91FXFwcYB6s7datm1VlVFfOnTtHz549AZgxYwZ6vZ7169dz00030bp1a+W4wYMHk5eXp5Q/JyeHX3/9lZkzZzJmzBjefPNN9uzZQ7t27ejZsyfbtm3jk08+IS0tjcDAQDp37sxdd91FVFRUJRveeust/vrrL2JiYrjtttsq7a8owrVVIufPn3d6mXQ6Hd26dePChQtWb8/1LZMkSU6/VyaTiZCQEO6//34mTpxI27Zt7VKuutAsRcHiTjhv3jyCgoL4+uuv8fb25p133sHT05PDhw8r6l5QUEBaWpriuqnX6zl37hy7d+/m6NGj/Pe//yUwMJCJEyeSlJREYGAgnTp1QpIkfv/9dzp16kR2djbDhg2z+uf379+f5cuX4+HhwYQJEwgKCkKWZUaNGsUNN9xAbGwszz//fLVlyMnJ4fz587z66qtW3VXlsVyvqKiIlJQUlixZwuzZs7l8+TJnzpzhoYceQqPR8Morr3Dvvffy3nvv8dhjj/HOO+8we/Zsxbd5//793HzzzYDZU0itVrN69WpefPFFevToQVBQEBEREdx+++1MmTJFuf7OnTsZOHAgv/zyi+LWqVar6devX4Pv33fffYe7uzsqlYqXX36Zjz/+mJUrV9KjRw+GDx+OLMu88847+Pn5ceedd/Lrr7+ybNkyZs+ezaxZs0hOTmb58uV07tyZHj168Nxzz/Hzzz/z7LPPMnr0aCIiIvDx8WHcuHGMGzdOuXZBQQGZmZm8//77+Pv7Ex0dzYABA6qslAoKCvDy8qq2y9Ei2nl5eZSWlpKZmUlgYCD79u1j27Zt+Pv7ExMTQ6dOncjLyyMlJYUBAwYo3y8sLCQhIYF58+bh5+dHQEAAfn5+yrnz8vK48cYb6dq1a6UfflXeLZbKrrxtFixlaEplateuXY3lLV8GVy+XpVtVq9Xy/vvvW+2zDD5bvldTtl570CxFoaSkBIPBQGhoKDk5OcpbMJjfhLds2cK8efN44YUX8PHx4cKFC8o/fMmSJbRt25aJEydy+vRp3nzzTWbMmEFERATR0dEcOnSIo0eP0rVrV9q0aUNwcDBHjx7l+uuvx2AwcOTIEXx9fTly5AjHjh2je/fuiqtm+T7I2tBqtSQkJChdSFWRnp7Oo48+ilar5f7776dr165kZ2ezcuVK1Go1mZmZ6PV62rdvT1FREQDTp09nzpw5LF++nLlz53Lu3DnGjx/PAw88oJz3uuuuo2fPnuTm5lq9gVlaPJYB5fnz55OYmEhJSQn9+/evNIZQHlsG7yzk5uby448/kpKSwksvvcQvv/xCcnIyx48fJzMzkzZt2hAQEMCSJUuUAb0HHniAefPm8cgjj7B161Z0Oh3vv/++0l0UFRXFQw89VOVbk8lkQpZl1Go1jz32GJmZmahUKgIDA62Otww8nz17loCAAPbt28fo0aPp2rWrVSVbscLt3Lkz27ZtY9u2bcrAYWJiotIV6ebmRlhYGFlZWVZdhcXFxXTt2pXU1FRat25NamqqMq6kVqu59dZbrf6/ULWHS1VY9jeXMlXc19TKVdF+g8GgCIGj07E3yzgFPz8/xo4dy0cffcS2bdtYu3atciNatmzJQw89xCeffML+/fsBCAwMVHzqDx48iFarxdPTk+DgYE6dOsXZs2cB8wOTn5+vpKSWJIkZM2Zw+PBhFi1axLp168jOziYmJoa7776b3bt385///IeQkJA62W8ymfDz88PLy4t//OMfJCYmApXHFkJDQ3nwwQcpLi4mPT2df//734SFhXHp0iU0Gg0Gg4EbbriB+Ph4hgwZAkDfvn2ZOXMmy5cvp7CwkNjYWO68805lMNmCl5dXlU1ywMrvOS4ujkGDBtUoCJb/VUVByMnJUfz+La52JpOJFi1asHDhQmJjY5U8L61atUKj0RAVFUXfvn3p3bs3MTExbN++HVmWCQ0N5cUXX6Rjx45cvnwZd3d3vvjiC6sfW/mKoDwqlapScFBQUFCl43/44QeOHTtG+/bt0Wq1nD9/Xrk3Fcta/m9GRgY7duzgu+++4+LFi9x0001MmjSJrl27Kg4HkiQRGBjI8ePHlcCj4uJifH19lajWXr16KW/HlrfF8hVMXboJmmOZmlO5LNdxBk2ipVAf10WDwUBQUBCPPvoo77zzDhEREfTp0wcfHx8mTpzIpk2beOaZZ/jss8+IjIxUKu6vv/6aEydOsHHjRgYPHqz0o0dFRdG1a1fCwsJYtWoVSUlJxMXF0a5dO7p06UJmZqZV/5+3t3eDygvmfstVq1YxZ84cvv32W7y9vZWBawvDhw9n48aNShBNZGQknTt35vjx47Rs2VL5X+zbt48BAwbg7u7O2LFjcXd3V77z3HPP1dtWW0lJSeF///sf4eHhPPDAA6hUKjIyMoiOjgYqN4lbtGhBcHCw4vIXFBTEtGnTePbZZxk3bhyenp5K1LdlADo4OJj33nsPtVqN0WhUzl2Ruj5LlgH41atX8+233yrbe/bsyU8//WR1zv3795Ofn8/evXvp3bs3AwYMoFevXtx1110kJyfTq1cvfHx8aNeuHYcOHeLEiRN07NgRMN/LXbt2kZSURMeOHenYsSNjxozh999/Jz8/X0ko2JCyNOcyNedyORKXbSno9Xo2b97MpEmT+PTTTyvtr80jx/Lm9+STT+Lu7s4bb7wBmAOMIiIimDNnDrm5ubz00ktcvnxZ6WLKycnh2WefJTo6mri4ONzc3Pj9999Zv349Go2GOXPmMHDgQLp166a8Cfj6+lY5IFRf1Go1Bw4cICMjg/nz5xMSEqL4xFsE4ciRI5SWluLp6UloaCiyLCs5TmbOnElQUBALFizg448/Zu3atURFRVn5VN94440EBQUp6405hWdxcTHvvvsu58+fJzMzk1mzZgFw7NgxpdWRmJjIokWLuO+++8jMzMTDw4PY2FjS0tKUvFBjxozh3XffJSEhgT///JNWrVoRExMDXP3BVReQ11CCgoJISUlRBhfB3Joq37137Ngx/vzzT8LDw3nyySdZuXKlMvlJ7969yc3NVVqnnTp1oqCggMLCQoxGI5s3b2bgwIG0bduWjRs3AuZndeTIkTz++OPceeedTJgwQZTpGi6Xo3CZlkJOTg56vZ6QkBBlwMfT05OOHTsyduxYoHpvm6qw7I+JiWHOnDncfvvt/PDDD0olNGDAAKZNm8bTTz9Nnz59lBH95ORkMjIySEpKIj8/n/vuu48DBw7Qs2dPpY+vfP9gYzFnzhzee+89oqOjGTp0KMXFxaxcuZIffviBM2fOEBISwptvvkn79u3p3LkzK1euBFAGtd98801OnTqFn5+flddEecr/Pxtz8Gr//v10796de+65BzC3xpKSkpSWzKxZs9i1axdDhw7lwQcfVLKxZmVlsWzZMi5fvszjjz9Oq1atGDhwoJIPxpF06tQJk8nEpEmTiIyMVNxdhwwZQu/evZEkiU8//ZTo6GhiY2MBuOeee3jrrbfYuHEjAwYMwNvbW0mx0aFDBwIDA/nhhx9ITU2lU6dOaLVabrvtNu666y7uuOMOQkNDAZRWrL2C/ZpzmZpzuRyF01sK2dnZLF26VPEht2SmDA4OZsiQIYwbN45Tp05Zfae0tJTTp09b5dupiUGDBnHzzTezdOlSkpOT0ev1qNVqHnnkEaKjozl79qxSKfbo0YP//Oc/+Pj40KlTJ3r16sX06dOVh8lRBAUFWaXB9ff3JzQ0lAkTJrBx40bWr19P+/bt0el0JCUlkZGRgVqtVkRPq9XSpUuXagUBHNec9fHxUYLFwJy3ZePGjUqitnnz5vHbb7/x0ksvMXToUNRqNTqdjkGDBvHVV1/x8ssvVxrfcMZMWd9++y3dunUjMTGRoqIi3n77bSZNmsSXX36ppDUp3xrr2bMnFy5cIDw8nICAALy8vNi0aRMbN27Ex8eHZ555hs6dO9O6dWtiY2OVgfj777+fzZs3c/DgQavrN8b9ao5las7lcgiyk9Hr9bIsy/LBgwfl0aNHy/3795eXLl2q7D937pz81VdfKesnTpyQlyxZIm/atEneunWrPGbMGDktLa3a85tMJlmWZTkpKUmOi4uT//3vf8uyLMslJSWyLMvyqlWr5HXr1tm9XA0hLy9PnjdvnrxixQpZlmX5559/Vuy2YDQalbKVlpY63Ma6YDQaZVk23wuLzTt37pRffPFF5RiTySQbDAblWFcnOztbzs/Pl1etWiW3bdtWlmVZfvrpp+V7773X6rhx48bJFy5ckPPy8uR77rlHnj59uvzHH3/IBoOhxvMfPXpU/v777+Xi4uJGK0NFmmOZZLn5lquxkGTZNSYnvXz5MidOnCArK4v77ruPV199lRkzZiDLMh9++CH3338/Go2G+++/n1deeUV5c3zsscdwd3dn5syZNb4VA3zyySfs3buXd999t1LzruK6MzGZTKxYsYLdu3czffp0duzYwYwZM9BoNNUmf5Pr4PLpCmRkZHDixAnFK6o6XOm+FBYW4uXlZZWqA8zpiZcuXap4tg0dOpS2bdty8OBBunXrxs0331xj0r6aylhaWmqVa8veNMcyQfMtlyNwmTEFLy8vSkpKuPXWW3n//fd5+OGHUavVPPDAA4SFhSkTSBQVFZGWlqaIwsiRI1m5cmW187SWZ9q0aUo0Z8Ub6yoVD5j79++88048PT2RJInx48crA8zVPawNiR52Bnv27CEyMrLW41zpvuTm5rJjxw66dOmiBBTu3LlTyYfl4eHBhx9+yPHjx/H19WXkyJGV7lvFNBlQcxkbu5JpjmWC5lsuR+AyouDp6YmXlxepqancddddSJLEokWLOHPmDBMnTlQmnbAkkLKkhOjevTvz5s2zykpYHWq12qnzAtcVRwxoOxrLm1ZBQQGnT5+me/fuLtUaqImwsDDef/99Nm3axE033URGRgYXLlygZ8+eeHh4AOaXG0uKDKg+MaGr0BzLBM23XI7AZbqPwOxmqdPplBv166+/8swzz6DVannmmWcYNWoUH3/8McuXL2fTpk1cuHCBhIQEunfvzvAKOcybC02lwqwrycnJXLx40SpVgKsjyzI7d+6ksLAQNzc3ZQpUy76meJ+aY5mg+ZbLEbiUKFgmnLBMSQfmymPUqFFMnjyZZ555Bk9PTzZu3Mjp06fp1q0b3t7exMXFNXo+EIGgOppjJdMcywTNt1z2xKVEwWAwcPjwYXr16gWgJII6ePAghw4dYtiwYTb1QwsEjU1zrFyaY5mg+ZarsXApUTh48CBvvvkmrVu35pFHHiE8PFzcUIFAIHAgLtPnYjQaOXToEKGhodxyyy3KxOdCEAQCgcBxuFRLQSAQCATOxWVaCgKBQCBwPkIUBAKBQKAgREEgEAgECkIUBAKBQKAgREEgEAgECkIUBAKBQKAgREEgEAgECkIUBAKBQKAgREEgEAgECkIUBAKBQKAgREEgEAgECkIUBAKBQKAgREEgEAgECv8Ps5QAEaKh6rgAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 400x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig_scale = 1\n",
    "plt.figure(figsize=(4 * fig_scale, 3 * fig_scale))\n",
    "plot_sample_df(\n",
    "    mtbsdf[~mtbsdf.show_wp_name.str.contains(\"woh\")],\n",
    "    \"bleu\",\n",
    "    \"BLEU\",\n",
    "    show_wp=filter(lambda x: \"woh\" not in x, show_wp),\n",
    "    #show_mean=True,\n",
    ")\n",
    "plt.ylim(18.2, 23)\n",
    "plt.tight_layout()\n",
    "plt.savefig(\n",
    "    \"figures/Violin_MT_BLEU.pdf\",\n",
    "    format=\"pdf\",\n",
    "    bbox_inches=\"tight\",\n",
    "    pad_inches=0.0,\n",
    ")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0b82d35d",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "4292d1429c89080e4a98e8c5b192a0c44fbb5e481943ce812f923de47b1f7601"
  },
  "kernelspec": {
   "display_name": "env: (machinelearning)",
   "language": "python",
   "name": "machinelearning"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.8"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}