experiments/sl/media/uci-experiments-create-figs+tabs.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "3d305798-f018-4ab3-88d7-40f063000b32",
   "metadata": {},
   "source": [
    "# Create tables/figures for UCI experiments in ICLR submission\n",
    "In this notebook we'll create the tables and figures for the UCI experiments in the ICRL paper. We logged runs using W&B which were run with:\n",
    "<!-- ```sh\n",
    "python train_and_inference.py -m ++wandb.use_wandb=True +experiment=australian_uci,breast_cancer_uci,digits_uci,glass_uci,ionosphere_uci,satellite_uci,vehicle_uci,waveform_uci ++random_seed=101,29,93,17,69\n",
    "```\n",
    "```sh\n",
    "python train_and_inference.py -m ++wandb.use_wandb=True +experiment=australian_uci,breast_cancer_uci,digits_uci,glass_uci,ionosphere_uci,satellite_uci,vehicle_uci,waveform_uci ++random_seed=101,29,93,17,69,68,117,36,187,329\n",
    "``` -->\n",
    "<!-- ```sh\n",
    "python train_and_inference.py -m ++wandb.use_wandb=True +experiment=australian_uci,breast_cancer_uci,digits_uci,glass_uci,ionosphere_uci,satellite_uci,vehicle_uci,waveform_uci ++random_seed=29,68,36,187,329\n",
    "``` -->\n",
    "```sh\n",
    "python train_and_inference.py -m ++wandb.use_wandb=True +experiment=australian_uci,breast_cancer_uci,digits_uci,glass_uci,ionosphere_uci,satellite_uci,vehicle_uci,waveform_uci ++random_seed=68,117,36,187,17 hydra/launcher=lumi_5hrs\n",
    "```\n",
    "This submits 5x seeds for each of the 8 UCI data sets. Each job logs test Accuracy, test NLPD, and test ECE for:\n",
    "- Baselines\n",
    "    - NN MAP\n",
    "    - Laplace BNN/GLM with kron/diag Hessian structures both with/without prior precision $\\delta$ tuning\n",
    "    - GP predictive (Immer et al 2021) for M=3200 with/without prior precision $\\delta$ tuning\n",
    "- Ablation\n",
    "    - SFR for M=1%/2%/5%/10%/15%/20%/40%/60%/80%/100% of N with/without prior precision $\\delta$ tuning\n",
    "    - GP subset for M=1%/2%/5%/10%/15%/20%/40%/60%/80%/100% of N with/without prior precision $\\delta$ tuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "425d8c3e-d7d1-49eb-9c04-5048ffd11b2c",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/m7/4s90cb416f5gppt52n88ddn80000gp/T/ipykernel_36382/3968878420.py:5: DeprecationWarning: \n",
      "Pyarrow will become a required dependency of pandas in the next major release of pandas (pandas 3.0),\n",
      "(to allow more performant data types, such as the Arrow string type, and better interoperability with other libraries)\n",
      "but was not found to be installed on your system.\n",
      "If this would cause problems for you,\n",
      "please provide us feedback at https://github.com/pandas-dev/pandas/issues/54466\n",
      "        \n",
      "  import pandas as pd\n"
     ]
    }
   ],
   "source": [
    "import copy\n",
    "import os\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import tikzplotlib\n",
    "import wandb\n",
    "from scipy.stats import ttest_ind"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2d18cf7a-39b8-4d7c-82a4-fd6b6ae35417",
   "metadata": {},
   "source": [
    "## Load data\n",
    "\n",
    "Either download the runs from W&B (and save as csv) or load data from a csv."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "4f255075-8065-4334-a6a2-7bb81e6c0a78",
   "metadata": {},
   "outputs": [],
   "source": [
    "download_runs = False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "4973a840-78eb-48e7-a5b9-5817db95cde2",
   "metadata": {},
   "outputs": [],
   "source": [
    "WANDB_ENTITY = \"aalto-ml\"\n",
    "WANDB_PROJECT = \"sfr-experiments\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "f1b4fbeb-5618-48ef-bcf6-333eefa1a83e",
   "metadata": {},
   "outputs": [],
   "source": [
    "WANDB_RUNS = [\n",
    "    # Australian\n",
    "    \"rmkfvvou\", # seed=117\n",
    "    \"escm5wb8\", # seed=36\n",
    "    \"b4olmy84\", # seed=17\n",
    "    \"vxl3juse\", # seed=187\n",
    "    \"ei3daai7\", # seed=68 \n",
    "    # Breast cancer\n",
    "    \"7uvnv5ih\", # seed=117\n",
    "    \"wvq4m8ag\", # seed=36\n",
    "    \"uqskhy0g\", # seed=17\n",
    "    \"2j2xzzk0\", # seed=187\n",
    "    \"9v8jgz4d\", # seed=68 \n",
    "    # Digits\n",
    "    \"m6xlewnb\", # seed=117\n",
    "    \"zrz4m0w7\", # seed=36\n",
    "    \"00o1czvn\", # seed=17\n",
    "    \"xjzwvtt5\", # seed=187\n",
    "    \"5wtejpuy\", # seed=68 \n",
    "    # Glass\n",
    "    \"qtdqucoh\", # seed=117\n",
    "    \"fczrmy41\", # seed=36\n",
    "    \"ezqvq88o\", # seed=17\n",
    "    \"9m9ai97h\", # seed=187\n",
    "    \"l9reator\", # seed=68 \n",
    "    # Ionosphere\n",
    "    \"yeg2yiik\", # seed=117\n",
    "    \"2lahdhi1\", # seed=36\n",
    "    \"ujy4gnir\", # seed=17\n",
    "    \"ra030cb9\", # seed=187\n",
    "    \"vs0utisj\", # seed=68\n",
    "    # Satellite\n",
    "    \"lrlll90x\", # seed=117\n",
    "    \"mdxjoasu\", # seed=36\n",
    "    \"ycvf0stp\", # seed=17\n",
    "    \"58axnjkv\", # seed=187\n",
    "    \"44qgjpzc\", # seed=68\n",
    "    # Vehicle\n",
    "    \"eoapfnjg\", # seed=117\n",
    "    \"mxud33bw\", # seed=36\n",
    "    \"ne8y4mji\", # seed=17\n",
    "    \"te2g2psp\", # seed=187\n",
    "    \"5zf4emjq\", # seed=68\n",
    "    # Waveform\n",
    "    \"a61o9vtt\", # seed=117\n",
    "    \"giw9eozu\", # seed=36\n",
    "    \"pynagve9\", # seed=17\n",
    "    \"jovevb8d\", # seed=187\n",
    "    \"0m5xdnd1\", # seed=68\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "674c733a-44ba-47c1-b3d5-5c6654b53901",
   "metadata": {},
   "outputs": [],
   "source": [
    "def load_table_as_dataframe():\n",
    "    api = wandb.Api(timeout=19)\n",
    "    dfs = []\n",
    "    for run_id in WANDB_RUNS:\n",
    "        print(\"Getting data for seed with run_id: {}\".format(run_id))\n",
    "        table_artifact = api.artifact(f\"{WANDB_ENTITY}/{WANDB_PROJECT}/run-{run_id}-Metrics:latest\")\n",
    "        table_artifact.download()\n",
    "        table = table_artifact.get(\"Metrics\")\n",
    "        df = pd.DataFrame(data=table.data, columns=table.columns)\n",
    "        dfs.append(df)\n",
    "    df_all = pd.concat(dfs)\n",
    "    return df_all\n",
    "\n",
    "def save_wandb_runs_as_csv(save_path: str = \"./csv/uci.csv\"):\n",
    "    df = load_table_as_dataframe()\n",
    "    df.to_csv(save_path)\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "9fd14abf-5354-4791-967d-4f8fe2745343",
   "metadata": {},
   "outputs": [],
   "source": [
    "save_path: str = \"./csv/uci.csv\"\n",
    "\n",
    "if download_runs:\n",
    "    df = save_wandb_runs_as_csv(save_path=save_path)\n",
    "else:\n",
    "    df = pd.read_csv(save_path)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9e746a7b-fcd6-4918-84d2-7177b3e8e4cd",
   "metadata": {},
   "source": [
    "## Helper functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "b94bcb16-2210-4ead-8d3c-2c456757817a",
   "metadata": {},
   "outputs": [],
   "source": [
    "def bold_if_significant(row):\n",
    "    if row[\"pvalue\"] < 0.05:\n",
    "        # reject the null hypothesis of identical avg NLPDs\n",
    "        mean = f\"{row['mean']:.2f}\"\n",
    "        std = f\"{(row['std']):.2f}\"\n",
    "    else:\n",
    "        mean = \"\\mathbf{\" + f\"{row['mean']:.2f}\" + \"}\"\n",
    "        std = \"\\mathbf{\" + f\"{(row['std']):.2f}\" + \"}\"\n",
    "    if \"NN MAP\" in row[\"model\"]:\n",
    "        # don't bold the NN MAP result\n",
    "        mean = f\"{row['mean']:.2f}\"\n",
    "        std = f\"{(row['std']):.2f}\"\n",
    "    return \"\\\\val{\" + mean + \"}{\" + std + \"}\"\n",
    "\n",
    "\n",
    "def add_num_data(row):\n",
    "    return NUM_DATAS[row[\"dataset\"]]\n",
    "\n",
    "\n",
    "def add_num_classes(row):\n",
    "    return NUM_CLASSES[row[\"dataset\"]]\n",
    "\n",
    "\n",
    "def add_input_dims(row):\n",
    "    return INPUT_DIMS[row[\"dataset\"]]\n",
    "\n",
    "\n",
    "def add_posthoc(row):\n",
    "    if \"BO\" in row[\"model\"]:\n",
    "        return \"Prior precision tuning\"\n",
    "    else:\n",
    "        return \"Fixed prior precision\""
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a27f97cd-d3a7-4105-b7f0-701964be6433",
   "metadata": {},
   "source": [
    "## Create Table A5 - SFR for uncertainty qunatification (UCI)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "940fd15e-59eb-49f3-b31e-fbfe0c11858f",
   "metadata": {},
   "outputs": [],
   "source": [
    "DATASETS = {\n",
    "    \"australian\": \"\\sc Australian\",\n",
    "    \"breast_cancer\": \"\\sc Breast cancer\",\n",
    "    \"ionosphere\": \"\\sc Ionosphere\",\n",
    "    \"glass\": \"\\sc Glass\",\n",
    "    \"vehicle\": \"\\sc Vehicle\",\n",
    "    \"waveform\": \"\\sc Waveform\",\n",
    "    \"digits\": \"\\sc Digits\",\n",
    "    \"satellite\": \"\\sc Satellite\",\n",
    "}\n",
    "NUM_DATAS = {\n",
    "    \"australian\": 690,\n",
    "    \"breast_cancer\": 683,\n",
    "    \"ionosphere\": 351,\n",
    "    \"glass\": 214,\n",
    "    \"vehicle\": 846,\n",
    "    \"waveform\": 1000,\n",
    "    \"digits\": 1797,\n",
    "    \"satellite\": 6435,\n",
    "}\n",
    "INPUT_DIMS = {\n",
    "    \"australian\": 14,\n",
    "    \"breast_cancer\": 10,\n",
    "    \"ionosphere\": 34,\n",
    "    \"glass\": 9,\n",
    "    \"vehicle\": 18,\n",
    "    \"waveform\": 21,\n",
    "    \"digits\": 64,\n",
    "    \"satellite\": 35,\n",
    "}\n",
    "NUM_CLASSES = {\n",
    "    \"australian\": 2,\n",
    "    \"breast_cancer\": 2,\n",
    "    \"ionosphere\": 2,\n",
    "    \"glass\": 6,\n",
    "    \"vehicle\": 4,\n",
    "    \"waveform\": 3,\n",
    "    \"digits\": 10,\n",
    "    \"satellite\": 6,\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "d5cf971c-9cc2-4756-a15e-408b22cbaab7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataset is australian\n",
      "Model is BNN full\n",
      "Model is GLM full\n",
      "Model is GP Subset (NN)\n",
      "Model is GP Subset (GP)\n",
      "Model is SFR (GP)\n",
      "Model is SFR (NN)\n",
      "Best model is SFR (GP) for australian\n",
      "Dataset is breast_cancer\n",
      "Model is BNN full\n",
      "Model is GLM full\n",
      "Model is GP Subset (NN)\n",
      "Model is GP Subset (GP)\n",
      "Model is SFR (GP)\n",
      "Model is SFR (NN)\n",
      "Best model is SFR (NN) for breast_cancer\n",
      "Dataset is ionosphere\n",
      "Model is BNN full\n",
      "Model is GLM full\n",
      "Model is GP Subset (NN)\n",
      "Model is GP Subset (GP)\n",
      "Model is SFR (GP)\n",
      "Model is SFR (NN)\n",
      "Best model is SFR (NN) for ionosphere\n",
      "Dataset is glass\n",
      "Model is BNN full\n",
      "Model is GLM full\n",
      "Model is GP Subset (NN)\n",
      "Model is GP Subset (GP)\n",
      "Model is SFR (GP)\n",
      "Model is SFR (NN)\n",
      "Best model is SFR (NN) for glass\n",
      "Dataset is vehicle\n",
      "Model is BNN full\n",
      "Model is GLM full\n",
      "Model is GP Subset (NN)\n",
      "Model is GP Subset (GP)\n",
      "Model is SFR (GP)\n",
      "Model is SFR (NN)\n",
      "Best model is SFR (NN) for vehicle\n",
      "Dataset is waveform\n",
      "Model is BNN full\n",
      "Model is GLM full\n",
      "Model is GP Subset (NN)\n",
      "Model is GP Subset (GP)\n",
      "Model is SFR (GP)\n",
      "Model is SFR (NN)\n",
      "Best model is SFR (NN) for waveform\n",
      "Dataset is digits\n",
      "Model is BNN full\n",
      "Model is GLM full\n",
      "Model is GP Subset (NN)\n",
      "Model is GP Subset (GP)\n",
      "Model is SFR (GP)\n",
      "Model is SFR (NN)\n",
      "Best model is SFR (NN) for digits\n",
      "Dataset is satellite\n",
      "Model is BNN full\n",
      "Model is GLM full\n",
      "Model is GP Subset (NN)\n",
      "Model is GP Subset (GP)\n",
      "Model is SFR (GP)\n",
      "Model is SFR (NN)\n",
      "Best model is SFR (NN) for satellite\n",
      "\\begin{tabular}{lccc|c|cccccc}\n",
      "\\toprule\n",
      " & $N$ & $D$ & $C$ & \\sc nn map & \\sc Laplace & \\sc Laplace glm & {\\sc gp} predictive & {\\sc gp} subset & \\our & \\our-{\\sc nn} \\\\\n",
      " &  &  &  &  & full & full & $M = 20\\% \\text{ of } N$ & $M = 20\\% \\text{ of } N$ & $M = 20\\% \\text{ of } N$ & $M = 20\\% \\text{ of } N$ \\\\\n",
      "\\midrule\n",
      "\\sc Australian & 690 & 14 & 2 & \\val{0.33}{0.08} & \\val{0.71}{0.02} & \\val{0.41}{0.03} & \\val{\\mathbf{0.34}}{\\mathbf{0.05}} & \\val{0.42}{0.05} & \\val{\\mathbf{0.33}}{\\mathbf{0.07}} & \\val{\\mathbf{0.33}}{\\mathbf{0.06}} \\\\\n",
      "\\sc Breast cancer & 683 & 10 & 2 & \\val{0.10}{0.04} & \\val{0.73}{0.06} & \\val{0.36}{0.05} & \\val{0.22}{0.01} & \\val{\\mathbf{0.20}}{\\mathbf{0.07}} & \\val{\\mathbf{0.16}}{\\mathbf{0.02}} & \\val{\\mathbf{0.16}}{\\mathbf{0.02}} \\\\\n",
      "\\sc Digits & 351 & 34 & 2 & \\val{0.09}{0.01} & \\val{2.34}{0.02} & \\val{3.12}{0.22} & \\val{\\mathbf{1.09}}{\\mathbf{0.05}} & \\val{\\mathbf{1.10}}{\\mathbf{0.06}} & \\val{\\mathbf{1.07}}{\\mathbf{0.05}} & \\val{\\mathbf{1.06}}{\\mathbf{0.06}} \\\\\n",
      "\\sc Glass & 214 & 9 & 6 & \\val{0.94}{0.16} & \\val{1.79}{0.03} & \\val{1.65}{0.19} & \\val{1.12}{0.09} & \\val{1.21}{0.15} & \\val{\\mathbf{0.86}}{\\mathbf{0.11}} & \\val{\\mathbf{0.97}}{\\mathbf{0.10}} \\\\\n",
      "\\sc Ionosphere & 846 & 18 & 4 & \\val{0.33}{0.06} & \\val{0.70}{0.02} & \\val{\\mathbf{0.34}}{\\mathbf{0.04}} & \\val{\\mathbf{0.34}}{\\mathbf{0.05}} & \\val{0.43}{0.04} & \\val{\\mathbf{0.32}}{\\mathbf{0.06}} & \\val{\\mathbf{0.33}}{\\mathbf{0.05}} \\\\\n",
      "\\sc Satellite & 1000 & 21 & 3 & \\val{0.27}{0.02} & \\val{1.80}{0.02} & \\val{0.80}{0.03} & \\val{\\mathbf{0.29}}{\\mathbf{0.01}} & \\val{0.34}{0.01} & \\val{\\mathbf{0.26}}{\\mathbf{0.02}} & \\val{\\mathbf{0.27}}{\\mathbf{0.02}} \\\\\n",
      "\\sc Vehicle & 1797 & 64 & 10 & \\val{0.35}{0.05} & \\val{1.39}{0.02} & \\val{1.52}{0.02} & \\val{\\mathbf{0.77}}{\\mathbf{0.03}} & \\val{\\mathbf{0.78}}{\\mathbf{0.03}} & \\val{\\mathbf{0.72}}{\\mathbf{0.04}} & \\val{\\mathbf{0.76}}{\\mathbf{0.03}} \\\\\n",
      "\\sc Waveform & 6435 & 35 & 6 & \\val{0.33}{0.01} & \\val{1.10}{0.01} & \\val{1.03}{0.03} & \\val{0.62}{0.02} & \\val{0.62}{0.05} & \\val{\\mathbf{0.52}}{\\mathbf{0.05}} & \\val{\\mathbf{0.55}}{\\mathbf{0.03}} \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>$N$</th>\n",
       "      <th>$D$</th>\n",
       "      <th>$C$</th>\n",
       "      <th>\\sc nn map</th>\n",
       "      <th>\\sc Laplace</th>\n",
       "      <th>\\sc Laplace glm</th>\n",
       "      <th>{\\sc gp} predictive</th>\n",
       "      <th>{\\sc gp} subset</th>\n",
       "      <th>\\our</th>\n",
       "      <th>\\our-{\\sc nn}</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>full</th>\n",
       "      <th>full</th>\n",
       "      <th>$M = 20\\% \\text{ of } N$</th>\n",
       "      <th>$M = 20\\% \\text{ of } N$</th>\n",
       "      <th>$M = 20\\% \\text{ of } N$</th>\n",
       "      <th>$M = 20\\% \\text{ of } N$</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>\\sc Australian</th>\n",
       "      <td>690</td>\n",
       "      <td>14</td>\n",
       "      <td>2</td>\n",
       "      <td>\\val{0.33}{0.08}</td>\n",
       "      <td>\\val{0.71}{0.02}</td>\n",
       "      <td>\\val{0.41}{0.03}</td>\n",
       "      <td>\\val{\\mathbf{0.34}}{\\mathbf{0.05}}</td>\n",
       "      <td>\\val{0.42}{0.05}</td>\n",
       "      <td>\\val{\\mathbf{0.33}}{\\mathbf{0.07}}</td>\n",
       "      <td>\\val{\\mathbf{0.33}}{\\mathbf{0.06}}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>\\sc Breast cancer</th>\n",
       "      <td>683</td>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "      <td>\\val{0.10}{0.04}</td>\n",
       "      <td>\\val{0.73}{0.06}</td>\n",
       "      <td>\\val{0.36}{0.05}</td>\n",
       "      <td>\\val{0.22}{0.01}</td>\n",
       "      <td>\\val{\\mathbf{0.20}}{\\mathbf{0.07}}</td>\n",
       "      <td>\\val{\\mathbf{0.16}}{\\mathbf{0.02}}</td>\n",
       "      <td>\\val{\\mathbf{0.16}}{\\mathbf{0.02}}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>\\sc Digits</th>\n",
       "      <td>351</td>\n",
       "      <td>34</td>\n",
       "      <td>2</td>\n",
       "      <td>\\val{0.09}{0.01}</td>\n",
       "      <td>\\val{2.34}{0.02}</td>\n",
       "      <td>\\val{3.12}{0.22}</td>\n",
       "      <td>\\val{\\mathbf{1.09}}{\\mathbf{0.05}}</td>\n",
       "      <td>\\val{\\mathbf{1.10}}{\\mathbf{0.06}}</td>\n",
       "      <td>\\val{\\mathbf{1.07}}{\\mathbf{0.05}}</td>\n",
       "      <td>\\val{\\mathbf{1.06}}{\\mathbf{0.06}}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>\\sc Glass</th>\n",
       "      <td>214</td>\n",
       "      <td>9</td>\n",
       "      <td>6</td>\n",
       "      <td>\\val{0.94}{0.16}</td>\n",
       "      <td>\\val{1.79}{0.03}</td>\n",
       "      <td>\\val{1.65}{0.19}</td>\n",
       "      <td>\\val{1.12}{0.09}</td>\n",
       "      <td>\\val{1.21}{0.15}</td>\n",
       "      <td>\\val{\\mathbf{0.86}}{\\mathbf{0.11}}</td>\n",
       "      <td>\\val{\\mathbf{0.97}}{\\mathbf{0.10}}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>\\sc Ionosphere</th>\n",
       "      <td>846</td>\n",
       "      <td>18</td>\n",
       "      <td>4</td>\n",
       "      <td>\\val{0.33}{0.06}</td>\n",
       "      <td>\\val{0.70}{0.02}</td>\n",
       "      <td>\\val{\\mathbf{0.34}}{\\mathbf{0.04}}</td>\n",
       "      <td>\\val{\\mathbf{0.34}}{\\mathbf{0.05}}</td>\n",
       "      <td>\\val{0.43}{0.04}</td>\n",
       "      <td>\\val{\\mathbf{0.32}}{\\mathbf{0.06}}</td>\n",
       "      <td>\\val{\\mathbf{0.33}}{\\mathbf{0.05}}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>\\sc Satellite</th>\n",
       "      <td>1000</td>\n",
       "      <td>21</td>\n",
       "      <td>3</td>\n",
       "      <td>\\val{0.27}{0.02}</td>\n",
       "      <td>\\val{1.80}{0.02}</td>\n",
       "      <td>\\val{0.80}{0.03}</td>\n",
       "      <td>\\val{\\mathbf{0.29}}{\\mathbf{0.01}}</td>\n",
       "      <td>\\val{0.34}{0.01}</td>\n",
       "      <td>\\val{\\mathbf{0.26}}{\\mathbf{0.02}}</td>\n",
       "      <td>\\val{\\mathbf{0.27}}{\\mathbf{0.02}}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>\\sc Vehicle</th>\n",
       "      <td>1797</td>\n",
       "      <td>64</td>\n",
       "      <td>10</td>\n",
       "      <td>\\val{0.35}{0.05}</td>\n",
       "      <td>\\val{1.39}{0.02}</td>\n",
       "      <td>\\val{1.52}{0.02}</td>\n",
       "      <td>\\val{\\mathbf{0.77}}{\\mathbf{0.03}}</td>\n",
       "      <td>\\val{\\mathbf{0.78}}{\\mathbf{0.03}}</td>\n",
       "      <td>\\val{\\mathbf{0.72}}{\\mathbf{0.04}}</td>\n",
       "      <td>\\val{\\mathbf{0.76}}{\\mathbf{0.03}}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>\\sc Waveform</th>\n",
       "      <td>6435</td>\n",
       "      <td>35</td>\n",
       "      <td>6</td>\n",
       "      <td>\\val{0.33}{0.01}</td>\n",
       "      <td>\\val{1.10}{0.01}</td>\n",
       "      <td>\\val{1.03}{0.03}</td>\n",
       "      <td>\\val{0.62}{0.02}</td>\n",
       "      <td>\\val{0.62}{0.05}</td>\n",
       "      <td>\\val{\\mathbf{0.52}}{\\mathbf{0.05}}</td>\n",
       "      <td>\\val{\\mathbf{0.55}}{\\mathbf{0.03}}</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    $N$ $D$ $C$        \\sc nn map       \\sc Laplace  \\\n",
       "                                                               full   \n",
       "\\sc Australian      690  14   2  \\val{0.33}{0.08}  \\val{0.71}{0.02}   \n",
       "\\sc Breast cancer   683  10   2  \\val{0.10}{0.04}  \\val{0.73}{0.06}   \n",
       "\\sc Digits          351  34   2  \\val{0.09}{0.01}  \\val{2.34}{0.02}   \n",
       "\\sc Glass           214   9   6  \\val{0.94}{0.16}  \\val{1.79}{0.03}   \n",
       "\\sc Ionosphere      846  18   4  \\val{0.33}{0.06}  \\val{0.70}{0.02}   \n",
       "\\sc Satellite      1000  21   3  \\val{0.27}{0.02}  \\val{1.80}{0.02}   \n",
       "\\sc Vehicle        1797  64  10  \\val{0.35}{0.05}  \\val{1.39}{0.02}   \n",
       "\\sc Waveform       6435  35   6  \\val{0.33}{0.01}  \\val{1.10}{0.01}   \n",
       "\n",
       "                                      \\sc Laplace glm  \\\n",
       "                                                 full   \n",
       "\\sc Australian                       \\val{0.41}{0.03}   \n",
       "\\sc Breast cancer                    \\val{0.36}{0.05}   \n",
       "\\sc Digits                           \\val{3.12}{0.22}   \n",
       "\\sc Glass                            \\val{1.65}{0.19}   \n",
       "\\sc Ionosphere     \\val{\\mathbf{0.34}}{\\mathbf{0.04}}   \n",
       "\\sc Satellite                        \\val{0.80}{0.03}   \n",
       "\\sc Vehicle                          \\val{1.52}{0.02}   \n",
       "\\sc Waveform                         \\val{1.03}{0.03}   \n",
       "\n",
       "                                  {\\sc gp} predictive  \\\n",
       "                             $M = 20\\% \\text{ of } N$   \n",
       "\\sc Australian     \\val{\\mathbf{0.34}}{\\mathbf{0.05}}   \n",
       "\\sc Breast cancer                    \\val{0.22}{0.01}   \n",
       "\\sc Digits         \\val{\\mathbf{1.09}}{\\mathbf{0.05}}   \n",
       "\\sc Glass                            \\val{1.12}{0.09}   \n",
       "\\sc Ionosphere     \\val{\\mathbf{0.34}}{\\mathbf{0.05}}   \n",
       "\\sc Satellite      \\val{\\mathbf{0.29}}{\\mathbf{0.01}}   \n",
       "\\sc Vehicle        \\val{\\mathbf{0.77}}{\\mathbf{0.03}}   \n",
       "\\sc Waveform                         \\val{0.62}{0.02}   \n",
       "\n",
       "                                      {\\sc gp} subset  \\\n",
       "                             $M = 20\\% \\text{ of } N$   \n",
       "\\sc Australian                       \\val{0.42}{0.05}   \n",
       "\\sc Breast cancer  \\val{\\mathbf{0.20}}{\\mathbf{0.07}}   \n",
       "\\sc Digits         \\val{\\mathbf{1.10}}{\\mathbf{0.06}}   \n",
       "\\sc Glass                            \\val{1.21}{0.15}   \n",
       "\\sc Ionosphere                       \\val{0.43}{0.04}   \n",
       "\\sc Satellite                        \\val{0.34}{0.01}   \n",
       "\\sc Vehicle        \\val{\\mathbf{0.78}}{\\mathbf{0.03}}   \n",
       "\\sc Waveform                         \\val{0.62}{0.05}   \n",
       "\n",
       "                                                 \\our  \\\n",
       "                             $M = 20\\% \\text{ of } N$   \n",
       "\\sc Australian     \\val{\\mathbf{0.33}}{\\mathbf{0.07}}   \n",
       "\\sc Breast cancer  \\val{\\mathbf{0.16}}{\\mathbf{0.02}}   \n",
       "\\sc Digits         \\val{\\mathbf{1.07}}{\\mathbf{0.05}}   \n",
       "\\sc Glass          \\val{\\mathbf{0.86}}{\\mathbf{0.11}}   \n",
       "\\sc Ionosphere     \\val{\\mathbf{0.32}}{\\mathbf{0.06}}   \n",
       "\\sc Satellite      \\val{\\mathbf{0.26}}{\\mathbf{0.02}}   \n",
       "\\sc Vehicle        \\val{\\mathbf{0.72}}{\\mathbf{0.04}}   \n",
       "\\sc Waveform       \\val{\\mathbf{0.52}}{\\mathbf{0.05}}   \n",
       "\n",
       "                                        \\our-{\\sc nn}  \n",
       "                             $M = 20\\% \\text{ of } N$  \n",
       "\\sc Australian     \\val{\\mathbf{0.33}}{\\mathbf{0.06}}  \n",
       "\\sc Breast cancer  \\val{\\mathbf{0.16}}{\\mathbf{0.02}}  \n",
       "\\sc Digits         \\val{\\mathbf{1.06}}{\\mathbf{0.06}}  \n",
       "\\sc Glass          \\val{\\mathbf{0.97}}{\\mathbf{0.10}}  \n",
       "\\sc Ionosphere     \\val{\\mathbf{0.33}}{\\mathbf{0.05}}  \n",
       "\\sc Satellite      \\val{\\mathbf{0.27}}{\\mathbf{0.02}}  \n",
       "\\sc Vehicle        \\val{\\mathbf{0.76}}{\\mathbf{0.03}}  \n",
       "\\sc Waveform       \\val{\\mathbf{0.55}}{\\mathbf{0.03}}  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def create_table_a5(df):\n",
    "    MODELS = [\n",
    "        \"NN MAP\",\n",
    "        \"BNN full\",\n",
    "        \"GLM full\",\n",
    "        \"GP Subset (NN)\", # GP predictive\n",
    "        \"GP Subset (GP)\",\n",
    "        \"SFR (GP)\",\n",
    "        \"SFR (NN)\",\n",
    "    ]\n",
    "    \n",
    "    COLUMNS_TITLES_MULTI = [\n",
    "        (\"$N$\", \"\"),\n",
    "        (\"$D$\", \"\"),\n",
    "        (\"$C$\", \"\"),\n",
    "        (\"\\sc nn map\", \"\"),\n",
    "        (\"\\sc Laplace\", \"full\"),\n",
    "        (\"\\sc Laplace glm\", \"full\"),\n",
    "        (\"{\\sc gp} predictive\", \"$M = 20\\% \\\\text{ of } N$\"),\n",
    "        (\"{\\sc gp} subset\", \"$M = 20\\% \\\\text{ of } N$\"),\n",
    "        (\"\\our\", \"$M = 20\\% \\\\text{ of } N$\"),\n",
    "        (\"\\our-{\\sc nn}\", \"$M = 20\\% \\\\text{ of } N$\"),\n",
    "    ]\n",
    "    \n",
    "    COLUMNS_TITLES_DICT = {\n",
    "        \"NN MAP\": \"\\sc nn map\",\n",
    "        \"BNN full\": \"\\sc bnn\",\n",
    "        \"GLM full\": \"\\sc glm\",\n",
    "        \"GP Subset (NN)\": \"{\\sc gp} predictive\",\n",
    "        \"GP Subset (GP)\": \"{\\sc gp} subset\",\n",
    "        \"SFR (GP)\": \"\\our\",\n",
    "        \"SFR (NN)\": \"\\our-{\\sc nn}\",\n",
    "    }\n",
    "\n",
    "    # Only keeps models we want in table\n",
    "    df = df[df[\"model\"].isin(MODELS)]\n",
    "\n",
    "    # Drop num_inducings we don't want in table\n",
    "    df = df[df[\"num_inducing_percent\"].isin([20, np.NaN])]\n",
    "\n",
    "    # Calculate mean and std (over seeds) for each combination of dataset and model\n",
    "    df_with_stats = (\n",
    "        df.groupby([\"dataset\", \"model\"])\n",
    "        .agg(\n",
    "            mean=(\"nlpd\", \"mean\"),\n",
    "            std=(\"nlpd\", \"std\"),\n",
    "            count=(\"nlpd\", \"count\"),\n",
    "        )\n",
    "        .reset_index()\n",
    "    )\n",
    "    # return df_with_stats\n",
    "\n",
    "    df_with_stats[\"N\"] = df_with_stats.apply(add_num_data, axis=1)\n",
    "    df_with_stats[\"D\"] = df_with_stats.apply(add_input_dims, axis=1)\n",
    "    df_with_stats[\"C\"] = df_with_stats.apply(add_num_classes, axis=1)\n",
    "    df_with_stats[\"prior_prec_tuning\"] = df_with_stats.apply(add_posthoc, axis=1)\n",
    "\n",
    "    # Calculate pvalues for bolding\n",
    "    df_with_stats[\"pvalue\"] = np.nan\n",
    "    df_with_stats[\"tstatistic\"] = np.nan\n",
    "    best_nlpd = np.inf\n",
    "    MODELS_WITHOUT_MAP = copy.deepcopy(MODELS)\n",
    "    MODELS_WITHOUT_MAP.remove(\"NN MAP\")\n",
    "    for dataset_name in DATASETS.keys():\n",
    "        print(f\"Dataset is {dataset_name}\")\n",
    "        df_dataset = df_with_stats[df_with_stats[\"dataset\"] == dataset_name]\n",
    "        # Find the best NLPD to use as base for paired t test\n",
    "        for model_name in MODELS_WITHOUT_MAP:\n",
    "            print(f\"Model is {model_name}\")\n",
    "            # print(df_dataset.loc[df_dataset.model == model_name, \"mean\"])\n",
    "            nlpd = df_dataset.loc[df_dataset.model == model_name, \"mean\"].values[0]\n",
    "            if nlpd < best_nlpd:\n",
    "                best_nlpd = nlpd\n",
    "                best_model_name = model_name\n",
    "        print(f\"Best model is {best_model_name} for {dataset_name}\")\n",
    "\n",
    "        group2 = df[(df[\"dataset\"] == dataset_name) & (df[\"model\"] == best_model_name)]\n",
    "        assert len(group2) == len(df.seed.unique()) # make sure there's one entry per seed\n",
    "        for model_name in MODELS:\n",
    "            group1 = df[(df[\"dataset\"] == dataset_name) & (df[\"model\"] == model_name)]\n",
    "            assert len(group1) == len(df.seed.unique()) # make sure there's one entry per seed\n",
    "            if model_name == best_model_name:\n",
    "                pvalue = np.inf\n",
    "                t_statistic = np.inf\n",
    "            else:\n",
    "                result = ttest_ind(group2[\"nlpd\"], group1[\"nlpd\"])\n",
    "                t_statistic = result.statistic\n",
    "                pvalue = result.pvalue\n",
    "            df_dataset.loc[df_dataset.model == model_name, \"tstatistic\"] = t_statistic\n",
    "            df_dataset.loc[df_dataset.model == model_name, \"pvalue\"] = pvalue\n",
    "            df_with_stats[df_with_stats[\"dataset\"] == dataset_name] = df_dataset\n",
    "\n",
    "    # Apply the function to the DataFrame to create the final LaTeX table\n",
    "    df_with_stats[\"mean_pm_std\"] = df_with_stats.apply(bold_if_significant, axis=1)\n",
    "\n",
    "    uci_table = df_with_stats.pivot(\n",
    "        index=[\"dataset\"],\n",
    "        # index=[\"dataset\", \"N\", \"D\", \"C\"],\n",
    "        columns=\"model\",\n",
    "        values=\"mean_pm_std\",\n",
    "    )\n",
    "    uci_table[\"$N$\"] = pd.Series(NUM_DATAS.values(), index=uci_table.index)\n",
    "    uci_table[\"$D$\"] = pd.Series(INPUT_DIMS.values(), index=uci_table.index)\n",
    "    uci_table[\"$C$\"] = pd.Series(NUM_CLASSES.values(), index=uci_table.index)\n",
    "\n",
    "    # uci_table.index.names = [\"Dataset\", \"N\", \"D\", \"C\"]\n",
    "    uci_table.index.names = [\"Dataset\"]\n",
    "    uci_table.index.names = [None]\n",
    "    # uci_table.index.names = None\n",
    "    uci_table.rename(index=DATASETS, inplace=True)\n",
    "    # uci_table = uci_table.reset_index().to_string(index=False)\n",
    "    uci_table.fillna(\"-\", inplace=True)\n",
    "    uci_table = uci_table.reindex(\n",
    "        columns=[\"$N$\", \"$D$\", \"$C$\"] + MODELS\n",
    "    ).rename_axis(columns=None)\n",
    "\n",
    "    uci_table.columns = pd.MultiIndex.from_tuples(COLUMNS_TITLES_MULTI)\n",
    "\n",
    "    uci_table.rename(columns=COLUMNS_TITLES_DICT, inplace=True)\n",
    "\n",
    "    # Print the LaTeX table\n",
    "    print(uci_table.to_latex(column_format=\"lccc|c|cccccc\", escape=False))\n",
    "\n",
    "    with open(\"./tabs/uci_table_no_tuning.tex\", \"w\") as file:\n",
    "        file.write(\n",
    "            uci_table.to_latex(\n",
    "                column_format=\"lccc|c|cccccc\", escape=False, multicolumn_format=\"c\"\n",
    "            )\n",
    "        )\n",
    "    return uci_table\n",
    "\n",
    "table_a5 = create_table_a5(df)\n",
    "table_a5"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "49ac930a-a46d-45b7-a819-a3e325e9d017",
   "metadata": {},
   "source": [
    "## Create Table A4 - UCI classification with prior precision tuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "ed6fe5d1-9af3-4d13-b9dd-6b4a3b57c55f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best model is SFR (GP) BO for australian\n",
      "Best model is GP Subset (NN) BO for breast_cancer\n",
      "Best model is GP Subset (NN) BO for ionosphere\n",
      "Best model is GP Subset (NN) BO for glass\n",
      "Best model is GP Subset (NN) BO for vehicle\n",
      "Best model is GP Subset (NN) BO for waveform\n",
      "Best model is GP Subset (NN) BO for digits\n",
      "Best model is GP Subset (NN) BO for satellite\n",
      "\\begin{tabular}{lccc|c|cccccc}\n",
      "\\toprule\n",
      " & $N$ & $D$ & $C$ & \\sc nn map & \\sc Laplace & \\sc Laplace glm & {\\sc gp} predictive & {\\sc gp} subset & \\our & \\our-{\\sc nn} \\\\\n",
      " &  &  &  &  & full & full & $M = 20\\% \\text{ of } N$ & $M = 20\\% \\text{ of } N$ & $M = 20\\% \\text{ of } N$ & $M = 20\\% \\text{ of } N$ \\\\\n",
      "\\midrule\n",
      "\\sc Australian & 690 & 14 & 2 & \\val{0.33}{0.08} & \\val{\\mathbf{0.33}}{\\mathbf{0.05}} & \\val{\\mathbf{0.33}}{\\mathbf{0.06}} & \\val{\\mathbf{0.42}}{\\mathbf{0.15}} & \\val{0.43}{0.04} & \\val{\\mathbf{0.33}}{\\mathbf{0.07}} & \\val{\\mathbf{0.33}}{\\mathbf{0.07}} \\\\\n",
      "\\sc Breast cancer & 683 & 10 & 2 & \\val{0.10}{0.04} & \\val{\\mathbf{0.10}}{\\mathbf{0.04}} & \\val{\\mathbf{0.10}}{\\mathbf{0.04}} & \\val{\\mathbf{0.10}}{\\mathbf{0.04}} & \\val{\\mathbf{0.12}}{\\mathbf{0.03}} & \\val{\\mathbf{0.10}}{\\mathbf{0.03}} & \\val{\\mathbf{0.10}}{\\mathbf{0.04}} \\\\\n",
      "\\sc Digits & 351 & 34 & 2 & \\val{0.09}{0.01} & \\val{0.14}{0.02} & \\val{\\mathbf{0.10}}{\\mathbf{0.01}} & \\val{\\mathbf{0.09}}{\\mathbf{0.01}} & \\val{0.22}{0.09} & \\val{0.11}{0.01} & \\val{\\mathbf{0.09}}{\\mathbf{0.01}} \\\\\n",
      "\\sc Glass & 214 & 9 & 6 & \\val{0.94}{0.16} & \\val{\\mathbf{0.92}}{\\mathbf{0.15}} & \\val{\\mathbf{0.92}}{\\mathbf{0.14}} & \\val{\\mathbf{0.92}}{\\mathbf{0.16}} & \\val{1.18}{0.16} & \\val{\\mathbf{0.92}}{\\mathbf{0.05}} & \\val{\\mathbf{0.93}}{\\mathbf{0.17}} \\\\\n",
      "\\sc Ionosphere & 846 & 18 & 4 & \\val{0.33}{0.06} & \\val{\\mathbf{0.36}}{\\mathbf{0.04}} & \\val{\\mathbf{0.32}}{\\mathbf{0.05}} & \\val{\\mathbf{0.34}}{\\mathbf{0.06}} & \\val{\\mathbf{0.41}}{\\mathbf{0.05}} & \\val{\\mathbf{0.32}}{\\mathbf{0.07}} & \\val{\\mathbf{0.33}}{\\mathbf{0.06}} \\\\\n",
      "\\sc Satellite & 1000 & 21 & 3 & \\val{0.27}{0.02} & \\val{\\mathbf{0.30}}{\\mathbf{0.01}} & \\val{\\mathbf{0.27}}{\\mathbf{0.02}} & \\val{\\mathbf{0.51}}{\\mathbf{0.57}} & \\val{\\mathbf{0.34}}{\\mathbf{0.01}} & \\val{\\mathbf{0.26}}{\\mathbf{0.02}} & \\val{\\mathbf{0.27}}{\\mathbf{0.02}} \\\\\n",
      "\\sc Vehicle & 1797 & 64 & 10 & \\val{0.35}{0.05} & \\val{\\mathbf{0.37}}{\\mathbf{0.04}} & \\val{\\mathbf{0.35}}{\\mathbf{0.05}} & \\val{\\mathbf{0.34}}{\\mathbf{0.05}} & \\val{0.46}{0.04} & \\val{\\mathbf{0.35}}{\\mathbf{0.05}} & \\val{\\mathbf{0.34}}{\\mathbf{0.05}} \\\\\n",
      "\\sc Waveform & 6435 & 35 & 6 & \\val{0.33}{0.01} & \\val{\\mathbf{0.33}}{\\mathbf{0.01}} & \\val{\\mathbf{0.33}}{\\mathbf{0.01}} & \\val{\\mathbf{0.33}}{\\mathbf{0.01}} & \\val{0.40}{0.05} & \\val{\\mathbf{0.32}}{\\mathbf{0.02}} & \\val{\\mathbf{0.33}}{\\mathbf{0.01}} \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>$N$</th>\n",
       "      <th>$D$</th>\n",
       "      <th>$C$</th>\n",
       "      <th>\\sc nn map</th>\n",
       "      <th>\\sc Laplace</th>\n",
       "      <th>\\sc Laplace glm</th>\n",
       "      <th>{\\sc gp} predictive</th>\n",
       "      <th>{\\sc gp} subset</th>\n",
       "      <th>\\our</th>\n",
       "      <th>\\our-{\\sc nn}</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>full</th>\n",
       "      <th>full</th>\n",
       "      <th>$M = 20\\% \\text{ of } N$</th>\n",
       "      <th>$M = 20\\% \\text{ of } N$</th>\n",
       "      <th>$M = 20\\% \\text{ of } N$</th>\n",
       "      <th>$M = 20\\% \\text{ of } N$</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>\\sc Australian</th>\n",
       "      <td>690</td>\n",
       "      <td>14</td>\n",
       "      <td>2</td>\n",
       "      <td>\\val{0.33}{0.08}</td>\n",
       "      <td>\\val{\\mathbf{0.33}}{\\mathbf{0.05}}</td>\n",
       "      <td>\\val{\\mathbf{0.33}}{\\mathbf{0.06}}</td>\n",
       "      <td>\\val{\\mathbf{0.42}}{\\mathbf{0.15}}</td>\n",
       "      <td>\\val{0.43}{0.04}</td>\n",
       "      <td>\\val{\\mathbf{0.33}}{\\mathbf{0.07}}</td>\n",
       "      <td>\\val{\\mathbf{0.33}}{\\mathbf{0.07}}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>\\sc Breast cancer</th>\n",
       "      <td>683</td>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "      <td>\\val{0.10}{0.04}</td>\n",
       "      <td>\\val{\\mathbf{0.10}}{\\mathbf{0.04}}</td>\n",
       "      <td>\\val{\\mathbf{0.10}}{\\mathbf{0.04}}</td>\n",
       "      <td>\\val{\\mathbf{0.10}}{\\mathbf{0.04}}</td>\n",
       "      <td>\\val{\\mathbf{0.12}}{\\mathbf{0.03}}</td>\n",
       "      <td>\\val{\\mathbf{0.10}}{\\mathbf{0.03}}</td>\n",
       "      <td>\\val{\\mathbf{0.10}}{\\mathbf{0.04}}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>\\sc Digits</th>\n",
       "      <td>351</td>\n",
       "      <td>34</td>\n",
       "      <td>2</td>\n",
       "      <td>\\val{0.09}{0.01}</td>\n",
       "      <td>\\val{0.14}{0.02}</td>\n",
       "      <td>\\val{\\mathbf{0.10}}{\\mathbf{0.01}}</td>\n",
       "      <td>\\val{\\mathbf{0.09}}{\\mathbf{0.01}}</td>\n",
       "      <td>\\val{0.22}{0.09}</td>\n",
       "      <td>\\val{0.11}{0.01}</td>\n",
       "      <td>\\val{\\mathbf{0.09}}{\\mathbf{0.01}}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>\\sc Glass</th>\n",
       "      <td>214</td>\n",
       "      <td>9</td>\n",
       "      <td>6</td>\n",
       "      <td>\\val{0.94}{0.16}</td>\n",
       "      <td>\\val{\\mathbf{0.92}}{\\mathbf{0.15}}</td>\n",
       "      <td>\\val{\\mathbf{0.92}}{\\mathbf{0.14}}</td>\n",
       "      <td>\\val{\\mathbf{0.92}}{\\mathbf{0.16}}</td>\n",
       "      <td>\\val{1.18}{0.16}</td>\n",
       "      <td>\\val{\\mathbf{0.92}}{\\mathbf{0.05}}</td>\n",
       "      <td>\\val{\\mathbf{0.93}}{\\mathbf{0.17}}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>\\sc Ionosphere</th>\n",
       "      <td>846</td>\n",
       "      <td>18</td>\n",
       "      <td>4</td>\n",
       "      <td>\\val{0.33}{0.06}</td>\n",
       "      <td>\\val{\\mathbf{0.36}}{\\mathbf{0.04}}</td>\n",
       "      <td>\\val{\\mathbf{0.32}}{\\mathbf{0.05}}</td>\n",
       "      <td>\\val{\\mathbf{0.34}}{\\mathbf{0.06}}</td>\n",
       "      <td>\\val{\\mathbf{0.41}}{\\mathbf{0.05}}</td>\n",
       "      <td>\\val{\\mathbf{0.32}}{\\mathbf{0.07}}</td>\n",
       "      <td>\\val{\\mathbf{0.33}}{\\mathbf{0.06}}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>\\sc Satellite</th>\n",
       "      <td>1000</td>\n",
       "      <td>21</td>\n",
       "      <td>3</td>\n",
       "      <td>\\val{0.27}{0.02}</td>\n",
       "      <td>\\val{\\mathbf{0.30}}{\\mathbf{0.01}}</td>\n",
       "      <td>\\val{\\mathbf{0.27}}{\\mathbf{0.02}}</td>\n",
       "      <td>\\val{\\mathbf{0.51}}{\\mathbf{0.57}}</td>\n",
       "      <td>\\val{\\mathbf{0.34}}{\\mathbf{0.01}}</td>\n",
       "      <td>\\val{\\mathbf{0.26}}{\\mathbf{0.02}}</td>\n",
       "      <td>\\val{\\mathbf{0.27}}{\\mathbf{0.02}}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>\\sc Vehicle</th>\n",
       "      <td>1797</td>\n",
       "      <td>64</td>\n",
       "      <td>10</td>\n",
       "      <td>\\val{0.35}{0.05}</td>\n",
       "      <td>\\val{\\mathbf{0.37}}{\\mathbf{0.04}}</td>\n",
       "      <td>\\val{\\mathbf{0.35}}{\\mathbf{0.05}}</td>\n",
       "      <td>\\val{\\mathbf{0.34}}{\\mathbf{0.05}}</td>\n",
       "      <td>\\val{0.46}{0.04}</td>\n",
       "      <td>\\val{\\mathbf{0.35}}{\\mathbf{0.05}}</td>\n",
       "      <td>\\val{\\mathbf{0.34}}{\\mathbf{0.05}}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>\\sc Waveform</th>\n",
       "      <td>6435</td>\n",
       "      <td>35</td>\n",
       "      <td>6</td>\n",
       "      <td>\\val{0.33}{0.01}</td>\n",
       "      <td>\\val{\\mathbf{0.33}}{\\mathbf{0.01}}</td>\n",
       "      <td>\\val{\\mathbf{0.33}}{\\mathbf{0.01}}</td>\n",
       "      <td>\\val{\\mathbf{0.33}}{\\mathbf{0.01}}</td>\n",
       "      <td>\\val{0.40}{0.05}</td>\n",
       "      <td>\\val{\\mathbf{0.32}}{\\mathbf{0.02}}</td>\n",
       "      <td>\\val{\\mathbf{0.33}}{\\mathbf{0.01}}</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    $N$ $D$ $C$        \\sc nn map  \\\n",
       "                                                    \n",
       "\\sc Australian      690  14   2  \\val{0.33}{0.08}   \n",
       "\\sc Breast cancer   683  10   2  \\val{0.10}{0.04}   \n",
       "\\sc Digits          351  34   2  \\val{0.09}{0.01}   \n",
       "\\sc Glass           214   9   6  \\val{0.94}{0.16}   \n",
       "\\sc Ionosphere      846  18   4  \\val{0.33}{0.06}   \n",
       "\\sc Satellite      1000  21   3  \\val{0.27}{0.02}   \n",
       "\\sc Vehicle        1797  64  10  \\val{0.35}{0.05}   \n",
       "\\sc Waveform       6435  35   6  \\val{0.33}{0.01}   \n",
       "\n",
       "                                          \\sc Laplace  \\\n",
       "                                                 full   \n",
       "\\sc Australian     \\val{\\mathbf{0.33}}{\\mathbf{0.05}}   \n",
       "\\sc Breast cancer  \\val{\\mathbf{0.10}}{\\mathbf{0.04}}   \n",
       "\\sc Digits                           \\val{0.14}{0.02}   \n",
       "\\sc Glass          \\val{\\mathbf{0.92}}{\\mathbf{0.15}}   \n",
       "\\sc Ionosphere     \\val{\\mathbf{0.36}}{\\mathbf{0.04}}   \n",
       "\\sc Satellite      \\val{\\mathbf{0.30}}{\\mathbf{0.01}}   \n",
       "\\sc Vehicle        \\val{\\mathbf{0.37}}{\\mathbf{0.04}}   \n",
       "\\sc Waveform       \\val{\\mathbf{0.33}}{\\mathbf{0.01}}   \n",
       "\n",
       "                                      \\sc Laplace glm  \\\n",
       "                                                 full   \n",
       "\\sc Australian     \\val{\\mathbf{0.33}}{\\mathbf{0.06}}   \n",
       "\\sc Breast cancer  \\val{\\mathbf{0.10}}{\\mathbf{0.04}}   \n",
       "\\sc Digits         \\val{\\mathbf{0.10}}{\\mathbf{0.01}}   \n",
       "\\sc Glass          \\val{\\mathbf{0.92}}{\\mathbf{0.14}}   \n",
       "\\sc Ionosphere     \\val{\\mathbf{0.32}}{\\mathbf{0.05}}   \n",
       "\\sc Satellite      \\val{\\mathbf{0.27}}{\\mathbf{0.02}}   \n",
       "\\sc Vehicle        \\val{\\mathbf{0.35}}{\\mathbf{0.05}}   \n",
       "\\sc Waveform       \\val{\\mathbf{0.33}}{\\mathbf{0.01}}   \n",
       "\n",
       "                                  {\\sc gp} predictive  \\\n",
       "                             $M = 20\\% \\text{ of } N$   \n",
       "\\sc Australian     \\val{\\mathbf{0.42}}{\\mathbf{0.15}}   \n",
       "\\sc Breast cancer  \\val{\\mathbf{0.10}}{\\mathbf{0.04}}   \n",
       "\\sc Digits         \\val{\\mathbf{0.09}}{\\mathbf{0.01}}   \n",
       "\\sc Glass          \\val{\\mathbf{0.92}}{\\mathbf{0.16}}   \n",
       "\\sc Ionosphere     \\val{\\mathbf{0.34}}{\\mathbf{0.06}}   \n",
       "\\sc Satellite      \\val{\\mathbf{0.51}}{\\mathbf{0.57}}   \n",
       "\\sc Vehicle        \\val{\\mathbf{0.34}}{\\mathbf{0.05}}   \n",
       "\\sc Waveform       \\val{\\mathbf{0.33}}{\\mathbf{0.01}}   \n",
       "\n",
       "                                      {\\sc gp} subset  \\\n",
       "                             $M = 20\\% \\text{ of } N$   \n",
       "\\sc Australian                       \\val{0.43}{0.04}   \n",
       "\\sc Breast cancer  \\val{\\mathbf{0.12}}{\\mathbf{0.03}}   \n",
       "\\sc Digits                           \\val{0.22}{0.09}   \n",
       "\\sc Glass                            \\val{1.18}{0.16}   \n",
       "\\sc Ionosphere     \\val{\\mathbf{0.41}}{\\mathbf{0.05}}   \n",
       "\\sc Satellite      \\val{\\mathbf{0.34}}{\\mathbf{0.01}}   \n",
       "\\sc Vehicle                          \\val{0.46}{0.04}   \n",
       "\\sc Waveform                         \\val{0.40}{0.05}   \n",
       "\n",
       "                                                 \\our  \\\n",
       "                             $M = 20\\% \\text{ of } N$   \n",
       "\\sc Australian     \\val{\\mathbf{0.33}}{\\mathbf{0.07}}   \n",
       "\\sc Breast cancer  \\val{\\mathbf{0.10}}{\\mathbf{0.03}}   \n",
       "\\sc Digits                           \\val{0.11}{0.01}   \n",
       "\\sc Glass          \\val{\\mathbf{0.92}}{\\mathbf{0.05}}   \n",
       "\\sc Ionosphere     \\val{\\mathbf{0.32}}{\\mathbf{0.07}}   \n",
       "\\sc Satellite      \\val{\\mathbf{0.26}}{\\mathbf{0.02}}   \n",
       "\\sc Vehicle        \\val{\\mathbf{0.35}}{\\mathbf{0.05}}   \n",
       "\\sc Waveform       \\val{\\mathbf{0.32}}{\\mathbf{0.02}}   \n",
       "\n",
       "                                        \\our-{\\sc nn}  \n",
       "                             $M = 20\\% \\text{ of } N$  \n",
       "\\sc Australian     \\val{\\mathbf{0.33}}{\\mathbf{0.07}}  \n",
       "\\sc Breast cancer  \\val{\\mathbf{0.10}}{\\mathbf{0.04}}  \n",
       "\\sc Digits         \\val{\\mathbf{0.09}}{\\mathbf{0.01}}  \n",
       "\\sc Glass          \\val{\\mathbf{0.93}}{\\mathbf{0.17}}  \n",
       "\\sc Ionosphere     \\val{\\mathbf{0.33}}{\\mathbf{0.06}}  \n",
       "\\sc Satellite      \\val{\\mathbf{0.27}}{\\mathbf{0.02}}  \n",
       "\\sc Vehicle        \\val{\\mathbf{0.34}}{\\mathbf{0.05}}  \n",
       "\\sc Waveform       \\val{\\mathbf{0.33}}{\\mathbf{0.01}}  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def create_table_a4_only_prior_prec_tuning():\n",
    "    COLUMNS_TITLES = [\n",
    "        \"NN MAP\",\n",
    "        \"BNN full GRID\",\n",
    "        \"GLM full GRID\",\n",
    "        \"GP Subset (NN) BO\",\n",
    "        \"GP Subset (GP) BO\",\n",
    "        \"SFR (GP) BO\",\n",
    "        \"SFR (NN) BO\",\n",
    "    ]\n",
    "    COLUMNS_TITLES_MULTI = [\n",
    "        (\"$N$\", \"\"),\n",
    "        (\"$D$\", \"\"),\n",
    "        (\"$C$\", \"\"),\n",
    "        (\"\\sc nn map\", \"\"),\n",
    "        (\"\\sc Laplace\", \"full\"),\n",
    "        (\"\\sc Laplace glm\", \"full\"),\n",
    "        (\"{\\sc gp} predictive\", \"$M = 20\\% \\\\text{ of } N$\"),\n",
    "        (\"{\\sc gp} subset\", \"$M = 20\\% \\\\text{ of } N$\"),\n",
    "        (\"\\our\", \"$M = 20\\% \\\\text{ of } N$\"),\n",
    "        (\"\\our-{\\sc nn}\", \"$M = 20\\% \\\\text{ of } N$\"),\n",
    "    ]\n",
    "    COLUMNS_TITLES_DICT = {\n",
    "        \"NN MAP\": \"\\sc nn map\",\n",
    "        \"BNN full GRID\": \"\\sc Laplace\",\n",
    "        \"GLM full GRID\": \"\\sc Laplace glm\",\n",
    "        \"GP Subset (NN)\": \"{\\sc gp} predictive\",\n",
    "        \"GP Subset (GP) BO\": \"{\\sc gp} subset\",\n",
    "        \"SFR (GP) BO\": \"\\our\",\n",
    "        \"SFR (NN) BO\": \"\\our-{\\sc nn}\",\n",
    "    }\n",
    "\n",
    "    \n",
    "    # Only keeps models we want in table\n",
    "    df_ = df[\n",
    "        df[\"model\"].isin(\n",
    "            [\n",
    "                \"NN MAP\",\n",
    "                \"BNN full GRID\",\n",
    "                \"GLM full GRID\",\n",
    "                \"GP Subset (NN) BO\", # GP predictive\n",
    "                \"GP Subset (GP) BO\",\n",
    "                \"SFR (GP) BO\",\n",
    "                \"SFR (NN) BO\",\n",
    "            ]\n",
    "        )\n",
    "    ]\n",
    "\n",
    "    # Drop num_inducings we don't want in table\n",
    "    num_inducings_to_drop = [1, 2, 5, 10, 15, 40, 60, 80, 100]\n",
    "    for num_inducing in num_inducings_to_drop:\n",
    "        df_ = df_[df_.num_inducing_percent != num_inducing]\n",
    "\n",
    "    # return df_\n",
    "    # Calculate mean and std (over seeds) for each combination of dataset and model\n",
    "    df_with_stats = (\n",
    "        df_.groupby([\"dataset\", \"model\"])\n",
    "        .agg(\n",
    "            mean=(\"nlpd\", \"mean\"),\n",
    "            std=(\"nlpd\", \"std\"),\n",
    "            count=(\"nlpd\", \"count\"),\n",
    "        )\n",
    "        .reset_index()\n",
    "    )\n",
    "\n",
    "    df_with_stats[\"N\"] = df_with_stats.apply(add_num_data, axis=1)\n",
    "    df_with_stats[\"D\"] = df_with_stats.apply(add_input_dims, axis=1)\n",
    "    df_with_stats[\"C\"] = df_with_stats.apply(add_num_classes, axis=1)\n",
    "\n",
    "    # Calculate pvalues for bolding\n",
    "    df_with_stats[\"pvalue\"] = np.nan\n",
    "    df_with_stats[\"tstatistic\"] = np.nan\n",
    "    best_nlpd = np.inf\n",
    "    for dataset_name in DATASETS.keys():\n",
    "        df_dataset = df_with_stats[df_with_stats[\"dataset\"] == dataset_name]\n",
    "        # Find the best NLPD to use as base for paired t test\n",
    "        for model_name in COLUMNS_TITLES:\n",
    "            nlpd = df_dataset.loc[df_dataset.model == model_name, \"mean\"].values[0]\n",
    "            if nlpd < best_nlpd and model_name != \"NN MAP\":\n",
    "                best_nlpd = nlpd\n",
    "                best_model_name = model_name\n",
    "        print(f\"Best model is {best_model_name} for {dataset_name}\")\n",
    "\n",
    "        group2 = df_[(df_[\"dataset\"] == dataset_name) & (df_[\"model\"] == best_model_name)]\n",
    "        # print(len(group2))\n",
    "        assert len(group2) == len(df.seed.unique()) # make sure there's one entry per seed\n",
    "        for model_name in COLUMNS_TITLES:\n",
    "            group1 = df_[(df_[\"dataset\"] == dataset_name) & (df_[\"model\"] == model_name)]\n",
    "            # print(len(group1))\n",
    "            assert len(group1) == len(df.seed.unique())\n",
    "            if model_name == best_model_name:\n",
    "                pvalue = np.inf\n",
    "                t_statistic = np.inf\n",
    "            else:\n",
    "                result = ttest_ind(group2[\"nlpd\"], group1[\"nlpd\"])\n",
    "                t_statistic = result.statistic\n",
    "                pvalue = result.pvalue\n",
    "            df_dataset.loc[df_dataset.model == model_name, \"tstatistic\"] = t_statistic\n",
    "            df_dataset.loc[df_dataset.model == model_name, \"pvalue\"] = pvalue\n",
    "            df_with_stats[df_with_stats[\"dataset\"] == dataset_name] = df_dataset\n",
    "\n",
    "    # Apply the function to the DataFrame to create the final LaTeX table\n",
    "    # df_with_stats[\"pvalue\"] = pvalues\n",
    "    df_with_stats[\"mean_pm_std\"] = df_with_stats.apply(bold_if_significant, axis=1)\n",
    "\n",
    "    # Reformat the table for paper\n",
    "    uci_table = df_with_stats.pivot(index=[\"dataset\"], columns=\"model\", values=\"mean_pm_std\")\n",
    "    uci_table[\"$N$\"] = pd.Series(NUM_DATAS.values(), index=uci_table.index)\n",
    "    uci_table[\"$D$\"] = pd.Series(INPUT_DIMS.values(), index=uci_table.index)\n",
    "    uci_table[\"$C$\"] = pd.Series(NUM_CLASSES.values(), index=uci_table.index)\n",
    "\n",
    "    uci_table.index.names = [None]\n",
    "    uci_table.rename(index=DATASETS, inplace=True)\n",
    "    uci_table.fillna(\"-\", inplace=True)\n",
    "    uci_table = uci_table.reindex(\n",
    "        columns=[\"$N$\", \"$D$\", \"$C$\"] + COLUMNS_TITLES\n",
    "    ).rename_axis(columns=None)\n",
    "\n",
    "\n",
    "    uci_table.columns = pd.MultiIndex.from_tuples(COLUMNS_TITLES_MULTI)\n",
    "    uci_table.rename(columns=COLUMNS_TITLES_DICT, inplace=True)\n",
    "\n",
    "    # Print the LaTeX table\n",
    "    print(\n",
    "        uci_table.to_latex(\n",
    "            column_format=\"lccc|c|cccccc\", escape=False, multicolumn_format=\"c\"\n",
    "        )\n",
    "    )\n",
    "\n",
    "    with open(\"./tabs/uci_table_with_only_prior_prec_tuning.tex\", \"w\") as file:\n",
    "        file.write(\n",
    "            uci_table.to_latex(\n",
    "                column_format=\"lccc|c|cccccc\", escape=False, multicolumn_format=\"c\"\n",
    "            )\n",
    "        )\n",
    "    return uci_table\n",
    "\n",
    "create_table_a4_only_prior_prec_tuning()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e2f91f39-a9c0-4864-8f1a-1f6e39951e01",
   "metadata": {},
   "source": [
    "## Create Figure 3 - Num inducing vs NLPD figure"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "8d987c41-031c-49cf-a6ae-ef5c17d7503a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiMAAAGzCAYAAAD9pBdvAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAA9hAAAPYQGoP6dpAABkIElEQVR4nO3deZwT9f0/8NdnJufeyw0CgiAKKhZF8UYLioq2Hj20WLEerd+v1lLb+tOvra31a7VaW2tr67faam3V1qMe9aha8aICCoi3HAKC3LD3bq6Z+fz++Mwkk2Sym+wmm93l9Xw8lmQnk2R22E1eeX8uIaWUICIiIioTrdwHQERERHs2hhEiIiIqK4YRIiIiKiuGESIiIiorhhEiIiIqK4YRIiIiKiuGESIiIiorhhEiIiIqK4YRIiIiKiuGEaI9xMaNGzF37lwMGjQImjbw/vRfeeUVCCHw6KOPdrnvBRdcgHHjxhX8HOPGjcMFF1xQ+MEB+MlPfoJAIIBJkybhN7/5Tbceg2igGnivSET9xL777otrr722157vhz/8IZ577jlcfPHFuPfee5PbTzzxRMybN6/XjgMApk6dirFjx6Kz1SiOPvpoDB8+HIZh9OKRlc5ZZ52FO+64A/X19bjiiivwySeflPuQiPoMhhGiMjn11FPx7LPP9trzrVixAocccghuueUWzJ8/P+04nn/+eViW1WvHMm/ePGzatAmvv/665+0bNmzA4sWL8dWvfhU+n6/oz3/33Xdj1apVRX/czkydOhWXXnopfvWrXwEA3nnnnV59fqK+jGGEqEzmzp2LlStXYvPmzb3yfO3t7Rg+fLjncezevRtLlizpleMAgK997WsQQuDBBx/0vP2hhx6ClLJkFRu/349gMFiSx+7KiBEjAACtra1leX6ivohhhKhMZs6cicrKSs/qyLJlyzBnzhwMGTIE4XAY48ePx4UXXpi2j2VZ+PWvf42DDjoIoVAIQ4cOxcknn4xly5Z5Pp+UEkKIrO2TJk3ChAkT8Mwzz2TdtmbNGpx99tkYMWIEQqEQRo8ejXPOOQfNzc1p+/31r3/F4YcfjoqKCtTX1+O4447DCy+8kPNnHzNmDI477jg8+uijSCQSWbc/+OCDmDBhAmbMmAEA2Lx5My688EIMHz4cwWAQBxxwAP70pz95PrZlWbjxxhsxevRohEIhzJo1C2vXrk3bx6vPSKHn09HU1IQFCxZgzJgxCAaDmDhxIn7+85/nrDQ5/XW4YDpRSvHrn0SUl2AwiFmzZuGZZ57BJZdckty+Y8cOnHTSSRg6dCiuvvpq1NXVYcOGDfjHP/6Rdv+LLroI9913H0455RRcfPHFMAwDr7/+OpYsWYLp06dnPZ9lWTk7rs6dOxfPPPMMbrzxxuS2eDyOOXPmIBaL4dvf/jZGjBiBzZs34+mnn0ZTUxNqa2sBANdffz1+8pOf4KijjsJPf/pTBAIBLF26FAsXLsRJJ52U8+efN28evvnNb+L555/Haaedltz+3nvv4f3338d1110HANi+fTuOOOIICCFw+eWXY+jQoXjuuedw0UUXoaWlBQsWLEh73JtvvhmapuH73/8+mpubccstt2DevHlYunRpzmPpzvkEgI6ODsycORObN2/Gt771LYwdOxZvvPEGrrnmGmzduhW333571n2cQNibzWJEfZ4korK56667ZFVVlYzFYsltjz/+uAQg33rrrZz3W7hwoQQgr7jiiqzbLMvK2pZIJGQoFJJf//rXPR/vX//6lwQgP/vss+S2t99+WwKQjzzySM7jWLNmjdQ0TZ555pnSNM0uj8OtoaFBBoNBee6556Ztv/rqqyUAuWrVKimllBdddJEcOXKk3LVrV9p+55xzjqytrZUdHR1SSilffvllCUBOnjw57Xz++te/lgDke++9l9w2f/58uffeeye/z/d87r333nL+/PnJ72+44QZZWVkpV69enfUz6LouN27cmPV4jY2NEoD83//931ynhmiPw2YaojI69dRT0dbWhldffTW5ra6uDgDw9NNPezZhAMBjjz0GIQR+/OMfZ93mboqJxWJYv349fvjDHyIajWL27Nmej3f88cdnNRk5lY/nn38eHR0dnvd74oknYFkWrrvuuqyqi1eTkFt9fT1OPfVUPPXUU2hvbwegmi7+9re/Yfr06Zg0aRKklHjsscdw+umnQ0qJXbt2Jb/mzJmD5uZmrFixIu1xv/GNbyAQCCS/P/bYYwEA69aty3ks+Z7PTI888giOPfZY1NfXpx3b7NmzYZomXnvttaz71NXVYerUqfjjH/+IRYsWYffu3Z2eJ6I9AcMIURmNGTMGBx10UFp/jZkzZ+Lss8/G9ddfjyFDhuCLX/wi7r33XsRiseQ+n3zyCUaNGoVBgwZ1+vgPPfQQ9tlnH/z85z/HZZddhvPPP99zv2AwiM9//vNpxzF+/HhceeWVuOeeezBkyBDMmTMHd955Z1p/kU8++QSapmHKlCnd+vnnzZuH9vZ2PPnkkwCAN954Axs2bEh2XN25cyeamprwhz/8AUOHDk37+sY3vgFANWu5jR07Nu37+vp6AEBjY2PO48j3fGZas2YN/vWvf2UdmxP6Mo/N8fe//x3xeBzHHnssDj300IKek2ggYp8RojKbO3cuHn300WT/AmfiriVLluCf//wnnn/+eVx44YW47bbbsGTJElRVVeX92HPmzMHjjz+OBx98EL/73e8wa9YsnHnmmTmP43vf+x7i8XiysnDbbbfhggsuwJNPPokXXngBV1xxBW666SYsWbIEo0eP7vHPftppp6G2thYPPvggvva1r+HBBx+Erus455xzAKT6VZx33nlpw5Hdpk6dmva9ruue+8kSdBi1LAsnnngirrrqKs/bJ02a5Ln9kksuQTwex+9+9zsceOCBRT8uon6nvK1ERPTaa69JAFn9DtweeOABCUDefffdUkopL7vsMimEkLt3787rOSKRiAwGg/K8887Luc/GjRslAPnCCy/k3Oc///mPBCCvvfZaKaWUt956qwQg33777byOw8tFF10k/X6/3LZtmxw6dKg88cQTk7cZhiGrq6uz+pV4cfqMZPZxWb9+vQQg77333uS2zD4j+Z7PzD4jU6ZMkUceeWSXx+bW0NAgAcif/OQnBd2PaCBjMw1RmR111FGor69PNpE0NjZmfYr/3Oc+BwDJppqzzz4bUkpcf/31WY+XeV8ACIVCGDZsGJqamnIeR2aTUUtLS9bspwcddBA0TUsexxlnnAFN0/DTn/40a3SI13F4mTdvHhKJBL71rW9h586daXOL6LqOs88+G4899hjef//9rPvu3Lkzr+foSqHn0/GVr3wFixcvxvPPP591W1NTk+fssS0tLQDU+SYihc00RGWm6zpOOukkPPPMM1iwYAH+/Oc/43e/+x3OPPNMTJgwAa2trbj77rtRU1ODU089FQBwwgkn4Otf/zruuOMOrFmzBieffDIsy8Lrr7+OE044AZdffnnW82ia1mVAOPXUU/HYY4/h9ttvx8KFC3H55Zfjy1/+MiZNmgTDMPCXv/wlGRAAYOLEibj22mtxww034Nhjj8VZZ52FYDCIt956C6NGjcJNN93U5c8/c+ZMjB49Gk8++STC4TDOOuustNtvvvlmvPzyy5gxYwYuueQSTJkyBQ0NDVixYgX+/e9/o6GhId9TnVN3zicA/OAHP8BTTz2F0047DRdccAEOPfRQtLe347333sOjjz6KDRs2YMiQIWn3cf4PBuL6QETdVr6iDBE57r//fhkIBGRra6tcsWKFPPfcc+XYsWNlMBiUw4YNk6eddppctmxZ2n0Mw5C33nqr3H///WUgEJBDhw6Vp5xyily+fLnnc+yzzz5y1qxZnR6H02S0atUquW7dOnnhhRfKCRMmyFAoJAcNGiRPOOEE+e9//zvrfn/605/ktGnTZDAYlPX19XLmzJnyxRdfzPvn/8EPfiAByK985Suet2/fvl1edtllcsyYMdLv98sRI0bIWbNmyT/84Q/JfXrSTCNlfuczs5lGSilbW1vlNddcIydOnCgDgYAcMmSIPOqoo+QvfvELGY/Hs36WDz/8UAKQf/nLX/I8O0QDn5CS0wASldvOnTsxYsQIPPbYYzjjjDNK8hzHHXcc3n33XTzzzDPYd999MWzYsKx9TNPEkCFD8OMf/zhrMjHqmY6ODuzevRu//e1vccstt2DhwoU44YQTyn1YRH0C64REfcDQoUNx++23FzRSplALFixALBbDMccc47lGDaCajO64446ijJShdLfccgvGjh2LW265BUcffXRy/hMiAlgZIdqDtLW14cMPP0RraytmzZpV7sPZo6xbtw6bNm3CXnvthYkTJ5b7cIj6FIYRIiIiKis20xAREVFZMYwQERFRWTGMEBERUVn1i0nPLMvCli1bUF1d3eVKoERERNQ3SCnR2tqKUaNGdTrRX78II1u2bOHUyURERP3Upk2bOp0yoF+EkerqagDqh6mpqSnz0RAREVE+WlpaMGbMmOT7eC79Iow4TTM1NTUMI0RERP1MV10sCu7A+tprr+H000/HqFGjIITAE0880en+r7zyCoQQWV/btm0r9KmJiIhoACo4jLS3t+Pggw/GnXfeWdD9Vq1aha1btya/vNbFICIioj1Pwc00p5xyCk455ZSCn2jYsGGoq6sr+H5EREQ0sPXaPCOf+9znMHLkSJx44on4z3/+0+m+sVgMLS0taV9EREQ0MJU8jIwcORJ33XUXHnvsMTz22GMYM2YMjj/+eKxYsSLnfW666SbU1tYmvzisl4iIaODq0UJ5Qgg8/vjjOOOMMwq638yZMzF27Fj85S9/8bw9FoshFoslv3eGBjU3N3M0DRERUT/R0tKC2traLt+/yzK09/DDD8eiRYty3h4MBhEMBnvxiIiIiKhcyrI2zcqVKzFy5MhyPDURERH1MQVXRtra2rB27drk9+vXr8fKlSsxaNAgjB07Ftdccw02b96M+++/HwBw++23Y/z48TjggAMQjUZxzz33YOHChXjhhReK91MQERFRv1VwGFm2bBlOOOGE5PdXXnklAGD+/Pm47777sHXrVmzcuDF5ezwex/e+9z1s3rwZFRUVmDp1Kv7973+nPQYRERHtuXrUgbW35NsBhoiIiPqOfN+/y9JnhIiIiMjBMEJERERlxTCSj3grYMbLfRREREQDEsNIVywTiDQAZqzrfYmIiKhgDCNdMSLqS1rlPhIiIqIBiWGkK4l2wIwCfX/QERERUb/EMNIZywTibSqIsDJCRERUEgwjnTEiqiqiBxhGiIiISqQsC+X1G4l2QAgAApBmuY+GiIhoQGJlJBfLUEN69SAgNIYRIiKiEmEYycWIquG8egiAUOGEiIiIio5hJBeniUYIuzLCPiNERESlwDDixd1EA0D1GbE4vJeIiKgEGEa8pDXRQFVHpATAMEJERFRsDCNe3E00gH3JuUaIiIhKgWEkU7KJJuTaqLGZhoiIqEQYRjKZMbuJJpja5jTTsDJCRERUdAwjuThNNIAaTQML7DNCRERUfAwjeXH6jDCMEBERFRvDSD6So2nYTENERFRsDCOFYGWEiIio6BhG8sYOrERERKXAMJI3u98IERERFRXDSL44tJeIiKgkGEbylezESkRERMXEMJIvIQBplPsoiIiIBhyGkbxpgMVmGiIiomJjGMmXEIA0y30UREREAw7DSL6ExjBCRERUAgwjeRNqRV8iIiIqKoaRfCVX7uWIGiIiomJiGMmbs1geO7ESEREVE8NIvoRmV0VYGSEiIiomhpG8sTJCRERUCgwj+RKaCiLsM0JERFRUDCP5EqyMEBERlQLDSL6cygj7jBARERUVw0ihWBkhIiIqKoaRgrEyQkREVEwMI4ViB1YiIqKiYhgpFJtpiIiIiophpGCsjBARERUTw0ihLK7cS0REVEwMI4UQGgA20xARERUTw0ghhAZYRrmPgoiIaEBhGCmIACxWRoiIiIqJYaQQQgCSlREiIqJiYhgpiMahvUREREXGMFIIATXpGQMJERFR0TCMFEJoUCv3cq4RIiKiYmEYKYhQVRFWRoiIiIqGYaQQQrOrIqyMEBERFQvDSEEEVDMNKyNERETFwjBSCGcGVvYZISIiKhqGkUIIYQcRVkaIiIiKhWGkO1gZISIiKhqGke5gnxEiIqKiYRjpFlZGiIiIioVhpDtYGSEiIioahpHuYJ8RIiKiomEYKZQQrIwQEREVEcNIwTTAMsp9EERERAMGw0ihhADnGSEiIioehhG3lk+Bd/4PWPN47n0EKyNERETFxDDitvNd4NXvAR/9pZOd2GeEiIiomBhG3AYfoC6b1+eufggBWCZH1BARERUJw4hb7TjAVwFYCaBlQ46dhH3JMEJERFQMDCNuQgMG7a+uN6zKsY/dTMOmGiIioqJgGMk0eLK6bFydYwcNgGQzDRERUZEwjGQaPEVdsjJCRETUKxhGMg1yKiO5wohdGWGfESIioqJgGMnkNNM0rwPMuMcOrIwQEREVE8NIpqrRgL9SDe1tXp99u+BoGiIiomJiGMkkBFA3UV1v+Dj3fqyMEBERFQXDiJe6Ceoy54gacDQNERFRkTCMeKm1KyOdhhFWRoiIiIqBYcRLfR7NNOwzQkREVBQMI16cPiMtGwAj6r0PKyNERERFwTDiJTwUCNSqwNH8SfbtQgOk2fvHRURENAAxjHgRAhi0n7ruOROrvXIvERER9RjDSC71k9SlVxgRGsMIERFRkTCM5OJURrxG1AgBgH1GiIiIioFhJJd6J4x4NdNoaoZWIiIi6jGGkVycykjLRiARSb9NQE16xonPiIiIeoxhJJfwECA0CIAEmtak3+as3MvhvURERD3GMNKZ5IiazMnPhF0VYWWEiIiopxhGOlOfoxOr0FRVhJURIiKiHmMY6YwzvDerE6uAaqZhZYSIiKinGEY6M2h/dZk514gztJeVESIioh5jGOmMUxlp2wzE21LbhcY+I0REREXCMNKZUD1QMUxdz+w3IjmahoiIqBgYRrpSO0FdtmxI3y7sfiNERETUIwwjXakYqi4juzNuYGWEiIioGBhGuhIarC6jmWFEcDQNERFRETCMdCVsh5GsyghYGSEiIioChpGuhIeoy8iu9O1CMIwQEREVAcNIVzptpjF7/XCIiIgGGoaRriQrIxlhRGgMI0REREXAMNKV8CB16VUZsRhGiIiIeophpCtOZSTeChjR1HZWRoiIiIqCYaQrgVpA86nradURuwMrh/cSERH1CMNIV4RIdWKNNKRv5/o0REREPcYwko/kXCOu4b3OdPCsjBAREfUIw0g+nH4jUfdcI5rdTMO5RoiIiHqCYSQfIY9ZWNlMQ0REVBQMI/nwmhJeaABYGSEiIuophpF8hDz6jIB9RoiIiIqBYSQfyT4jmc00FgBWRoiIiHqCYSQfuaaEh2BlhIiIqIcYRvIRyjUlvGSfESIioh5iGMlHsjKyK+MGu98IERERdRvDSD6cMGJEgERHartkZYSIiKinGEby4a8E9KC6ntWJlZURIiKinmAYyYcQuaeE58q9REREPcIwkkWoi8yKh+dcIxpgMYwQERH1BMNIJqHZs6tmhJFkJ9bMlXsZRoiIiHqCYSST0JBcBM/NaaZxL5YnNIYRIiKiHmIYyeRURjLDiNdieRBspiEiIuohhpEswjXVu4vXXCPOfhxRQ0RE1G0MI5mEBs/JzLxW7nU6u3LiMyIiom5jGMkkBKDpuUfTpM0zYjfncOIzIiKibmMY8SJ8+TXTOBUUNtMQERF1G8OIF03PPZomsjsVPlgZISIi6jGGES+aDiDHaBorDiTa1HXh9C1hZYSIiKi7GEa8CF9204u/AvBVqOvJphrByggREVEPMYx4ETlOS+aIGsHRNERERD3FMOJFeAztBVKdWN0jaiBZGSEiIuoBhhEvuSojnovlCY6mISIi6gGGES/Jic8yeE58xsoIERFRTzCMeNLsqd5zrdzrMdcIERERdQvDiBch0PnKvbvTt3OxPCIiom5jGPGSa30ar5V7hYasOUmIiIgobwWHkddeew2nn346Ro0aBSEEnnjiiS7v88orr+CQQw5BMBjExIkTcd9993XjUHuR0FKzq7rlaqZhZYSIiKjbCg4j7e3tOPjgg3HnnXfmtf/69esxd+5cnHDCCVi5ciUWLFiAiy++GM8//3zBB9trcoWRXIvlMYwQERF1m6/QO5xyyik45ZRT8t7/rrvuwvjx43HbbbcBACZPnoxFixbhV7/6FebMmVPo0/cSkaMDqxNGGlRQEXZHVzCMEBERdVfJ+4wsXrwYs2fPTts2Z84cLF68OOd9YrEYWlpa0r56lRDefUGcMGIZQKzZ3sjKCBERUU+UPIxs27YNw4cPT9s2fPhwtLS0IBKJeN7npptuQm1tbfJrzJgxpT7MbF7r0+hBIFCtrjtNNQJqP841QkRE1C19cjTNNddcg+bm5uTXpk2bev8gNJ93wMgcUSM0qInPONcIERFRdxTcZ6RQI0aMwPbt29O2bd++HTU1NQiHw573CQaDCAaDpT60zmk6PIfshocALRsyVu6V4MRnRERE3VPyysiRRx6Jl156KW3biy++iCOPPLLUT90zQveujGROfOaMumEzDRERUbcUHEba2tqwcuVKrFy5EoAaurty5Ups3LgRgGpiOf/885P7X3rppVi3bh2uuuoqfPzxx/jd736Hhx9+GN/97neL8xOUSq71aULOXCPO8F57cjQ20xAREXVLwWFk2bJlmDZtGqZNmwYAuPLKKzFt2jRcd911AICtW7cmgwkAjB8/Hs888wxefPFFHHzwwbjttttwzz339OFhvbZcK/eGB6lLp5nGGXXDyggREVG3FNxn5Pjjj4fspArgNbvq8ccfj7fffrvQpyov4VEVAVyzsO5O7cc+I0RERN3WJ0fT9AnOKJlMThiJ7krfzsoIERFRtzCM5ORMCZ+5WJ7TTNOQsT8rI0RERN3BMJKL0ADku1geJz0jIiLqLoaRXJwp4XOFkWiDaxp4j3VsiIiIKC8MI7kkh/ZmhJFgvX1FArGm1HVWRoiIiLqFYSQXkaPPiO4HgnXqunt4L8MIERFRtzCM5CI017DdDM4srO4p4S2j1w6NiIhoIGEY6Uyu9WmcWVjdU8J77UdERERdYhjpVBfr07inhGdlhIiIqFsYRjqj+fJrpmGfESIiom5jGOmM5kNezTQQ9sq9HN5LRERUKIaRzniNpgG4Pg0REVERMYx0Juf6NM6U8K4wwrlGiIiIuoVhpDPJic8yZC2Wp7GZhoiIqJsYRjojcpyeUMZoGqeZhpURIiKigjGMdCZXGHEqI7EmwEy45hlhZYSIiKhQDCOdEnbVI3N9mrpUUIk2qP0g2UxDRETUDQwjnREaVH+QjJCh6akF86K7XKNp2ExDRERUKIaRziQXy/OahTVjeC/AyggREVE3MIx0xgkjXhWPrCnhwQ6sRERE3cAw0hkhkJxdNZMzoibqCiPswEpERFQwhpHOCM3VHyRDsplmV2obKyNEREQFYxjpiuaD9yysXs00rIwQEREVimGkK5ovRwfWjGYaIQBp9N5xERERDRAMI10Reo4+I5nNNBpgsZmGiIioUAwjXckVRjKbaYQALFZGiIiICsUw0hVN996e1UyTYwgwERERdYphpCtdrU8TbwXMGABWRoiIiLqDYaRL9rozmQK19kgbqKaa5Mq9HFFDRERUCIaRDFJKJExXc4vQvOcyEyI18VlkN1KL5bGphoiIqBAMIxlihoWWjnhqgzPxmZdkJ9ZddmiR4CysREREhWEYySQlpP0FoPPF8tKmhGdlhIiIqDsYRjxY0l3fEAC6Wrl3VyqwsM8IERFRQRhGPMlUpkiuT+NVGRmkLp0OrKyMEBERFYxhxIMaFJPRTOO5Po1dGYnudjXlsDJCRERUCIYRD5aUqUghNAA5KiNcuZeIiKjHGEZySFVGhGukTAavlXtZGSEiIioIw4iHrNyRa+XeUMaU8J53JiIios4wjHiQUCNqknwhQHpM9c5mGiIioh5jGPGSOa27Hui8mcaIAIkO584lPzwiIqKBhGHEg5QZkcJZgyaTvwrQg+p61B7ea5mlPjwiIqIBhWHEg4SE5a6ECJ/6ylyVVwjXXCO7oCZIYzMNERFRIRhGcnGXRnS/qo5khhHA1W/EnmvEax8iIiLKiWHEg5XZTCM0FUi6CiMQgMXKCBERUSEYRjxI93Twjlwjapxmmugue9p4VkaIiIgKwTDiRQJWZoVDyzWixl0ZybGgHhEREeXEMJJDVqTQ/d47pq3cC3soDgMJERFRvhhGcpBWRhVE+AChZ/cbSc7C2pBaUI+zsBIREeWNYSQHmTl5meYDNI9OrMn1aeyhvdJiZYSIiKgADCM5mJl5QtNVU43MmNTMHUaSC+qxMkJERJSvHFOL7pk+2daCRxZ/gqBPw7nHTszewRcCEm3p25w+I9EGO4OwzwgREVEhWBlxae6I4+X3t+DdTxu8u31ogeyg4fQZMWOA0QHAYp8RIiKiAjCMuFSGVKEokjAhJdKnhAfsNWpE+jZ/BeALq+vR3XYQYWWEiIgoXwwjLpVBO4zEDUjpMfGZ5rdH1GT2G3HPNQJWRoiIiArAMOJSYYeRWNy0qyJeI2p8gJVI357sN+KEEVZGiIiI8sUw4lIZVBObSQCRmOlRGckxoiZt5V7nEYiIiCgfDCMuAZ8Gn6b6hHTEE9l9RgDVPyRrrpHMZhpWRoiIiPLFMOIihEg21UTipvdOWsCjMuKe+AzsM0JERFQAhpEMThjpiBk5hvd6jKgJu6eEF6yMEBERFYBhJIPTb6Qjbnj3/PAaUeNeLA9adjMOERER5cQwkiEc1AGoZhrpVRrxGlHjbqYRApxnhIiIKH8MIxmcyoiaa8RjB01XYUR6VEaclXtZGSEiIsobw0gGp89ING5mr9zryBxRk+wzYs/Ayj4jREREeWMYyVDhmoU153QhejA9cDhhxDKARKvqT8IRNURERHlhGMlQGUgN7bVy5QnNlx429CAQqFbXI432RoYRIiKifDCMZAi716fJFSiSnVhd/UacTqyx3apqwqYaIiKivDCMZEhbLC9XaUTz29URj34jkSYAks00REREeWIYyeDuwGrmChTOiJq0TqyuxfJYGSEiIsobw0iGZGUkkWOeEUfmiJqQaxZWeKz4S0RERJ4YRjJUuOYZsTorbugB7xE1rIwQEREVhGEkQ4UzmiaWYzp4h+ZP7xeSuXIvKyNERER5YRjJ4J5nxDBzrNwL2CNqXGvUJJtpnJV7WRkhIiLKB8NIBieMWBKIJazc/UY0n10dsfuNJEfTNKhLjqYhIiLKC8NIhnBAV2vdoZOVe4HsuUbSVu4FKyNERER5YhjJoGkawn5Xv5HOChx6KLV6b3LSs0Y7oLAyQkRElA+GkQxCAOGgDgDoiBldDO91rVETGqQupQXEm1kZISIiyhPDSAYhRLIy0hFPdD2ixqH7gWCduh5tBGQnnV+JiIgoiWEkg+aqjETiXUx8pvkAobn6jTgjaprS160hIiKinBhGMgghELbnGunoqs+I5gOEa42akN2JNdbAMEJERJQnhpEMmhDJ4b2djqYBVBjR/R6VkQYA7DNCRESUD4YRD5X2lPDRmNn1fCF6KLVGTTKMNKavW0NEREQ5MYx4qHD6jCTMrgfo+oKp4OE000QbVIjhxGdERERdYhjxUBlSlZGOmAGrq0Ch+ZCcJc29WB4kh/cSERHlgWHEQ6XdgTWaMLueu0zzqxE10krvMyIlOPEZERFR1xhGPFSGAgDUYnldxglnRI2VSM3CGmlQ4YSVESIioi4xjHioCqVW7s2rmcYZUeOsT5NspmFlhIiIqCsMIx6cob2RuAlp5REonBE1ThiJNQFWjJURIiKiPDCMeEirjORzB2dETbAOgN2ZNdII9hkhIiLqGsOIh+Q8I/lWRjSfyiCanlowL7qblREiIqI8MIx4qLKH9iZMC3Ejj2ndhQ8Quj2ixuk3wsoIERFRPhhGPFSEUqvxtkXzmElV99sjagwgzMoIERFRIRhGPPh1gaBPnZr2WLzrOwhdNdFYhmsW1kaOpiEiIsoDw4gHIQTCzmJ5sXyaaQTgC9uVEffEZ6yMEBERdYVhxIMmBCrsWVjb85kSHgB8IUC65hqJNTKMEBER5YFhJIfkXCMxI7/WFs0HQKYqI5EGFU6IiIioUwwjOaQmPjOQ16gY4VNr1ASdDqwMI0RERPlgGMnBaaZRK/fmcQfdDwg/EKpT30cb1RTxRERE1CmGkRycykhH3IDMp53GGVETrFPfR3ezMkJERJQHhpEcKu25RiLxPAOFEKoTa6BGfZ9oA+LtQCJSoiMkIiIaGBhGcqh09RnJe7oQPQT4K+3OrAAi24GO7YCZKM1BEhERDQAMIzk4lZG8m2kAu9+IllqfxoyrCknHTg7zJSIiyoFhJAenMhKNm/mvMCN8qrkm5Azv3Q0EaoFYgz3UlzOyEhERZWIYyaHCXrm3I995RgB79V5/+sq9mg74q1RzTbylNAdLRETUjzGM5FAVsisjCRMy39qI5lPhwwkjkV3qUg8AehBo3w4kOkpwtERERP0Xw0gOVcnRNAVURpwRNcnhvQ2p2/wVqt9Ix3bVl4SIiIgAMIzk5FRGInGzsK4eeggI1avrTmXEEahWw307dnBCNCIiIhvDSA7OaJpYwoRhFTASRvdnN9M4hFAztEab1G3s0EpERMQwkoszmgYAOqIFzBMifK7F8nZ73K6pCklkFxBr6tlBEhERDQAMIzkE/T74dQEAaCskjGg+oGacur7zHWDds9n76H7Vt6R9OxBv6/nBEhER9WMMI50IuxbLy5vmAwZPASafB0ACL38b2PF29n6+sKqStG8HjGhxDpiIiKgfYhjphLNYXlu0gDDijKg5/Gpg7OdV0PjXBUDrpux9A1WAGVMjbKwCnoOIiGgAYRjpRKU98Vl7vMC1ZfSQupz1e1UliewCnjsfiDVn7xusBeKtnDKeiIj2WAwjnQjblZH2QiojQGqhvEAVcPKfgYoRQONq4MVvZi+aJ4Ra6Te6W00ZT0REtIdhGOlE+sq9BQzD1fwqZEgJVI0CTvkz4KsANi8CFl2TPaRX86nVfiM7gBinjCcioj0Lw0gnnD4jHTEj/8XyAHuNGh9g2VWQIQcCs3+vOqx+/BCw8s7s++hBFWLatwOJSI+PnYiIqL9gGOlEsjISK7Qy4lNf7llW954NHHWDuv7mTcDaJ7Pv568EpGFPGV9gPxUiIqJ+imGkE84srB2FTgkvhKp0yIy+JgdeABx0sbr+yneBbW9l3zdQAyTa1JTx7NBKRER7AIaRTqTWp0kU1kwDqHlEvIbrHnEdsPdJakjv8xcCzRvSbxcCCNQCsUY1gyunjCfqW6RUq2/HWjgkn6hIGEY6URF0Vu41C2umAewRNR730XRg1p3AkKlqVd9/nQ9EG7P38Vep4b5ew4GJqPeZCfX32LoJaPkUaN0INH+q/n4ZSoh6hGGkE6nKiFF4gUIPAFrAuzOqvwI4+T410qbpE+CFSwAznn1/PWh3aG3v1vETUQ85VZD27XYA2QQYEdW/K1gPQAJtm1Uo6djNvl5E3cQw0olKd2Wk0Dv7QkDlcMCKeU/3XjkcOPl+VQHZuhh47QfZTTL+MNSL3TbAiHXnRyCi7jDjanXtlk+Blg2qyVRoKoAEqu0Rc0J9sAjWq/t0bFX7d+zK/nBBRJ1iGOlEld2BNRIzutd3I1gLVI5UYcTrxWnwZODE/wOEDqx+FFhxu8dj1ABmVHVodY/OIaLikpaqQrZvV5WOts/U362/GgjVqQ8YQmTfTwj1wSE0SF3v2GZXSnbxQwRRnhhGOuEeTWNa3exIGqwDKoap1Xm9SrhjjgeO+Zm6vuwXwOrHPB6jVrVVR3YVt0OrlKqt24iqF+FYs/o0yPZv2pOYcdXvo2Wj6lAe2Q3oPhUuAlWqD1e+fHYo0XQVSlo+VR8kuBgmUad85T6AvsyZZySaMNAajSMc9EHz+mTUGSGA8GBAmkBkJyBqs1/cppynSsHv/B549ftA1V7AqCNcj6GpCknHTjWZWsXg/J9fWipcWKYaamyZajI2IwZYcXu7/aXuoNrDw0NUE1KhPy9RfyAt1Rck3qq+rLjqoxWsVX9vPeULqS8jqv5uo03qsYO1ajsRpWEY6YQTRqQEmjriqK0IIhzoxikTAqgYql4Ao7tVtSTzBW/G/6hPUeufBV64CDjjKaBuQup2zafapyM7AV9AtVs7ByfNjLBhqE97VgwwDTtsuJqahFChRuiA7gdEKBWQpFRVkpZNQKhefcrzBQv/mYn6IiNmVwGbVFAQUNWMQFVpns8JJWZMVTajTUCoVg3f94dL85xE/RDDSCeCfh2aELCkRDRmoj1qdC+MACp8VAxTwSDWpDq9uasOQgM+fwfwz63AjrfVKr9n/BMID0rt4wupoNG+PdUPxYy5qhsG1KsrVNDQdPsyCGiV+VU5hFAvzJahhh7H24CKIWoytkLK1UR9hWUChj0vSLwNkAm1snawpjhVkHzoQfVlxlUzUFqlJMwKJO3x2GekE0KI5Po0lpSIxBOIGz3oRKrpQMVw9cYea8ru/+ELA3PuBarHqGabFy7MbmsOVKWafBJt6jF0e6G90CC7mlGvXmj9lSrA6P7CX+w0n3ocTVNDF9s+4xBj6l+MqBpu27xB9QeJt6gqX2iQqjL2VhBx0wPq78oXtvupfAq0bVF/W5zgkPZgDCNdcMJILGHBsCQ6Yj3s3Kn7gcoRKih4TWhWMRQ45X4VWLa9Bbz6vexp4QPVqqknUK1KvXqwdFULX1g9VyKiXtDbuW4O9WGWqSogrZtVCOnYCkCq3+FgrQoDfYHuVyN0/JVAvDkVSuJtDCW0R2IY6YITRiJxAyG/jvZoAobZwzVj9IAKJHpQvXBmqp8EnPgHVZ1Y+4QaZVNOTgdaX1hVZFo2qiDFtXOoL5BSheWOXa4qSGtqZIu/ou82g2g+FZT8Vapy0/Ip0PqZHUr490V7DoaRLtRXqk9Sn+5sRcCnI25YiMSLMN+HLwRUjVQvRvG27NtHHwsce4u6vuLXwKq/9/w5e0oPqBd3WKqDa9sW7xlmiXqDZdhVkM/sIbTb1XanmVL3l/f4CqH5VOUmUGN3IN9oh5JWhhLaIzCMdOHo/UcCAF7+YAssKRHwaWiLxrs/74ibv0JVSJwppzPt/1Vg2hXq+mtXAZsX9fw5i8Ffac990qJeNDt2cW4S6h3O30rHTlUFad2oOqf6K1QI8Ze4M6gZB9Y/B7y8AFj2y+x1pXpK01WQClSraedbNqop6GMtDCU0oDGMdOG4KSMQ8uvY3hTBR581IujXEUtYiCaK9OYbqAKqRthzf3hUGQ77ATDhi+rN/oVvAo1rivO8PaXpqs1b9wPt29QLZryV7d1UGpmL1HXsUKHDPT17qUgJ7HwP+M91wF8PAV64GFj9CLD8NuChI1Uo8Wpu7QlNt/uG1aqOuC0bU82jnImZBqBuhZE777wT48aNQygUwowZM/Dmm2/m3Pe+++6DECLtKxTqP5P+hAM+zNh3GADgpfc2QwgBXRNojxqFr+SbS7BWVUiMmBqq6yY04PhfAsOnq45uz52v5ivIl7RUM1D7dqB5HbDrfWDLEmDjS8AnTwEf/w1474/AijuApTepT32F/Fy+kPpEasZV000719GhInHmvPFapM4ZkVLKKkjHTuDd/wMePRH4x8nA+39UlZCK4cCBFwGDJqsAvvw24MEjVHOqV5NrTwhNhZJQnfoba92kqkHRJoYSGlAK/jjx97//HVdeeSXuuusuzJgxA7fffjvmzJmDVatWYdiwYZ73qampwapVq5Lfi77amcyDJgRmThmJVz/cimWf7ERLRxxVIT864gZihoWQv0ijWIJ16sWlfRsQ0NLbu30hYM6fgCe+oIb8PnseMPo49UKd/OoAjHb1Ymh0pLZ7VVu6MvbzwLE3q5lg8yGEesG0DDWHgjM3SbFms6Q9ixlXv8+xJvW7LGGH3kFd3bMIzx0DPv03sOphYNPLqZmJ9SAwbg4w6SuqP5fmU0F//bPAstuAxtXAW7cA7/4B+Nx/AwdcoEJTsQjNHtYv1Tlp+wzQw2oeolJXhoh6gZAFfryfMWMGDjvsMPz2t78FAFiWhTFjxuDb3/42rr766qz977vvPixYsABNTU3dPsiWlhbU1taiubkZNTU13X6c7jAtiW2N7bjh0RVYv6MV846diLmH7o22aAJVIT8GVxexyiOlGq3SscOeZCzjBaZpLfDEF9WLdKGEpnrs+yvs+Ucq1aX7S1rAmn+oqbH9VWpW2ClfLzxQJCKAGVE/Q3hwcV+UaWCSlgrO8VYg1qpmD9YCqvpR6sn2pAR2vacCyNrH0/++hh0C7PcVYMLp6gODF8sE1j2tKiRNn6htocEqlEyZX5qZVqVU58uMpkKJv6p/ddqlPUK+798Fxel4PI7ly5fjmmuuSW7TNA2zZ8/G4sWLc96vra0Ne++9NyzLwiGHHIKf/exnOOCAA3LuH4vFEIulSv0tLUVujy2AJlQlZ+aUkVi/oxUL39+CUw8Zi5BfRyRmIB42EfAV6cVSCLUmjGWqCkMwYx2buonA6Y8CHz9khwsnSNgBw19lB42KjKBRoWaczKcidfClam6T7cuBRf+jmnJm3grU7pP/z+EPq8ml4q3qE64zrTxfKCmTEbNnR21OTarnrwACvRBg27er8LHqYaAxVblF5Qhg3y8B+31Z/c11RdOBiV8E9jlNDcVf/ktVwVxyA/DOXcC0y4HJ81SwKhYh1HnyhVWfktbNqnoUrAeC1X1nPhWiPBVUGdmyZQv22msvvPHGGzjyyCOT26+66iq8+uqrWLp0adZ9Fi9ejDVr1mDq1Klobm7GL37xC7z22mv44IMPMHr0aM/n+clPfoLrr78+a3s5KiMAsK2pHW1RA1fetxixhIkffekQTB5dj9ZIHPWVQdRWFnntFqe5JtbovY5NqVkm8OGfVR8So0MFmenfA6Z+s/BysBlXM8XqYRVKdD+g+dXjsLS8Z8q1SJ0vXPrfdSMKfPoisPphYNOrrmaYEDD+ZNUMs9cxPavGWIZafXvFr1QfDwCoGAFM+zYw+Vz1sxablPYSERF7qvs6e3gzQwmVV76VkZKHkUyJRAKTJ0/GueeeixtuuMFzH6/KyJgxY8oWRna3RtEeTeDBRWvx8vtbcNR+w3H5KQcibpiwJDCiLgxdK/KLqGUAbVvVJ8ZQfXkmbWrdpIYUf/aa+n7IVOD4XwCDc1e1PCVLyjEA0l4vx6+Cic+ZQdZnh5RuTF1P/UPaInUR9f/sqyj9G6aUwM6VwKpHgE+eTG+GGT5dNcPsc7p68y4mM6FCz4pfqyUVAKByJHDId4D9vlq6n9uIqC8tqDq+Bmq42CWVTUmaaYYMGQJd17F9+/a07du3b8eIESPyegy/349p06Zh7dq1OfcJBoMIBvvOH0844ENrJIHPHzgKL7+/BW+u3YHWSAJVIR/aoglE4iaqQkUOI5oPqBxuL6zXrF5Uelv1GODUB9UL6hvXA7veBf5xKvC5y9QLar6f8JySsr9CfS8tNZTZjNtrclgARKpi4gvZfQX8rpDCRfr6pZyL1PVC5+b2baoP1KqHgSbXkPjKkcCkLwOTvpS+Mnax6X7VPDPpS2rU2tt3AO1bgdevBt7+LXDoAtUcVOzmS1841XzTsV2NAEqGkv4zkpH2LAW9GgQCARx66KF46aWXktssy8JLL72UVinpjGmaeO+99zBy5MjCjrSMgn4Nfl1g9OBKjBtaDcOUeP2jrRBCwKerSdCsUsyvoQfUC6ce9F7HpjcIoT7FfeVlYNwpqmKz4tfAo3OAbcu6+Zia+pn8laqcHBqkLvWg+hTrXlvE+XImV4u1qBK/meCcJn2V02RQjkXqjCiw9kk14uyBw4ClN6og4gsBE88C5j4EfG0pcPj/K20QcdODwAHzgXP+Axx9g1q9u+0z4NXvAw/PVGGpFJMGOiOQdD/QvgNo/lT1k+GsydQHFTya5u9//zvmz5+P//u//8Phhx+O22+/HQ8//DA+/vhjDB8+HOeffz722msv3HTTTQCAn/70pzjiiCMwceJENDU14dZbb8UTTzyB5cuXY8qUKXk9ZzlH0zga22No6Yhjyert+NPCVdhrUAVu+foRkAA6ogaG1YURDpSoD0Qiosq80lTD+Mpp3TPAomvVqB8I4MALgcOvTlU9is0yVRXFMtQlpHojE361WrEeVi+6TgVF93M4cblYht0XxKmCGKn/n1I2vUkJ7FihJiJb+5Saj8cx4nC7Gea08v/tOIwI8OFfgJV3puYMqh0PHHqlmuCwVFVApxIpdFWZCtaVZqQPkUtJmmkA4Ktf/Sp27tyJ6667Dtu2bcPnPvc5/Otf/8Lw4cMBABs3boTm6j/R2NiISy65BNu2bUN9fT0OPfRQvPHGG3kHkb4i7NfRDIkjJg3HA6+tweaGDqza0oz996qDEEB7NIGQXy/NHCr+sOrh37ZFvdiX6o0/H/vMBUYdBSy+Xr34v/9H4NMXgONuUXOfFJumZ784S5kKKEYjEDUBCEDYnWKdzpBOZ1ndz86ypeJUQRLtqnpnRtWbnXP+S6l9q+oouvrh1JBaQM2PM+lLqimmdnxpj6E7fGHVGXzyecCH96tQ0rweWPhtVXU89Eo1lLjYoVoPqC8zrppuYs2q6SZUV/oJ5Ii6UHBlpBz6QmXEkhLbmzpgSeD+V1bhlQ+24pjJI/Dfcw6AaVmIJUwMq61AsFiToHmJtahA4nT8LLdNr6gOrk7nvP3OAY68Tn3qKgenemIZqbK30FRI0f32EGd2li0KpwriDMm1DLuvQomrIEYE2PC8atr47DWoGdGgnnf8XFUFGXVU/6qOJdqB9+8F3vl9qnPtoP1VKBl/Sul+FjOhJkqEcIWSPrzCMfVLJRlNUy59IYwAQEtHDA1tMWxviuC6vy+DX9dw5yXHoCrkR2skgZoKPwZVlbiDWLQJaN8CVQnQ7EuhLiHURdY2r8siibcBb94MfHAfAKmmyj7mZ2qYZF/gdJZ1gorz66751Sd4f1h1qEwOOWZn2ZycUVHuKohmB+NSrw2zfbmqgHzylBoO7Bh5hKqA7DO37zTDdFe8VVUa3/lDqqlp8BRg+veBvU8qXUiwDDX8HlChJFin+nMxlFARMIyUQCxhYkdzBwI+DT98aBk27mrD+TMn4eRpY5AwLJiWhWF1FfDrJf5UFmu2O6FZgGWpS+l8Of+d0r7uuoT7e6T2c1/V7JlaC31z2fom8Nr3U+XyfU4Djv5foGJot37EkpIyVT2xEqm5JoRuN/ME7CGn7oDi23NfnM1E+sRk0lKViHwn0uuuts2pZpjm9antVaNTo2Fqx5Xu+bsirdJULWLNwLt3A+/dnQoJQw5SoWTsrBKHknb19+Gsh+Ov7F9VJupzGEZKQEqJnS1RxA0Tr3+0Ffe9vBqjB1fi5+fNgBACrZEEBlUFUFNRxmHJmQEk12XmNmmp60YUiDbYo10K7JtiRIHlv1LlZmmqT1hHXQ/se3b/eCPP6iwLddxOXxRfhRoRovkHfmdZZw2UeJtqHuyt6dkTEWDDc6oZZvMipJphKlT1Y7+vqGpIuc67mVAVISuBZHXSX1WacxJtVAv1vfdH9X8BAMOmqVAyemYJQ4lphxJTrYcTrFeXA/V3nUqKYaRE2qIJ7GqJQBMCl92zCHHDwk++cigmjapDNGFCABheVwFd6wdvvl6kVKMhOnaqScoCNYW/0O56H3jlSmD3B+r7MfbCe9V5LrzXl7g7y1pGRhXFb3eWDalqykDoLJtrkbpSzk8hJbDtLdUh+pOnUtUAABh1pJoVdZ+55VvjyDJU0LbiahSXv0JNkCZ0FRjirepvpFRVhEiDCvgf3Jta+HL4dBVK9jqmdKFEWqn+QP5Ke1h2JZsxqSAMIyWSMC1sb+yA36fhnpc+xmsfbsVxU0bi0pOmQEqJtmgCQ2vDqAz283VYjJgadhhtVJ+GCx0CaCaAd+9SlRIzpj7ZjjgcGHKAmsF18AFqpEN/fWFL6yxrQg05FvaQY9fMsu6mnr5aHSrXInWtm4E1j6iZUVs2pLZXj001w9SMLd3zd0Za9vTqMbv6UWFPGlaRPpuplCo8RRvUudMDpetv0bETeOd3wAf3q+oMoKpE07+vQlupONP3Wwl1HkKDSlcNogGHYaSEdrVE0BEzsLmhAz95eBkCPg13XnwMKkN+dMQMBPwahtWESzPMtzdJS5XoO3aqF6JgTeGf/BrXqr4k297Kvs0XAgZNVsHECSmDJpd36HJPpHWWNeymL7iaeULpQ47L3Vk2uUhdk2oeEUiFqFJJdADrn1XNMFveQFozzITTVRVk5OHlaRKQUr3JG1EVJvQQEKi1F6Tron+MtFSYizaoaoIeKt1w2fbtajjwR3+1l1gAMOpo4LAfACMOK/7zOZzQ6ny4CA1SzTf9uRJIJccwUkIdMQM7WyKoCOi4+oE38dnudlxwwiScdPAYWJZEezyBEbUVCJVqErTeZkRVlSTWlJpquhDSAna+q5pvdn9gf32YKjmnEUDdPsCgKelVlIphfbey0Bl3Z1mZMeRY89udZcOpZp5Sd5YtxyJ1UgJbl6pmmHX/TK3OC6g30f2+DIw/tTzNMFKqN1czqq7rIRW6nRVxCz0nlqnOa2S3+v32d+PvJV9tW9S08h8/mOrjNHqmWtRy+KGleU4g1Z/IjKlJ7cKDVIdXhhLywDBSQqZlYVtjBzRNYOF7W3D/q6sxZkgVbp53OIQQaIsmUBn0Y0jNAFoHQlpqWHFkl3pTDXSjSuJmmao0v/sDYJcdTnZ/oNbS8BIekgomTkip3af/loqdzrLSucxcn8eZWbZInWWNmOqMGm/uvUXqWjcBqx9VIaTl09T2mr1TzTDVY0r3/J0x4ypkS8OugFSppodiNU1ZBhBtVitvm1G7eadErwetm4G3fw2s+nsq7I79vGq+GXpwaZ4TcC2AGVWhJLlScD9voqaiYhgpMWd6eAC47O5FSJgWfvrV6Zg4shaGaSFumBheV4GAr5++WeaSiKhAEm8uzQtsx85UMHGCSvMnqSYPt2Qzz5RUUBk8uXwdHXsq2VnWFVSkdE11n1FF6aqzbK5F6nyh0lVBEu1qyYDVDwNbFqe2+6vsZpgvq75D5ahyuUfCaHbfjkB1aWeLNeNqqG60UV0PVJUuALZsVDO4rn4k1dF675NUpWTIgaV5TiA1C68ZsStL9UCwuvSrMVO/wDBSYtG4gW3NHagM+nHXCx9i0UfbcPwBI/HNE9U0963ROOoqgqitCPT/viOZLFO9uEZ2A7DsKkkJf8ZEBGhc5aqiOM08HR47C9Ux1l1B6c/NPIBrThSv9Xk8OstKC4i3q2Y1M6r2dUJMKUhLNcOsehhY97Tr/0Wo0R6TvqxmEi1HXyDLVOfAjKm+O/5K9em9lOfDixFT/x/RJhUU/JWlC0DN64EVt6sVi50QP/5UNaPr4MmleU6HEVFfWhAI1dthr++swE69j2GkxNzTw2/Y0YqfPrIcQZ+GOy85FhVBHxKGhahhoCroR1U4gFApp4kvl0SHWjAv1lraT3xepKWaeZLh5ANg14dAxzbv/cND0isoQ5xmnn7azi0tj4nb7DlRIHtnkbqWT1PNMK2bUttrxqn5QPb9UnmGczsjYYyoanLxhdUSBU5oK2coTURUKIk1AZCAv7p0TY1Na4HltwNrn0Cyo/A+pwPTrwTqJ5XmOR1GVIVSLWCHkhqGkj0Uw0gvaOmIoaE9hqqgH1f9ZQk2N3Tgws/vh9lTRwNQfUsicRNCAJVBP6pC/tKuXVMOlqlGEER2pSaAKueLfWSXqpq4Q0rTWu9mHj2k1gAZfAAwxAkqU/p3M49Tni9VyIq3pZphti5JbQ9UA/t8QXVGHT69938HkiNh7NElelAFkGRH1D5UFXM6gEabVCgRemknFWtcDSz7peo8DAAQwMQzgEMXAHUTS/OcDiNqV0r89krBtaWds4b6HIaRXuBMDx/063hh5Wf4y2trsPfQKvzsa4enNc0YpgolugZUhlQoGXB9SeJtqkqSaFef9vpSJzYjAjQU0swzLruzbMXwvvWG1pukpYbhrn5EBZHkKCgBjD5WDccdd3J5lqM3Y+p4LMuugFSrMFnK0UHFIqX6e4k22BOn2ROqleq4d3+o5v1Z/6z6XmjAvmcBhywo/erGZkxVUoUPCNXaQ6b7wGKfVHIMI73APT28YUpcds/rMEyJ/z33MOwzPPs4DdNCJGHAr2moCvlRGfLDV+p1bHqTZah+JNHdqfb5vvoGXmgzT2hweh+U/t7Mk4/mDSqArH4ktTIzoH7u/b6i3siqytAM4x4JowVVVSHgjITph/8f0lJhPtqgJlDTg6VdPXfX+8Cy24BPX1DfC1316znkO6WfZM6MqwAmdFelpI9VrqioGEZ6iTM9fHU4gN8+9z7eWLUdJxw4CpfMzt1RLG6YiCZMBHQN1eEAKkM+6NoACSXOjJTtO1Tv+kBN/3qDiOxOVU52uZt5zOx99RAwaL/0gDJosnpj7K/iraoT6qqHgW1vprYHaoAJX1AhZNghvf/mkTkle6Cq9CNheptlqr+djl32yJRuzHxciJ3vAMt+AWxcqL7XfMB+XwWmfaf0fX3MhB1KhKqShGpLG8CobBhGeol7evg1W5txw6MrEPTr+N0lxyDcxaRnsYSJuGEi4NNRUxFAOODrv2vaZDITqkISaVBvFv21HwagmgEaV6dXUXZ/mD55l1vNuOwqSsWIvvtCKy21KN3qR1QJ37CnGhcasNdxKoCMO6l0k3flkjUSJqzeuHzhgd0Z0jLsidMaVCjpzkSDhdi+XFVKPntVfa/5gf2/BhzybaByZOmeF7BXhLb/jgJ2paQvV1SpYAwjvciZHr4i6MP371+CrY0duGjW/ph1UNefLqSUiBkWEoaJUEBHdViFEm0g/DFKqV5UO3aqN/RAVd9eo6UQ0lKjSdL6oXwAtOdq5hmU3Q+lbkJ5q0bN69S6MGseVbN5OuomppphSv1mlElaqhOqaU/J7gsD/hogUFn+kTC9zUyk5iixYvZw4BKGsK1vqkrJlv+o7/UgMPk84HOXAZXDS/e8gApgiXYkRxiF6hlKBgiGkV7kTA9fGfTh2RUb8cDrazF+WDVu/NrheT+GlBLRhAnTlAgHfagO+xHy6wNjjhIzrkrPibbUtNXCNduo1sPZRfuSgpp5gqnRPM5InsGTVfNDqcRa1KiK1Y+krxcUqAUmflH1HRg2rXffBJwp2Z2OsXrINRKmhBO09RdGTE0yGGlU/WT8VaVtmtryBvDWL4BtS9X3egg44HwVSsJDSve8gN1U1a5CaaDKFUr28N+BfoxhpBe5p4ePJkxcfs8iGKbEjV87DOOHFXa8lpSIxk1YUqIi6EO1PRy434eS5BotCfVlxu31LRIZ06HrqYDSn/qadCatmcc1u2xnzTyDM9bmqRzZ/YBgmerT7qqHgfXPpVZ8FRow+ng1HHfvk3p/yKUZszuiWiqYBVwjYfrrNP+lZERdE6dZpV2kTkrVdLfsVtWMA6j/lwO+ARz8X2o9mlJKhhLTHmFkLyqp6QA0+1LYk//Zl8i4TNvez18/+zGGkV7W0BZFaySBqpAfv3n2fSxevR2HTRyKS0+a0mXfES+WJdERN3ptjhIpJaR9mb69k/vk8ZiZhBAQcF4bBGAZEFYCQhoQZiK1Kqiz+i2QHk5KuYhcb5KWmr7bPe397g+A9q3e+4fqXRUUZzTPhM4/ITetTTXDuJuP6iep4bj7nglUjijuz9WVrCnZq9Rw3P46EqYcEh2q6SbeDECoc1iq8Cal6kuy7BfAjrfVNn8lcOCFwNRvqd/LUnImsJOW/YFFuuYMklBLTTuX8A4lznZNV6N4hO667hFg4BFwhOuxqCAMI73MPT38qs1NuOHRFQCAqpAPJ08bizkHj0ZlqPDSqmFaiCZMaCL/OUqklLCkcynVXFhp29SlYVmwLAlTSliW2u6VPjoPJIX/+giI9DyRFlAAAQuaNCGsBDSZgDCj6gsmNMuEgAR0HUL4IXQVUIT9QiGSL0r2y5NIbkk9vn0lLRTZ14Xr9rKINLg6ydqVlMY1uZt56vdLr6BUjwY2vqSqIDtWpPYN1gETvqj6ggw9uHcDnWXac4HYI2H8FeWZkn0gcSZOizQA8RYV5ErZnCGl+r1a9gtg13tqW6AaOPBiYOolqlmt3KSECisyI7hI122Wfd2+lNL+W3AFGyeEZIUSJ5jo6cFGaAVUaLTe/dvrAxhGepl7eviQX8eS1dvxyOJ12NqoJtaqCPgwZ9ponDJtLKq6GUo64mqOkuqwH36fngwWlmXBtKT6khYsy650SBUWnPChCEhIaMkAIKAJu2Ih0K2Os13dxYkDTnBxDsU5Pof6PvVNWjaSEtKM2006CQgzDhgRCCsOYRnqjhogodsLyfkA6JBCJp8/8zg7C0XO5uR50lz30QQ02KFFpB5HZDy2KxflDEXp24QrkLlCkRFVzTyZk7Yl2nKccedBdWDM8SqA7H1iSTs/Or9fqf8/CzIRVR0voakmoGAtpG6vFCwEZMZ9sx8T6Y/p2ig99lPbs+8ku9jPeX73HL3Scj+PfbvlcTxIhdrk/yEAYY+Kc6KB833m76JwP0iu213h2X2b80MJQ02cJuJtgOaHCDgdP9NDtrrewzdCKdX8JG/dCjR8pLYFaoGDvwkceFFp+zv1ls6CS1rAcQcdIK1CA6SqLBCQdpiR0AABSOFLhRhNV69bQrP309TvmCvASGlfOsFGIuNvyDmC9N//zm5PVsNdP3p1Cea+YhgpA2d6+OqQ+rRnWRJL1mzH40s3YHOD6h8Q8us48eDROPWQsaitKPxTYdwwEUuoT8lC1QiSIUIIV7iAyNo+4EipOvTZTT2wAwqsmB1QzOQLgBR2M4/wpYccwP6D9AhFrje+tL8SmV4PkhKAkBAy9f8hpYQQQn3wAnKGoqxtGVUbJ5Q4by0qFAkIaUHv+Ay+hg+hN30EveEjaI0fQWvfAqtuEhITvwRj/BmQFcPsn8H187neaKXHtuS/rndn98ttVvBwzoGUkGYMwoyqZ9TCkIFqSF+F6gRpn4/UY3pU4dyP57GfkCK5n3PO3C9hzu+5c94Bde5T24Rr36yn7/JNOzNEeB2v+5i8QpOETH4Yd/8NO/dJPrezv338OYMMAEgLWqINItYI3eiA9AUh9TCgaWn7Zt5Xc/2+AemhW33v+t1zpRohJfSNzyO48lfQmlarQwjWIXHAN2FOvsCu0mQcZyehSuS8LS15qX+z3mjTT35WiM0Kys5Fxv2y7p87sEpLev6tOI9h2dUZaQcYaang4nwvLNN5YHu75bzouH5uCSt5IpxQAzuMCDvEaIDuA6BBOs1Omg4phV0tFpDJFw4NEpodjFMVG+cZLEiMqKss+jpqDCNl4J4e3j2JmSUllq3diceXrsenu9Sn2YBPw+ypozH30LGorxzAcyaUg2UmO8oKK5EeUJx+KEKHTPZB6aV+KO4QI7NDTkGhKOMF1tkurAikCALOfDUyvzffXD+/5xufa18hJWDFoVmqXV/qIUh/tZrAypea2tzzuT1CWdoHy7R9B2CYLlB29cm5kKnfC8uESLSqSonRActXkdYxubM36qx3Ao9A5X4MCPWGW7HpedR89Bv4W9cBAMxAPVr3uxjtE+bB0sP222oqHeaqVHZ1W/bP4P3W5RUmvEKfeh7XBwbXMTp/N9nV1OQd07cXELzU9vTbcz6OlMnzLJxqjPohkKzYwIKQQLJSk3nAzg+eDCeufjRO5UXoaI9LDBs+BqFQcTuyM4yUgZQSO1oiSBiWZ6dVKSVWrNuFx99cj3XbWwEAfl3DCQeOwunT98bgai4gVTLSyggoUbsjpV1Vsdt5kgFlIA03LgUzDmHGAGlA6kHAVwkZqFazhnIkTPlZBkSsGYg3QpgxuzpVwg89lonAp08j/N4d0Fs3qE2hIYgccCliE89NBaKcgSrtZmTcDCCPN27XN12F7D1Ol/1oLERiUQzeawpCFcVtamMYKZO2aAI7myOo6aQJRkqJdz9twONL12P11mYAgE8XmDllFL4wfW8MreUCUr1CylRAkYZr0TU7sNjDjWUynPj27DdaywDMCIRlQGoBwFcJBKogfWF1fqjvMeMQ8WYg1gRhxSF9laXtNGwZCKx/AuH3fwu9bZPaFB5mh5JzShuIqPssC5H23QwjXelPYcQ9PXxXHYGklPjws0Y8vnQ9PvysCQCgawLH7D8CXzx8HEbUVfTCEVMWez6UrH4o0oQzqkW1z7qGGw9U9pTswooDmk91QE32A+FImH7DjEHEmlQwkab6/ytlgLQSCK77B0Lv/RZ6h5rd1wqPQOTA/0ZswpcYSvoahpH89KcwAqSmhy9kKO/Hmxvx+NINeG9jAwBVXTx6PxVK9hrUj9d1GSiy+qFEASvq6ociU/1QhFNJ6aclYmmpNy8zqn4mXxgI1KgOkXvalOwDjRFxhRIJWco5SgDAjCP4yaMIfXAndHtFbLNiFKIHXYbYPmezotZXMIzkp7+FkfZYArtaoqgM+grueLdmazOeeHM93l6/G4Bq+5yx7zCcMWM8xg7px6vBDkRp/VDsVWXNSP/shyKlCiBWTDUj+4KAvwbSX6H6gTCADBz2HCUqlLSoDoylnnLdjCG49u8If/B7aJEdalPVGEQOvBzx8WcM7Opif8Awkp/+Fkbc08N3NUFZLuu2t+CJNzdg2Sc7k9sOmzAUZ84Yj3HDBsBY/oEqOdzYaeZx90MxkmMA+0w/FDNmd0Q1VUdUf7X6tOwL993gRMUhJWC0Q0QbIRKt6nfSNQKqJIwogmsfUqEkqj5wmdV7I3LQFYjvffqe3SernBhG8tPfwgiQPj18T2zc2Yon3tyApWt2JHuWHzR2ECaMqMHQmhCG1oQxtDaMwVXBok9WQ0WU2Q/FjKbPhyIlIHy9M9zYaySME0D4CXXPIy2IRBsQa4RItNmdkytKWw0zOhBa/QBCH/4BWkw1TZs1+6hQMvZUhpLexjCSn/4YRqJxA7taojAsiaBf63aFxLG5oR1PvLkBb6za5jk9uxDA4KqQCii1YQyrCWFITRjDalVgqa8Kdmt2VSohy0yvotjNPCXph2IZdkfUBKD5VQdGpyMq2+0JUHOUGG1AdDeEEYHUQyqgllKiHaHVf1WhJN4EADBq90Vk6neQGDOH1bnewjCSn/4YRgA1CVpH3EBHNIGEKRHwaQj4tB5N4LS1sQPL1+3EjuYIdrZEsdO+TJhWp/fTNYEhNSEMqwm7Kir297Vh1IT9nFiqL3APN3aaeeyF5dRwY4nUcOMu+qFYpqq+mDF7JIzdEdUX5mgGys0y7InTdvfOHCUAkGhFaNX9CH10D7R4CwDAqNtfhZLRJ7LPUqkxjOSnv4YRR8K0EIkl0BY1EDdM+HQNQb9etEqFJSWaO+LY2RzFzpYIdrZEsCN5PYrdrVGYVuf/zUGfhiE1YdRWqIX4gn7Xl09zXXe2a67r2bf59Z6FLnLJ6ocSTy48lzbtvTPcGFJ1RIVQAcRfbXdEDfFFnfJnJeyJ05p6LZSIeAtCH9+L4Md/gmavvWQFatVz+8KQehjSFwL0EKQvnKzeSD2Uft0XUr/79qW6Tzi13dnHF7b/LvbwCgzDSH76exhxmJaFaNxEazSOWMKCABD06yXv62FaFhraYq6wEsWOlkiyqtLYFssxsXL3CYH0sGLPu+LTNfh1kbzu0zX4NAF/8nv39czbMu6rieR8Lj5N3abrKgj5dQ2681iaetwBF46c5hzLXjjQVMONAWEHkMo9/oXWkhKmqVaoNi0Jw7RgmBKmZcGwv+9yuyVhmtnbIQR8moCmCeiagE/TktfTv7Q8tqV/r9mP595Wtt9fMwYRb4aINal+Rr7KkjftiVgTQh/9EaFVf1YLAZaY1AJ2MHECizvYOKEnlB6GnOvJMJQZfpz7htXoNL0P98liGMnPQAkjDktKxBIm2qMJROIGTEstoOf3ledNI2FY2NWqgkp71EDMXowvljDt61bq++Q2e3vGvobZd3+d9LxDj7DDjGaHGWEHIXVbKhjZ32uaCkWax20Zj5G6zfU8xXyjcVbsKkEAsSz1pm44b+5m6nvTtT3zezPtTdx13WvfLh6j8+3OfVPPZfX9l7e8CYG0wONzhZbsEOQVgHIHI03zDlbubbo04bMi8MkoNEjo/hB0XbfDE+Bz9hMibZsmBHwaMsKW+zYBXcux/lCiDXr7FtXfyYhCmBHAvhRmVM2b4mzP2icKYbj2c10Ke8Xv3mYJHyw9DEsLwtJCMPUQTC2UvG5p6ntTC8LUwjC1IAw9BEuEYGhBGPa+hut7QwvBFEF7WwgJEYIFHyx7XRpL2uvuJK/bxVbXyujSshCPd2DukQdhr2FDivoz5/v+3Udj2sCmCYFwwIeQX0fcsNAeS6AjaiCSMBD06T3uV1Iov0/DyPoKjKzv+YyvpmWlwktaULHsNw0LCfvNI3ndkjAM53r6baYp7f3VbQlTJh/HyLotdV+vNyLTkjAtE7BXPe5LUsHEK/QIV/Wn89t0XcCypPebfc4Q4Hzyd6oAroqA641/oLyvOwHSp6s3Yud7PRkY7e3298ntmqq8OW+6Ekg718719C+r020q4Kltqeu5F4BLmBbQ9359i0KFLbXquDsEaULAkmFIGYJEnf1GmlogUCbfYCWcU2dJ6dru/XwaTPiRQAgxBEUMQdhfznUR99imLkOIIWDf7r4eFPb3rushEUs9pzSgGa0AWkt6Lk2pIfkTyaA6Ihl0bQtk3B6EJoNo3ueKooeRfDGMlJEQItmMUR1W/Upaowbaoomi9yvpLbqmoSKooSJY/l8ty3KFFfvNNeEVZLq6zfne8r7NKyglsm6zYBgyeT3zBVLta6I/vdM4n9Sz3tQ93+yd686buusNvsv7egSEHNvd36f2cQKFemPrD811Vp5hJp/g44Qc1WSlfteT25z72c1ZyUv3Y6XdZsGSqupkmgYsMwHTMGBK9QZoSmQfQ9o2wJTSMyCosKVuKH7Dscc5ho4YdMQQSl+Rz3VdwF7wFur1OnUd0Ozl+DSRWoVX89pfSgSEgSCiKqCIGAKIJ0NQAFEEkUAQUTgRISDiCMiouo4o/FBhJyBj8COKANTtfmebVPtpUJVRXVioQAQViGSvMNiJLfJrPTijPVP+dwwCoFbv9VcEURnyJ/uVdEQNCAGEAjp0bc9t9+8uTRMIairs9TWZQSlhppoYMgNP8jYjMzRlVIrsJgtNiGT1xCm5u9/cc7/5p6oszpt3emXAtd0u41NpaHZlwI8+/ncvJZBog4g5E6f51QKKXQQ+S7rCiRN4LAnLgisspYKL86umCftNHvabvuu62kfY+zj7uwKDHRqESA8Mmmtfd9joD6EVABL2V4c9Ek+YUbvJSjVhwUw1TSGt6cpu7rL3M6NNGDR8XNl+DoaRPkbXNFSGNISDvmS/ko64Acsyy9qvhIqrLwclorwJoear8VcCiVYg2ggRb1QdPjtZRkATApouwF//IhIC0AOQegBATWG1JacDa/WYEh1c1xhG+ih3v5JqV7+SaMJAoAz9SoiIchIaZKAW8FWpUBLbDRFrtIfkhsp9dNQPMIz0ce5+JVUhE9G4YfcrUU04AR/n9CCiPkLTIYN1gL8KiDUD8cZUKOFEe9QJhpF+RFVEdFSG/IglLETiBqJxA20JBhMi6kM0H2R4MBCoBuLNQKwJwuhQc5TogXIfHfVBDCP9kHvEijOUNpowEIkZaE8YAIMJEfUFegAyPBQI1KhAEm8CzAjXRKIsDCP9XFowqbAQN1TFhMGEiPoMPQhZMRwyUAMRbYRItACyQ60WzRV6CQwjA4quaQgHNIQDGcEkbqI9ymBCRGXmC0NWhiCNWohYE0S8RXV+9Vfu0csWEMPIgJUWTCyJuGEymBBR+QkB+CshfRWQwVpVKYm3qJWofRUMJXsohpE9gK6pYcLuYBJNmOiIGWnBxKdr/W7GVyLqp4QA/FWq/0iiDYg1QsSbIbWAHUr4WrQnYRjZw7iDSU04kBZMOmJGcrZDzV68TROumQmFYFghouISGmSgRs3cmmgFYg0Q8abkyrm0Z2AY2YNlBpOEYSJuqmnH1RTNFkwLsCy1HoUlATgrQMI9FbOautpZo0HTGFyo/3EWMJf2P87vubOOinRt64yzTonXr797m7Mf0rY5++2BfzvuOUoSrUDUnjhND6umm+Q5ybjM3L4nnrsBgGGEANjLewd8yJwrUa1+qYKIZbmuS5laX8JeDM6yAEuq9VKkfV9I9aIrkXqN8HrBzbzN2SCQ/dri9WKfun/6AzlhifquQkOAsxqr803myqzO0uju/SFSv4sQqecUQqjnFPb3SK1P4nC2pTYIZP5GuY8HGYEl7dg8wozsbD+Z/TstXTs6v9tSpo7Z+Vmk9P5bSf1c6Y+RdXt2Vkr7u8z9uNl/196Pm53C1PIyOmSgDsKZOC3RDFgJ+/bk2bIvOvteqEshnF+C9O2Zl2n7uQ9KJK/L1AuUxz7u23IFoxwBisGKYYQ6p9mrSukA0MUIPGkHFEva163UdVNKSHt9b8veF8k3FbVAlrrmLA0u096MnNvUFcv1nM5tGW8AHrcJ6X7Fy35DAlIv5M6rTvonWaT2T15Pv60/haj+GgLSVk4VumthM/s+mki+HajrIv3/SaT/XwjnCrL/39wLsSHzPu6f1f0zZ2xNDxmpc5XJaz/3xq5CSz77OX+DAOD8FTnb3P9alvu+9v5ef3eex+y6c2fH7L4983GdLaIG0Csh9OywoT7gOH/bru1SZocUV1hRf1cZ2y3n9zL9eaQ07V2cF6jUpZSAlnwey3Uc0n6e9Of1PJ5cl86P77x2Zf7GZbxOpba7Hibjb8fZIS1O9qGwwzBCRSOEgO4ElyJKe9OE83ed8caZsa/XbXC9WWa96LkuMm/rjyHKKwQkw40dArLekOE8Td8OAX2x0pX1gTn7mwEp7e8tuS1tj6xtPQlk2Y/sdUw5boB3EOrswbr9HF4VGwnIZJhJv93KEVqk5b1/8jUgFSmT+yVfg9JOppV8JnXhvp8EYMEnNOi+8k1ExzBCfZ5wvWGlrvT9F/pyhSgAe0wIoPJKqxJmXcn6hignhhGiEumvIYqIqLdxdhkiIiIqK4YRIiIiKiuGESIiIiorhhEiIiIqK4YRIiIiKiuGESIiIiorhhEiIiIqK4YRIiIiKiuGESIiIiorhhEiIiIqK4YRIiIiKiuGESIiIiorhhEiIiIqq36xaq+zvHpLS0uZj4SIiIjy5bxvO+/jufSLMNLa2goAGDNmTJmPhIiIiArV2tqK2tranLcL2VVc6QMsy8KWLVtQXV0NIUTRHrelpQVjxozBpk2bUFNTU7THpXQ8z72H57p38Dz3Dp7n3lHK8yylRGtrK0aNGgVNy90zpF9URjRNw+jRo0v2+DU1NfxF7wU8z72H57p38Dz3Dp7n3lGq89xZRcTBDqxERERUVgwjREREVFZ7dBgJBoP48Y9/jGAwWO5DGdB4nnsPz3Xv4HnuHTzPvaMvnOd+0YGViIiIBq49ujJCRERE5ccwQkRERGXFMEJERERlxTBCREREZcUwQkRERGW1R4eRO++8E+PGjUMoFMKMGTPw5ptvlvuQ+rWbbroJhx12GKqrqzFs2DCcccYZWLVqVdo+0WgUl112GQYPHoyqqiqcffbZ2L59e5mOeGC4+eabIYTAggULktt4notj8+bNOO+88zB48GCEw2EcdNBBWLZsWfJ2KSWuu+46jBw5EuFwGLNnz8aaNWvKeMT9j2ma+NGPfoTx48cjHA5jwoQJuOGGG9IWVuN57p7XXnsNp59+OkaNGgUhBJ544om02/M5rw0NDZg3bx5qampQV1eHiy66CG1tbcU/WLmH+tvf/iYDgYD805/+JD/44AN5ySWXyLq6Orl9+/ZyH1q/NWfOHHnvvffK999/X65cuVKeeuqpcuzYsbKtrS25z6WXXirHjBkjX3rpJbls2TJ5xBFHyKOOOqqMR92/vfnmm3LcuHFy6tSp8jvf+U5yO89zzzU0NMi9995bXnDBBXLp0qVy3bp18vnnn5dr165N7nPzzTfL2tpa+cQTT8h33nlHfuELX5Djx4+XkUikjEfev9x4441y8ODB8umnn5br16+XjzzyiKyqqpK//vWvk/vwPHfPs88+K6+99lr5j3/8QwKQjz/+eNrt+ZzXk08+WR588MFyyZIl8vXXX5cTJ06U5557btGPdY8NI4cffri87LLLkt+bpilHjRolb7rppjIe1cCyY8cOCUC++uqrUkopm5qapN/vl4888khyn48++kgCkIsXLy7XYfZbra2tct9995UvvviinDlzZjKM8DwXx//7f/9PHnPMMTlvtyxLjhgxQt56663JbU1NTTIYDMqHHnqoNw5xQJg7d6688MIL07adddZZct68eVJKnudiyQwj+ZzXDz/8UAKQb731VnKf5557Tgoh5ObNm4t6fHtkM008Hsfy5csxe/bs5DZN0zB79mwsXry4jEc2sDQ3NwMABg0aBABYvnw5EolE2nnff//9MXbsWJ73brjsssswd+7ctPMJ8DwXy1NPPYXp06fjy1/+MoYNG4Zp06bh7rvvTt6+fv16bNu2Le0819bWYsaMGTzPBTjqqKPw0ksvYfXq1QCAd955B4sWLcIpp5wCgOe5VPI5r4sXL0ZdXR2mT5+e3Gf27NnQNA1Lly4t6vH0i1V7i23Xrl0wTRPDhw9P2z58+HB8/PHHZTqqgcWyLCxYsABHH300DjzwQADAtm3bEAgEUFdXl7bv8OHDsW3btjIcZf/1t7/9DStWrMBbb72VdRvPc3GsW7cOv//973HllVfif/7nf/DWW2/hiiuuQCAQwPz585Pn0ut1hOc5f1dffTVaWlqw//77Q9d1mKaJG2+8EfPmzQMAnucSyee8btu2DcOGDUu73efzYdCgQUU/93tkGKHSu+yyy/D+++9j0aJF5T6UAWfTpk34zne+gxdffBGhUKjchzNgWZaF6dOn42c/+xkAYNq0aXj//fdx1113Yf78+WU+uoHj4YcfxgMPPIAHH3wQBxxwAFauXIkFCxZg1KhRPM97kD2ymWbIkCHQdT1rdMH27dsxYsSIMh3VwHH55Zfj6aefxssvv4zRo0cnt48YMQLxeBxNTU1p+/O8F2b58uXYsWMHDjnkEPh8Pvh8Prz66qu444474PP5MHz4cJ7nIhg5ciSmTJmStm3y5MnYuHEjACTPJV9HeuYHP/gBrr76apxzzjk46KCD8PWvfx3f/e53cdNNNwHgeS6VfM7riBEjsGPHjrTbDcNAQ0ND0c/9HhlGAoEADj30ULz00kvJbZZl4aWXXsKRRx5ZxiPr36SUuPzyy/H4449j4cKFGD9+fNrthx56KPx+f9p5X7VqFTZu3MjzXoBZs2bhvffew8qVK5Nf06dPx7x585LXeZ577uijj84amr569WrsvffeAIDx48djxIgRaee5paUFS5cu5XkuQEdHBzQt/a1I13VYlgWA57lU8jmvRx55JJqamrB8+fLkPgsXLoRlWZgxY0ZxD6io3WH7kb/97W8yGAzK++67T3744Yfym9/8pqyrq5Pbtm0r96H1W//1X/8la2tr5SuvvCK3bt2a/Oro6Ejuc+mll8qxY8fKhQsXymXLlskjjzxSHnnkkWU86oHBPZpGSp7nYnjzzTelz+eTN954o1yzZo184IEHZEVFhfzrX/+a3Ofmm2+WdXV18sknn5Tvvvuu/OIXv8ghpwWaP3++3GuvvZJDe//xj3/IIUOGyKuuuiq5D89z97S2tsq3335bvv322xKA/OUvfynffvtt+emnn0op8zuvJ598spw2bZpcunSpXLRokdx33305tLfYfvOb38ixY8fKQCAgDz/8cLlkyZJyH1K/BsDz6957703uE4lE5H//93/L+vp6WVFRIc8880y5devW8h30AJEZRniei+Of//ynPPDAA2UwGJT777+//MMf/pB2u2VZ8kc/+pEcPny4DAaDctasWXLVqlVlOtr+qaWlRX7nO9+RY8eOlaFQSO6zzz7y2muvlbFYLLkPz3P3vPzyy56vyfPnz5dS5nded+/eLc8991xZVVUla2pq5De+8Q3Z2tpa9GMVUrqmuSMiIiLqZXtknxEiIiLqOxhGiIiIqKwYRoiIiKisGEaIiIiorBhGiIiIqKwYRoiIiKisGEaIiIiorBhGiIiIqKwYRoiIiKisGEaIiIiorBhGiIiIqKz+PxRsvgw79u8hAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def create_figure_3(df, plot_samples: bool = False, save: bool = False):\n",
    "    COLORS = {\"SFR (GP) BO\": \"steelblue\", \"GP Subset (GP) BO\": \"darkorange\"}\n",
    "\n",
    "    # Remove the vehicle seeds which failed\n",
    "    df = df[~((df[\"seed\"].isin([17])) & (df[\"dataset\"] == \"vehicle\"))]\n",
    "    \n",
    "    # Calculate mean and std (over seeds) for each combination of dataset and model\n",
    "    df_with_stats = (\n",
    "        df.groupby([\"dataset\", \"model\", \"num_inducing_percent\"])\n",
    "        .agg(\n",
    "            mean=(\"nlpd\", \"mean\"),\n",
    "            std=(\"nlpd\", \"std\"),\n",
    "            count=(\"nlpd\", \"count\"),\n",
    "        )\n",
    "        .reset_index()\n",
    "    )\n",
    "    \n",
    "    for dataset in DATASETS.keys():\n",
    "        fig, ax = plt.subplots()\n",
    "        for model_name in COLORS.keys():\n",
    "            df_model = df_with_stats[(df_with_stats[\"model\"] == model_name) & (df_with_stats[\"dataset\"] == dataset)]\n",
    "            if \"SFR\" in model_name:\n",
    "                model_name_not_tuned = \"SFR (GP)\"\n",
    "            elif \"GP\" in model_name:\n",
    "                model_name_not_tuned = \"GP Subset (GP)\"\n",
    "            else:\n",
    "                model_name_not_tuned = None\n",
    "            \n",
    "            df_model_not_tuned = df_with_stats[(df_with_stats[\"model\"] == model_name_not_tuned) & (df_with_stats[\"dataset\"] == dataset)]\n",
    "\n",
    "            means_tuned = df_model[\"mean\"].to_numpy()\n",
    "            means_not_tuned = df_model_not_tuned[\"mean\"].to_numpy()\n",
    "            # print(means_tuned.shape)\n",
    "            # print(means_not_tuned.shape)\n",
    "            \n",
    "            means = np.min(np.stack([means_tuned, means_not_tuned], 0), 0)\n",
    "            means_idx = np.argmin(np.stack([means_tuned, means_not_tuned], 0), 0)\n",
    "            means_std = np.where(means_idx, df_model_not_tuned[\"std\"],  df_model[\"std\"])\n",
    "            \n",
    "            ax.plot(\n",
    "                df_model[\"num_inducing_percent\"].to_numpy(),\n",
    "                # df_model[\"mean\"].to_numpy(),\n",
    "                means,\n",
    "                color=COLORS[model_name],\n",
    "                label=model_name,\n",
    "            )\n",
    "            if plot_samples:\n",
    "                num_inducing_levels = len(df[\"num_inducing_percent\"].unique()) - 1\n",
    "                df_ = df[(df[\"model\"] == model_name) & (df[\"dataset\"] == dataset)]\n",
    "                ax.plot(\n",
    "                    df_[\"num_inducing_percent\"].to_numpy().reshape(-1, num_inducing_levels).T,\n",
    "                    df_[\"nlpd\"].to_numpy().reshape(-1, num_inducing_levels).T,\n",
    "                    # color=COLORS[model_name],\n",
    "                    label=df_[\"seed\"].to_numpy().reshape(-1, num_inducing_levels)[:, 0],\n",
    "                )\n",
    "                ax.legend()\n",
    "            ax.fill_between(\n",
    "                df_model[\"num_inducing_percent\"].to_numpy(),\n",
    "                means - means_std,\n",
    "                means + means_std,\n",
    "                # df_model[\"mean\"].to_numpy() - df_model[\"std\"].to_numpy(),\n",
    "                # df_model[\"mean\"].to_numpy() + df_model[\"std\"].to_numpy(),\n",
    "                alpha=0.1,\n",
    "                color=COLORS[model_name],\n",
    "            )\n",
    "        plt.locator_params(axis=\"y\", nbins=4)\n",
    "        ax.set_title(\"\\sc{\" + DATASETS[dataset] + \"}\")\n",
    "        # ax.set_xlabel(\"$M$ as % of N\")\n",
    "        # ax.set_xlabel(\"$\\\\frac{M}{N} \\\\times 100$\")\n",
    "        # ax.set_ylabel(\"NLPD\")\n",
    "        ax.set_xlim(-5, 105)\n",
    "        if save:\n",
    "            save_dir = \"./figs\"\n",
    "            plt.savefig(os.path.join(save_dir, dataset + \".png\"))\n",
    "            tikzplotlib.save(\n",
    "                os.path.join(save_dir, dataset + \".tex\"),\n",
    "                axis_width=\"\\\\figurewidth\",\n",
    "                axis_height=\"\\\\figureheight\",\n",
    "            )\n",
    "\n",
    "create_figure_3(df, plot_samples=False, save=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2c1b994e-a5be-4537-8a61-1d3b35806ce0",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}