{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "15e07b1a",
   "metadata": {},
   "source": [
    "# Factor analysis & predictive feature importance\n",
    "\n",
    "**Goal.** Identify which characteristics in the sample carry the most predictive signal for communication-apprehension (CA) scores, and summarize the latent factor structure of the PRCA Likert items.\n",
    "\n",
    "| Analysis | What it answers |\n",
    "|---|---|\n",
    "| PRCA item factor analysis (PCA + FA) | How the group / interpersonal items cluster; which items load most strongly |\n",
    "| Random Forest impurity importance | Which encoded predictors the RF relies on for `gt_group_ca` / `gt_interpersonal_ca` |\n",
    "| Permutation importance | Which **raw** Prolific/Qualtrics fields most hurt MAE when shuffled |\n",
    "| Predictor PCA | Dominant directions of variation among persona covariates |\n",
    "\n",
    "Supporting code: [`src/ca_personas/feature_importance.py`](../src/ca_personas/feature_importance.py)."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6976cca9",
   "metadata": {},
   "source": [
    "## 1. Setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "2cd232af",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-07-25T15:40:04.188558Z",
     "iopub.status.busy": "2026-07-25T15:40:04.188439Z",
     "iopub.status.idle": "2026-07-25T15:40:05.241119Z",
     "shell.execute_reply": "2026-07-25T15:40:05.240444Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Project root: /workspace\n",
      "Data source: /tmp/sibling_data File A/B/C\n",
      "Prolific: ['/tmp/sibling_data/PRCAProlificExport_FileA.csv', '/tmp/sibling_data/PRCAProlificExport_FileB.csv']\n",
      "Qualtrics: /tmp/sibling_data/PRCAQualtricsExport_FileC.csv\n"
     ]
    }
   ],
   "source": [
    "from __future__ import annotations\n",
    "\n",
    "import sys\n",
    "from pathlib import Path\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "\n",
    "ROOT = Path.cwd()\n",
    "if not (ROOT / \"src\").exists() and (ROOT.parent / \"src\").exists():\n",
    "    ROOT = ROOT.parent\n",
    "sys.path.insert(0, str(ROOT / \"src\"))\n",
    "\n",
    "from ca_personas.feature_importance import (\n",
    "    likert_item_matrix,\n",
    "    predictor_feature_importance,\n",
    "    predictor_pca,\n",
    "    run_factor_and_importance_bundle,\n",
    "    run_item_factor_analysis,\n",
    ")\n",
    "from ca_personas.load import load_and_prepare, load_full_cohort, load_qualtrics\n",
    "from ca_personas.ml_baseline import prepare_modeling_frame\n",
    "from ca_personas.paths import cohort_source_label, full_cohort_paths\n",
    "\n",
    "PROLIFIC, QUALTRICS = full_cohort_paths()\n",
    "OUT_DIR = ROOT / \"outputs\" / \"feature_importance\"\n",
    "TIER = \"transit\"  # fullest tabular persona tier for importance\n",
    "\n",
    "pd.set_option(\"display.max_colwidth\", 80)\n",
    "print(\"Project root:\", ROOT)\n",
    "print(\"Data source:\", cohort_source_label())\n",
    "print(\"Prolific:\", [str(p) for p in PROLIFIC])\n",
    "print(\"Qualtrics:\", QUALTRICS)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "99a3b140",
   "metadata": {},
   "source": [
    "## 2. Load data\n",
    "\n",
    "- **Likert factor analysis** uses all Qualtrics rows with complete PRCA items (larger N).\n",
    "- **Predictor importance** uses the Prolific\u2194Qualtrics inner join (same modeling frame as stage-one ML)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e4ff2c6c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-07-25T15:40:05.242776Z",
     "iopub.status.busy": "2026-07-25T15:40:05.242608Z",
     "iopub.status.idle": "2026-07-25T15:40:05.636732Z",
     "shell.execute_reply": "2026-07-25T15:40:05.636210Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Complete PRCA Likert rows for factor analysis: 241\n",
      "Joined modeling rows (transit tier): 241\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>participant_id</th>\n",
       "      <th>Age</th>\n",
       "      <th>Sex</th>\n",
       "      <th>Employment status</th>\n",
       "      <th>gt_group_ca</th>\n",
       "      <th>gt_interpersonal_ca</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ebab12e9878344fdc93a8c1cad27ff2f709e12b90f0496de3332f508f3b2b9c9</td>\n",
       "      <td>36.0</td>\n",
       "      <td>Male</td>\n",
       "      <td>Part-Time</td>\n",
       "      <td>14</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>b6954799ca47bc836ac02ab86e4710c3b0e096d7ace449b81aaa1d5845431821</td>\n",
       "      <td>41.0</td>\n",
       "      <td>Male</td>\n",
       "      <td>Other</td>\n",
       "      <td>17</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>6ac07c3845b0ce5cac4ab3569ea70236b7cfa8fd51576e12226279a282182dcb</td>\n",
       "      <td>39.0</td>\n",
       "      <td>Male</td>\n",
       "      <td>Full-Time</td>\n",
       "      <td>11</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>c67b3266ecd4af12979c1f67d1efffd0d3e9bfbc567d211e5b7a0724b7890b7d</td>\n",
       "      <td>61.0</td>\n",
       "      <td>Male</td>\n",
       "      <td>Other</td>\n",
       "      <td>10</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1cefb8d34b92489e8ffd9cf34c469964f0527b1cae33e00a8a52f4aef0d66679</td>\n",
       "      <td>45.0</td>\n",
       "      <td>Male</td>\n",
       "      <td>Full-Time</td>\n",
       "      <td>8</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>236</th>\n",
       "      <td>35318e96d9b881e070c4e3197b0078f084d77f314e457e48d4a138467510c821</td>\n",
       "      <td>53.0</td>\n",
       "      <td>Male</td>\n",
       "      <td>Other</td>\n",
       "      <td>15</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>237</th>\n",
       "      <td>5dcb913ef602a61446ee04e04fc6db188af5a2dfcaaf3abd9e9bc80078ba6623</td>\n",
       "      <td>63.0</td>\n",
       "      <td>Female</td>\n",
       "      <td>Full-Time</td>\n",
       "      <td>12</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>238</th>\n",
       "      <td>2d4ac125c9c68840a8575bd1668214b40367b4f534f739c84503e268fdd52d5a</td>\n",
       "      <td>23.0</td>\n",
       "      <td>Female</td>\n",
       "      <td>Full-Time</td>\n",
       "      <td>15</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>239</th>\n",
       "      <td>9e261399ca55b7c3ff0b2c3bc8d28b112142fc2147c4fdca061066cfd2db545e</td>\n",
       "      <td>32.0</td>\n",
       "      <td>Male</td>\n",
       "      <td>Full-Time</td>\n",
       "      <td>13</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>240</th>\n",
       "      <td>32284ade76d6973e58641037ac723bba4bb21173da4999ad6f751179232b46e5</td>\n",
       "      <td>64.0</td>\n",
       "      <td>Male</td>\n",
       "      <td>Other</td>\n",
       "      <td>16</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>241 rows \u00d7 6 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                       participant_id   Age  \\\n",
       "0    ebab12e9878344fdc93a8c1cad27ff2f709e12b90f0496de3332f508f3b2b9c9  36.0   \n",
       "1    b6954799ca47bc836ac02ab86e4710c3b0e096d7ace449b81aaa1d5845431821  41.0   \n",
       "2    6ac07c3845b0ce5cac4ab3569ea70236b7cfa8fd51576e12226279a282182dcb  39.0   \n",
       "3    c67b3266ecd4af12979c1f67d1efffd0d3e9bfbc567d211e5b7a0724b7890b7d  61.0   \n",
       "4    1cefb8d34b92489e8ffd9cf34c469964f0527b1cae33e00a8a52f4aef0d66679  45.0   \n",
       "..                                                                ...   ...   \n",
       "236  35318e96d9b881e070c4e3197b0078f084d77f314e457e48d4a138467510c821  53.0   \n",
       "237  5dcb913ef602a61446ee04e04fc6db188af5a2dfcaaf3abd9e9bc80078ba6623  63.0   \n",
       "238  2d4ac125c9c68840a8575bd1668214b40367b4f534f739c84503e268fdd52d5a  23.0   \n",
       "239  9e261399ca55b7c3ff0b2c3bc8d28b112142fc2147c4fdca061066cfd2db545e  32.0   \n",
       "240  32284ade76d6973e58641037ac723bba4bb21173da4999ad6f751179232b46e5  64.0   \n",
       "\n",
       "        Sex Employment status  gt_group_ca  gt_interpersonal_ca  \n",
       "0      Male         Part-Time           14                   16  \n",
       "1      Male             Other           17                   17  \n",
       "2      Male         Full-Time           11                   12  \n",
       "3      Male             Other           10                    6  \n",
       "4      Male         Full-Time            8                   12  \n",
       "..      ...               ...          ...                  ...  \n",
       "236    Male             Other           15                   14  \n",
       "237  Female         Full-Time           12                    6  \n",
       "238  Female         Full-Time           15                   17  \n",
       "239    Male         Full-Time           13                    6  \n",
       "240    Male             Other           16                    9  \n",
       "\n",
       "[241 rows x 6 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "qualtrics = load_qualtrics(QUALTRICS)\n",
    "participants = load_and_prepare(PROLIFIC, QUALTRICS, how=\"inner\")\n",
    "# Restrict FA to the analytic matched cohort (not Qualtrics-only rows).\n",
    "analytic_ids = set(participants[\"participant_id\"].astype(str))\n",
    "qual_analytic = qualtrics[qualtrics[\"participant_id\"].astype(str).isin(analytic_ids)].copy()\n",
    "items = likert_item_matrix(qual_analytic, scored_direction=True)\n",
    "model_df = prepare_modeling_frame(participants, tier=TIER)\n",
    "\n",
    "assert len(participants) >= 100, f\"Expected full cohort; got N={len(participants)}\"\n",
    "print(f\"Complete PRCA Likert rows for factor analysis: {len(items)}\")\n",
    "print(f\"Joined modeling rows ({TIER} tier): {len(model_df)}\")\n",
    "model_df[[\"participant_id\", \"Age\", \"Sex\", \"Employment status\", \"gt_group_ca\", \"gt_interpersonal_ca\"]]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7efd5b36",
   "metadata": {},
   "source": [
    "## 3. Factor structure of PRCA items\n",
    "\n",
    "Items are scored so higher values = higher apprehension (comfort items reverse-coded). We report PCA variance explained and FactorAnalysis loadings."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "20cffd3c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-07-25T15:40:05.638476Z",
     "iopub.status.busy": "2026-07-25T15:40:05.638340Z",
     "iopub.status.idle": "2026-07-25T15:40:05.652638Z",
     "shell.execute_reply": "2026-07-25T15:40:05.652103Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using n_factors=2 on n=241 rows\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>component</th>\n",
       "      <th>variance_explained</th>\n",
       "      <th>cumulative_variance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>PC1</td>\n",
       "      <td>0.662968</td>\n",
       "      <td>0.662968</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>PC2</td>\n",
       "      <td>0.078097</td>\n",
       "      <td>0.741065</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  component  variance_explained  cumulative_variance\n",
       "0       PC1            0.662968             0.662968\n",
       "1       PC2            0.078097             0.741065"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "item_fa = run_item_factor_analysis(items, n_factors=2)\n",
    "print(f\"Using n_factors={item_fa['n_factors']} on n={item_fa['n_rows']} rows\")\n",
    "item_fa[\"pca_variance\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "a8608488",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-07-25T15:40:05.653637Z",
     "iopub.status.busy": "2026-07-25T15:40:05.653530Z",
     "iopub.status.idle": "2026-07-25T15:40:05.658388Z",
     "shell.execute_reply": "2026-07-25T15:40:05.657266Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Factor1</th>\n",
       "      <th>Factor2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Group: dislike discussions</th>\n",
       "      <td>0.694467</td>\n",
       "      <td>0.303157</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Group: tense/nervous</th>\n",
       "      <td>0.820549</td>\n",
       "      <td>-0.067303</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Group: tense with new people</th>\n",
       "      <td>0.803653</td>\n",
       "      <td>-0.254942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Group: comfortable (rev)</th>\n",
       "      <td>0.843880</td>\n",
       "      <td>0.262264</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Group: like involvement (rev)</th>\n",
       "      <td>0.844849</td>\n",
       "      <td>0.379075</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Group: calm/relaxed (rev)</th>\n",
       "      <td>0.852267</td>\n",
       "      <td>0.008962</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Interpersonal: nervous w/ acquaintance</th>\n",
       "      <td>0.741367</td>\n",
       "      <td>-0.460102</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Interpersonal: tense in conversations</th>\n",
       "      <td>0.785573</td>\n",
       "      <td>-0.227920</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Interpersonal: afraid to speak up</th>\n",
       "      <td>0.766809</td>\n",
       "      <td>-0.112527</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Interpersonal: no fear speaking (rev)</th>\n",
       "      <td>0.750037</td>\n",
       "      <td>-0.005941</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Interpersonal: calm in conversations (rev)</th>\n",
       "      <td>0.842288</td>\n",
       "      <td>-0.109984</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Interpersonal: relaxed w/ acquaintance (rev)</th>\n",
       "      <td>0.810580</td>\n",
       "      <td>-0.142373</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                               Factor1   Factor2\n",
       "Group: dislike discussions                    0.694467  0.303157\n",
       "Group: tense/nervous                          0.820549 -0.067303\n",
       "Group: tense with new people                  0.803653 -0.254942\n",
       "Group: comfortable (rev)                      0.843880  0.262264\n",
       "Group: like involvement (rev)                 0.844849  0.379075\n",
       "Group: calm/relaxed (rev)                     0.852267  0.008962\n",
       "Interpersonal: nervous w/ acquaintance        0.741367 -0.460102\n",
       "Interpersonal: tense in conversations         0.785573 -0.227920\n",
       "Interpersonal: afraid to speak up             0.766809 -0.112527\n",
       "Interpersonal: no fear speaking (rev)         0.750037 -0.005941\n",
       "Interpersonal: calm in conversations (rev)    0.842288 -0.109984\n",
       "Interpersonal: relaxed w/ acquaintance (rev)  0.810580 -0.142373"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fa_loadings = item_fa[\"fa_loadings\"]\n",
    "fa_loadings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "5b3d3c1a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-07-25T15:40:05.659754Z",
     "iopub.status.busy": "2026-07-25T15:40:05.659630Z",
     "iopub.status.idle": "2026-07-25T15:40:05.821004Z",
     "shell.execute_reply": "2026-07-25T15:40:05.820520Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>item</th>\n",
       "      <th>communality_pca</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Group: dislike discussions</td>\n",
       "      <td>0.344216</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Interpersonal: nervous w/ acquaintance</td>\n",
       "      <td>0.317527</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Group: like involvement (rev)</td>\n",
       "      <td>0.267517</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Group: comfortable (rev)</td>\n",
       "      <td>0.205303</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Interpersonal: tense in conversations</td>\n",
       "      <td>0.153162</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Group: tense with new people</td>\n",
       "      <td>0.149656</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Interpersonal: calm in conversations (rev)</td>\n",
       "      <td>0.103213</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Interpersonal: relaxed w/ acquaintance (rev)</td>\n",
       "      <td>0.102112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Group: calm/relaxed (rev)</td>\n",
       "      <td>0.097897</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Interpersonal: afraid to speak up</td>\n",
       "      <td>0.091931</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Group: tense/nervous</td>\n",
       "      <td>0.089731</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Interpersonal: no fear speaking (rev)</td>\n",
       "      <td>0.077734</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                            item  communality_pca\n",
       "0                     Group: dislike discussions         0.344216\n",
       "6         Interpersonal: nervous w/ acquaintance         0.317527\n",
       "4                  Group: like involvement (rev)         0.267517\n",
       "3                       Group: comfortable (rev)         0.205303\n",
       "7          Interpersonal: tense in conversations         0.153162\n",
       "2                   Group: tense with new people         0.149656\n",
       "10    Interpersonal: calm in conversations (rev)         0.103213\n",
       "11  Interpersonal: relaxed w/ acquaintance (rev)         0.102112\n",
       "5                      Group: calm/relaxed (rev)         0.097897\n",
       "8              Interpersonal: afraid to speak up         0.091931\n",
       "1                           Group: tense/nervous         0.089731\n",
       "9          Interpersonal: no fear speaking (rev)         0.077734"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(8, 5))\n",
    "abs_load = fa_loadings.abs()\n",
    "abs_load.plot(kind=\"barh\", ax=ax)\n",
    "ax.set_xlabel(\"|Factor loading|\")\n",
    "ax.set_title(\"PRCA item |loadings| (FactorAnalysis)\")\n",
    "ax.invert_yaxis()\n",
    "fig.tight_layout()\n",
    "plt.show()\n",
    "\n",
    "item_fa[\"communalities\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "56298cab",
   "metadata": {},
   "source": [
    "## 4. Predictive feature importance (persona covariates)\n",
    "\n",
    "For each CA target we fit a Random Forest on the `transit` tier feature set and report:\n",
    "\n",
    "1. **Impurity importance** on one-hot/scaled encoded columns  \n",
    "2. **Permutation importance** on the original survey fields (MAE-based)\n",
    "\n",
    "::: {.callout-warning}\n",
    "With a small joined excerpt, importances are unstable. Re-run on the full cohort before strong substantive claims; the ranking API stays the same.\n",
    ":::"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "d6764c20",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-07-25T15:40:05.822740Z",
     "iopub.status.busy": "2026-07-25T15:40:05.822602Z",
     "iopub.status.idle": "2026-07-25T15:40:25.324922Z",
     "shell.execute_reply": "2026-07-25T15:40:25.324043Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Top raw features \u2014 group CA\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>feature</th>\n",
       "      <th>permutation_importance_mean</th>\n",
       "      <th>permutation_importance_std</th>\n",
       "      <th>target</th>\n",
       "      <th>tier</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Q28</td>\n",
       "      <td>1.899627</td>\n",
       "      <td>0.125196</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>LocationLatitude</td>\n",
       "      <td>1.364317</td>\n",
       "      <td>0.081851</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Age</td>\n",
       "      <td>1.081285</td>\n",
       "      <td>0.064105</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>LocationLongitude</td>\n",
       "      <td>1.011894</td>\n",
       "      <td>0.044757</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Employment status</td>\n",
       "      <td>0.957730</td>\n",
       "      <td>0.078861</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Q26</td>\n",
       "      <td>0.697526</td>\n",
       "      <td>0.049419</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Sex</td>\n",
       "      <td>0.200735</td>\n",
       "      <td>0.023698</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Q21</td>\n",
       "      <td>0.191117</td>\n",
       "      <td>0.033292</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Country of residence</td>\n",
       "      <td>0.166201</td>\n",
       "      <td>0.017538</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Q27</td>\n",
       "      <td>0.156788</td>\n",
       "      <td>0.023155</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                feature  permutation_importance_mean  \\\n",
       "0                   Q28                     1.899627   \n",
       "1      LocationLatitude                     1.364317   \n",
       "2                   Age                     1.081285   \n",
       "3     LocationLongitude                     1.011894   \n",
       "4     Employment status                     0.957730   \n",
       "5                   Q26                     0.697526   \n",
       "6                   Sex                     0.200735   \n",
       "7                   Q21                     0.191117   \n",
       "8  Country of residence                     0.166201   \n",
       "9                   Q27                     0.156788   \n",
       "\n",
       "   permutation_importance_std       target     tier  \n",
       "0                    0.125196  gt_group_ca  transit  \n",
       "1                    0.081851  gt_group_ca  transit  \n",
       "2                    0.064105  gt_group_ca  transit  \n",
       "3                    0.044757  gt_group_ca  transit  \n",
       "4                    0.078861  gt_group_ca  transit  \n",
       "5                    0.049419  gt_group_ca  transit  \n",
       "6                    0.023698  gt_group_ca  transit  \n",
       "7                    0.033292  gt_group_ca  transit  \n",
       "8                    0.017538  gt_group_ca  transit  \n",
       "9                    0.023155  gt_group_ca  transit  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Top raw features \u2014 interpersonal CA\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>feature</th>\n",
       "      <th>permutation_importance_mean</th>\n",
       "      <th>permutation_importance_std</th>\n",
       "      <th>target</th>\n",
       "      <th>tier</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Age</td>\n",
       "      <td>1.525612</td>\n",
       "      <td>0.102048</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Employment status</td>\n",
       "      <td>1.232257</td>\n",
       "      <td>0.102118</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Q28</td>\n",
       "      <td>1.146547</td>\n",
       "      <td>0.075376</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>LocationLatitude</td>\n",
       "      <td>1.114654</td>\n",
       "      <td>0.057168</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>LocationLongitude</td>\n",
       "      <td>1.063704</td>\n",
       "      <td>0.062338</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Q26</td>\n",
       "      <td>0.744586</td>\n",
       "      <td>0.049029</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Sex</td>\n",
       "      <td>0.307918</td>\n",
       "      <td>0.030717</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Student status</td>\n",
       "      <td>0.219858</td>\n",
       "      <td>0.030537</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Country of residence</td>\n",
       "      <td>0.158381</td>\n",
       "      <td>0.017978</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Q21</td>\n",
       "      <td>0.135057</td>\n",
       "      <td>0.021949</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                feature  permutation_importance_mean  \\\n",
       "0                   Age                     1.525612   \n",
       "1     Employment status                     1.232257   \n",
       "2                   Q28                     1.146547   \n",
       "3      LocationLatitude                     1.114654   \n",
       "4     LocationLongitude                     1.063704   \n",
       "5                   Q26                     0.744586   \n",
       "6                   Sex                     0.307918   \n",
       "7        Student status                     0.219858   \n",
       "8  Country of residence                     0.158381   \n",
       "9                   Q21                     0.135057   \n",
       "\n",
       "   permutation_importance_std               target     tier  \n",
       "0                    0.102048  gt_interpersonal_ca  transit  \n",
       "1                    0.102118  gt_interpersonal_ca  transit  \n",
       "2                    0.075376  gt_interpersonal_ca  transit  \n",
       "3                    0.057168  gt_interpersonal_ca  transit  \n",
       "4                    0.062338  gt_interpersonal_ca  transit  \n",
       "5                    0.049029  gt_interpersonal_ca  transit  \n",
       "6                    0.030717  gt_interpersonal_ca  transit  \n",
       "7                    0.030537  gt_interpersonal_ca  transit  \n",
       "8                    0.017978  gt_interpersonal_ca  transit  \n",
       "9                    0.021949  gt_interpersonal_ca  transit  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "importances = predictor_feature_importance(\n",
    "    participants,\n",
    "    tier=TIER,\n",
    "    n_repeats=30,\n",
    "    random_state=42,\n",
    ")\n",
    "\n",
    "group_perm = importances[\"gt_group_ca__raw_permutation\"]\n",
    "inter_perm = importances[\"gt_interpersonal_ca__raw_permutation\"]\n",
    "print(\"Top raw features \u2014 group CA\")\n",
    "display(group_perm.head(10))\n",
    "print(\"Top raw features \u2014 interpersonal CA\")\n",
    "display(inter_perm.head(10))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "4c1526cf",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-07-25T15:40:25.326731Z",
     "iopub.status.busy": "2026-07-25T15:40:25.326590Z",
     "iopub.status.idle": "2026-07-25T15:40:25.490024Z",
     "shell.execute_reply": "2026-07-25T15:40:25.489108Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABKYAAAIDCAYAAAA+HxaBAAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjExLjEsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvctoD+AAAAAlwSFlzAAAPYQAAD2EBqD+naQAA2clJREFUeJzs3Xd8T+f///HnOwkRIQkJgtixg9YmNrH3VqNWS6s2FV3oR40abT/a0potH9Te1KZG7L33ihWyFAnJ+f3hl/P1loSE8DYe99stt5v3da5zndd1qs2rr/d1rmMxDMMQAAAAAAAA8IrZ2ToAAAAAAAAAvJsoTAEAAAAAAMAmKEwBAAAAAADAJihMAQAAAAAAwCYoTAEAAAAAAMAmKEwBAAAAAADAJihMAQAAAAAAwCYoTAEAAAAAAMAmKEwBAAAAAADAJihMAQDeCf7+/nJwcLB1GHhFSpcurUqVKlm1eXl5qX379kl6nZcxZkIYhqEhQ4Yoe/bssre313vvvffKY3iT7Ny5U8mSJdPx48dtHcpTXbt2TY0bN1aaNGlksVg0evRorVq1ShaLRRs3bkz0eBs3bpTFYtHatWuf2ff06dOyWCyaNGlSoq6xefNmOTo66syZM4mODwAAicIUACCRRowYIYvFIovFoh07dsTZ57333pPFYpG3t/dLi6NXr15KkSLFSxv/dfWq5/2u3ufHeXh4qHPnzrYOw8r8+fM1ePBgTZgwQQ8ePND+/ftfynVex7k/j969e6tVq1bKly+frUN5ql69eunEiRM6evSoDMNQv379bB3SM1WoUEEVKlTQ559/butQAABvKL46BgA8F2dnZ02bNk2lSpWyat+3b58OHDggZ2dnG0UWtxEjRmjEiBG2DgM2dPny5TdizITYtGmT0qZNq5o1a9rk+m+SjRs3atu2bfrxxx9tHcozrV27Vq1bt1bGjBnNtpo1a8owDBtG9WzdunVTo0aNdPz48de++AcAeP2wYgoA8FwaN26s2bNn6/79+1btU6dOlbe3N48WAS9RUFCQnJycbB3GG2HChAnKmzevSpQoYetQnioqKkq3bt16I/+51q5dW2nSpNFvv/1m61AAAG8gClMAgOfSrl07hYWFafHixWZbZGSkZs6c+dQ9d9avX6/KlSsrderUSpkypUqVKqUFCxZY9Tl79qw++OADZc6cWSlTppSPj4+GDh2qu3fvSpIaNmyon376SREREeZjhRaLRefPn4/3unHtMdW1a1e5ubkpJCREH3zwgVxdXZUxY0aNHTtWknT79m21bt1abm5ucnd31+eff67o6Og4xwgODlarVq3k6uqqtGnTqkOHDrp9+/Zzzf/xuNq3b6+0adMqf/78z5z3jz/+aNXu4uKiihUrauXKlXGOf+fOHXXs2FFubm5yc3NT27ZtFRYWZvZ7nvucmPsR3zxjTJ8+XSVLllTKlCmVKlUq+fn5aefOnVZjREREqG/fvvL09JSzs7Nq1Kihs2fPxhlbfPtBzZgxQ2XLllXq1Knl7u6uBg0a6ODBg7pz544sFotu3bqlyZMnm/MvV65cvGMWLVo03qJshQoVlDdvXqu2hMzxcZcvX5bFYtHs2bN15coVM6YJEyYkaszixYub5zo4OChTpkxq166drly5IknPnHtISIgsFkucqxCLFy+uatWqWbWlSpVKn332mbZu3aqyZcvKyclJQ4cONcfq06ePcuTIoeTJkytTpkzq1q2bQkNDzfPv37+vgQMHytvbW05OTsqRI4c6duyoCxcuxHuvpEfFnpUrV6pKlSqxji1atEgWi0Xbt2/X999/r+zZs8vOzk7nz5/X8ePHrf7Op0iRQj4+Pho9erT534B79+4pRYoU+uabb6yuF7M/1ON/D//66y9ZLBbt27cvzjhHjBhh/vdp5MiRVteO+Xlyj6mE3Lf4HD16VDVr1lTKlCmVIUMGDRgwQFFRUbH6JfS+J0+eXOXKldOSJUueeW0AAJ5EYQoA8Fy8vLxUrVo1TZs2zWxbsmSJgoOD1a5duzjPWbp0qfz8/JQtWzYdPnxYZ86cUcWKFdWkSROrb9pr1qypq1evav369bp9+7bmzZunBw8emEWwRYsWqWfPnnJ0dJRhGOZP9uzZn2suPXr0UMeOHXXp0iV999136tu3r+bMmaNOnTrpww8/1MWLF/XDDz9o1KhRmjp1apxjdO3aVW3bttWlS5c0f/58rV27VjVq1NCDBw8SPf/Hx2zatKnOnDmjzz///Jnz7tWrl9kWFRWlQ4cO6f3331eDBg20d+/eWON3795dTZs21aVLl/TXX39p8eLFVvvEvMh9Tsj9iG+ekjRw4EB17txZrVq10rlz53Tq1Cnlz59fFSpU0O7du81zP/jgA02aNEk///yzrl69qiFDhujTTz/VvXv3nhmjJPXv318dO3ZUnTp1dOTIEZ0+fVofffSRfvrpJ6VKlUqGYcjd3V2dOnUy579ly5Z4x+vUqZMOHDigPXv2WLWfOnVK//zzjzp16mS2JXSOj/Py8pJhGGrRooUyZ85sxtS1a9dEjbl7927z3PDwcC1YsEBHjx5VgwYN9PDhw+ea+7OcPn1a33//vSZPnqyTJ0+qSJEiCgsLk6+vr5YvX67Jkyfr9u3bWrp0qbZs2SI/Pz/z70v//v01efJks8+WLVtUqVIljRs37qnXPHjwoMLCwlS8ePF4+4wZM0aGYWjXrl1aunSpHBwclC9fPqu/84GBgerXr5+++eYbs3Dt5OQkX19f/f3331b3NSQkRE5OTlq9erXZvnbtWnl4eMRbtPT39zfnOmDAAKtrP1lYlpTg+xaXy5cvq0KFCgoLC9POnTt1/PhxZc+eXV9//XWsvom57yVLltTZs2fN4iYAAAlmAACQCMOHDzckGceOHTNmzpxp2NvbG1euXDEMwzBq165t+Pn5GYZhGL6+vkauXLmszs2TJ4+RL18+Iyoqyqq9atWqhpubm3H37l3j8uXLhiRj8uTJT42jZ8+ehqOjY4LjHjBggGFvb2/V1qVLF0OSMW/ePKv2kiVLGs7Ozsbs2bOt2suUKWOUKVMmzjGmT59u1b5s2TJDkvHnn3+abQmZ/+Nj/vHHH7Hmkdh5G4ZhZM+e3fjss89ixTxnzhyrft27dzeSJ09uPHjw4Lmvl5j7Ed88jx07ZlgsFuPLL7+0ao+OjjaKFi1q1KhRwzAMw9ixY4chyfjhhx+s+gUEBBiSjIoVK1q1Z86c2fjwww/Nz4cOHTIsFovx+eefP3VO7u7uRqdOneI89uSYwcHBRooUKYxPPvnEqp+/v7/h4OBgXL16NVFzjE+LFi2MzJkzW7W96JhbtmwxJBkBAQFmW3xzDw4ONiQZw4cPj3WsWLFiRtWqVa3anJ2djVSpUhkhISFW7V9//bVhZ2dnHD582Kr98OHDVn+PfHx8jGbNmj01/rgsWrTIkGSsWLEi1rGFCxcakoxWrVoleLyuXbsa3t7e5ufhw4cbdnZ2xu3btw3DMIz//Oc/Ru7cuY2mTZsajRs3Nvtly5bNaNmy5VPHfvDggSHJGDBggFX7ypUrDUnGhg0bzLaE3rcNGzYYkow1a9aYfWL+Pb98+bLVuZ999pkhyZg4caLZlpj7PmnSJEOSsXXr1gT1BwAgBiumAADPrVGjRkqVKpWmT5+uq1ev6u+//1aHDh3i7Hv58mWdPHlSDRs2lJ2d9a+fpk2bKiQkRHv37lWGDBnk6emp7777TtOnT9eNGzde+jxq1apl9Tlfvnz6999/Y7Xnz58/3sfE6tevb/W5du3aSp48udavXy8p4fN/2pjP8u+//+rLL79Uvnz5lCJFCqtH706fPh2rf+3ata0++/j4KDIyMkk29H7W/Xha3+XLl8swDDVr1syq3WKxqEqVKtq0aZMkad26dXGeX6pUKXl6ej4zxpUrV8owDH3wwQfPnlACubm5qUmTJpo1a5a5/1pUVJT+/PNP1alTx4wroXNMjMSMefjwYTVr1kwZM2aUg4OD1WN6cf1dSQoVK1aUq6urVdvSpUtVoEABFSxY0Kq9YMGC8vT0NGMuUqSIli5dquHDh+vEiRMJvmZISIgkKXXq1PH2ie/fs2nTpqls2bJycXGxemTy7Nmz5uN8fn5+io6O1tq1ayVJa9askZ+fn/z8/LR+/XpFRUXp1KlTunDhgvz8/BIc97Mk9L7FZd26dSpWrJgyZ85s1d6wYcNYfRNz311cXCT93z0HACChKEwBAJ5bihQp1KJFC/3xxx/6888/lSpVKjVq1CjOvrdu3ZKkOAsGMW1BQUFycHDQ6tWrVaBAAX388cfKkCGDChYsqKFDhyb48azEcHFxUcqUKa3aYvZ/ivkfrcfb4/qfrhQpUsTqa7FYlC5dOgUFBUlK+PxjpEyZUm5ubomaS+vWrTV+/HiNHDlSgYGBioqKkmEY8vHxifVoj4uLS6w3JybV/1gm5H7EiGue165dkyQVK1ZMDg4Osre3l52dnezs7DR69Gjdv39fd+/eNe9phgwZYsUQV9uTYoqeT/4P+ovq3LmzQkJCNH/+fEmPCmCBgYFWj/EldI6JkdAxAwMD5evrq5CQEK1cuVJhYWEyDEP79++XpKc+BpYQRjxvkIvrPl+7dk1HjhyRg4ODVcwWi0XXrl0z/xmPGzdOHTt21A8//KB8+fIpY8aM6ty58zP3mIr5u/X43mkJiWvChAnq0KGD6tatqyNHjujBgwcyDEP9+vVTdHS0WZh6//335e7urtWrV+vOnTvavn27/Pz8VL16dYWEhGjXrl1as2aNJMXad+tFJPS+xeXWrVsJ/ncmMfc95h6nSZPmBWYGAHgXUZgCALyQ9u3b69ixYxo5cqRatmypFClSxNkvbdq0kqTr16/HOhbT5uHhIUkqVKiQli5dqpCQEG3ZskU1a9bUN998o+7duyd5/BaLJVHtcbl//36s//E1DEM3b96Uu7u7pMTNX5KSJUuW4OtLUmhoqJYsWaLPPvtMDRo0UNq0aWVnZyfDMOL8n8jEzC+xEnI/YsQ1z5j7cOLECT18+FBRUVFmMcD4//vupEyZ0hzraff0adKlSydJSb4nTsWKFeXt7a0pU6ZIkqZMmaKMGTNarcBL6BwTI6FjLlq0SGFhYfrll1/03nvvmdc5d+5cgq+VKlUq2dvbKzw8PNax+O5nfP+sS5UqpYcPH1rFHBNvzIsB0qRJo19++UXXr1/X0aNH5e/vryVLlqhixYp6+PBhvHFmy5ZN0v8V7RIa159//qkSJUroiy++UJYsWcyNyZ+8R3Z2dqpatapWr16tTZs2yTAMValSRdmzZ5e3t7dWr16tNWvWKG/evMqaNWu8MSRWQu9bXNzd3RP870xi7vvVq1clKUnnCQB4N1CYAgC8kDJlyihfvnwKDg5+6tv4smTJoty5c2vx4sWxVlTMnz9fbm5uKlq0qFW7o6OjfH19NWbMGFWoUEGbN282jzk7O5urGF4HS5cutfq8cuVKRUZGqmrVqpKeb/5xiW/eFotFhmHI0dHRqn3hwoVxFg8S6nnv87Pux9PUqVNHFotFf/3111P7xbxp7clr7dy586mFiCevM3PmzKf2c3Z2VkRExDPHi2GxWNSxY0dt2LBBAQEBWrZsmT788EOrt0ImdI6Jkdgxn/y78ueff8bqE9/cHRwclDVrVh0+fNiqfffu3QkqCsaoV6+e9u7dm+DHBy0Wi/Lnz6+ePXuqT58+unDhgi5evBhv/0KFCsnFxSXezeSf5sn7ExQUpFWrVsXq5+fnp4sXL+rnn39WyZIlzdWCfn5+WrlypTZs2JCkj/FJib9vj6tSpYr27NljFpJiPP6G1Scl5L7v3LlTOXPmTPIViACAtx+FKQDACzt27JgMw1Dp0qWf2m/UqFE6fvy4OnfurIsXL+ratWv64osvtGbNGg0bNkxOTk7atWuXmjZtqrVr1+rGjRu6d++e/v77b+3bt0+VK1c2x/Lx8VF0dLSWL18e52vOXyVXV1ctXbpUK1euVHh4uDZt2qSuXbvq/fffV4sWLcx+CZn/s8Q3bxcXF1WsWFHjx4/Xzp07defOHS1dulQ//PCDfHx8nntuz3OfE3o/nnZNf39/DRkyRCNHjtTFixd17949HT16VGPGjNGnn34q6dFeUg0bNtSgQYO0YMEChYWFaceOHRo8eLAKFy78zOsULFhQffr00Q8//KBhw4bp0qVLCgkJ0bJly6weu/Px8dHOnTsVGBiYoPlLj1YS2tnZqXnz5nrw4IE6duz4XHNMjISOWb16daVIkUL9+/fX9evXdfXqVX3xxRdWhbOEzP2jjz7SihUrNHv2bIWHh2vr1q0aNmxYov6+DRgwQHny5FHdunW1fPlyBQcH6/bt29q6das6d+6shQsXSnr0GNyMGTN07tw5RURE6OjRo1qwYIFy5sz51BU69vb2ql27dpx7mz1N/fr1tXXrVv3xxx+6c+eO9u/fr8aNG8dZWI0pOq1atUrVq1c326tXr66AgACFhoYmeWEqofctLp9//rmcnZ3VrFkzHTlyRCEhIfrtt9/iLCgm9L5HRkZqy5Ytid4XDwAASbyVDwCQOI+/le9p4norn2EYxurVq40KFSoYzs7ORooUKYwSJUpYvR0uKirKWLBggVGjRg0jXbp0hrOzs1GwYEFj2LBhRkREhNnv4cOHRufOnQ0PDw/DYrEYkoxz587FG098b+VzdXWN1bdbt26Gs7NzrPa43lAXM8atW7eMZs2aGalTpzbc3NyMdu3aGTdv3kz0/J8W17PmfeXKFaN58+ZG2rRpDRcXF6NRo0bG5cuXY70lLb7xZ82aZUgy9u3bl6DrxSUx9+Np8zQMw5g7d65RsWJFw8XFxUiZMqXh4+NjfP7558alS5fMPvfu3TN69eplpE+f3nBycjKqVq1qnDp1yihVqtQz38oXY9q0aUaJEiUMJycnI126dEaDBg2MgwcPmsePHj1q+Pr6Gk5OToYkw9fX95ljGoZh1KtXz5BkVKhQ4YXmGJe43sqXmDFXrVplFC1a1HBycjKyZMlifPvtt8aRI0cMScbUqVMTNPfIyEijR48ehoeHh+Hs7GzUq1fPuHLlSrxv5evWrVuc8YaHhxtffvmlkS9fPsPR0dHw8PAwKlSoYEydOtX8d37v3r3Ghx9+aOTIkcNIkSKFkS1bNqNLly7GxYsXn3qfDMMwNm7caEgydu7cadUe81a+7du3xzrnwYMHxldffWVkyZLFcHJyMkqWLGmsX7/e+PLLLw1JVm+uNAzDyJ07d6w30oWEhBgODg6Gg4ODERoa+sw4E/NWPsNI2H2L6618hvHojZR+fn7m3/m+ffsax44di/VWvoTe95h7+azfCwAAxMViGK/JMxAAALyBunbtqtmzZ/Mmqv+P+4HXUbly5ZQzZ844H1fEi/Pz85OLi4u54T8AAInBo3wAAAB4q/3www+aNWuWjh07ZutQ3jqbN2/W5s2b9f3339s6FADAGyr2ZgIAAADAW6REiRJ68OCBrcN4K1WoUCFRLwcAAOBJrJgCAAAAAACATbDHFAAAAAAAAGyCFVMAAAAAAACwCQpTAAAAAAAAsAkKUwAAAAAAALAJClMAAAAAAACwCQpTAAAAAAAAsAkKUwAAAAAAALAJClMAAAAAAACwCQpTAAAAAAAAsAkKUwAAAAAAALAJClMAAAAAAACwCQpTAAAAAAAAsAkKUwAAAAAAALAJClMAAAAAAACwCQpTAAAAAAAAsAkKUwAAAAAAALAJClMAAAAAAACwCQpTAAAAAAAAsAkKUwAAAAAAALAJClMAAAAAAACwCQpTAAAAAAAAsAkKUwAAAAAAALAJClMAAAAAAACwCQpTAAAAAAAAsAkKUwAAAAAAALAJClMAAAAAAACwCQpTAAAAAAAAsAkKUwAAAAAAALAJClMAAAAAAACwCQpTAAAAAAAAsAkKUwAAAAAAALAJClMAAAAAAACwCQpTAAAAAAAAsAkKUwAAAAAAALAJClMAAAAAAACwCQpTAAAAAAAAsAkKUwAAAAAAALAJClMAAAAAAACwCQpTAAAAAAAAsAkKUwAAAAAAALAJClMAAAAAALwGFi1aJIvFot27d9s6FOCVoTAFAC/AMAzNnTtXdevWlaenp5InTy5PT08VL15c/fv31+HDh20d4gsLCgrSN998o/fff18uLi5KlSqV8ufPrzZt2mjTpk1xnnPw4EFZLBbZ29vr/PnzrzZgAABg2rhxoywWi37++efnOn/27NmyWCzav39/0gaGJBEYGKjPP/9cPj4+SpUqlVKnTi0fHx916tRJO3bsiPOcdevWyWKxyMnJSbdv337FEQOxUZgCgOd0//591a9fX+3bt5evr6/++ecf3blzR4cPH5a/v7/WrVun4sWL2zrMF7Jnzx75+Pho0aJFGjRokC5cuKCbN2/qzz//VHR0tCpVqqTLly/HOm/ixIlKkyaNUqRIoSlTptggcgAAgLfbxo0b5ePjo82bN2vUqFG6cuWKrl+/rt9++01BQUEqXbq07t+/H+u8SZMmydPTUw8ePNCMGTNsEDlgzcHWAQDAm6pXr15auXKlNmzYoPLly5vtHh4eatq0qRo1aiR/f38bRvhiQkNDVa9ePXl4eGjbtm1KlSqVeaxEiRKaOXOmKlWqJDs76+847t+/rxkzZqhTp04KDg7WlClTNGjQINnb27/qKQAAALyVAgMD1ahRIxUoUEDr16+Xo6OjeczX11eLFy/W8OHDZbFYrM67deuWFi5cqMGDB2v79u2aOHGievTo8arDB6ywYgoAnsPly5c1adIkNWvWzKoo9Th7e3uNGjXKqs3NzU1du3bV9u3b5evrKycnJw0ePFiSdPHiRbVt21YZMmRQ8uTJ5e3trcGDBysyMtI8f8KECbJYLLEejzt9+rQsFosmTZpkti1btkwWi0VbtmzR8OHDlTlzZjk5OalixYoJ2rdg4sSJunr1qoYNG2ZVlHrcxx9/rEyZMlm1zZ07V6Ghoeratau6deumK1euaNWqVc+8HgAAeDWmTZsmi8WiY8eOaejQocqUKZOcnJxUrVo1nTx50uw3YsQItWrVSpL0/vvvy2KxyGKxaPbs2Wafmzdvqlu3bsqaNauSJ08uLy8v9enTR3fv3o11vaNHj2rw4MHy8vKSnZ2dQkJCzGNHjhzRF198oQwZMihlypSqUaOGjh49Giv2F71eeHi4+vTpo5w5c8rJyUm5cuVS165ddfXqVavrnDx5Us2aNZOHh4ccHR2VP39+jRo1StHR0Ym+j5IUEBBg3r+Yx+jee+89/fLLL8/1z/C///2vQkJCNGrUKKui1OMGDhwY69iff/4pSerUqZM+/fRTHT58ON5H/oBXhcIUADyHtWvXKioqSjVq1Ej0uefOndPw4cM1ceJEnTp1SkWLFtX169dVunRpHThwQMuXL9fNmzf13XffaezYsWrWrNkLxfrDDz/IYrHowIEDOnjwoJycnFSpUiUdO3bsqeetWrVK9vb28vPzS9T1Jk2apOrVq8vb21tFixZVyZIlNXHixBeZAgAAeAlGjBghd3d3HT58WLt27dKlS5fUvHlzGYYhSfL399esWbMkSfv27ZNhGDIMQy1btpT0qEhUqlQpbdmyRTNnztTt27c1d+5cLVu2THXr1rUq4kjSkCFD5OHhoQMHDmjmzJlWq3mGDBkiLy8vHTt2TLt27VJ4eLgqVKhgtWVAUlzv008/1bx58zRz5kwFBwdrw4YNKlasmCZMmGCed+bMGZUqVUqXL1/Whg0bdO3aNfXr109ff/21OnfunOj7KEmlS5c2759hGLp48aI+/vhj9e7d2+qLxYRatWqVXF1dVaZMmUSdN3nyZDVt2lTp0qUz8zXyNNicAQBItCFDhhiSjLVr1ybqPFdXV8PZ2dm4deuWVXufPn0MOzs749ixY1btP/zwgyHJWLNmjWEYhjF+/HhDknHu3DmrfqdOnTIkGRMnTjTbli5dakgyWrRoYdU3JCTEcHV1NZo3b/7UWL29vQ1PT89Eze/EiROGJGPx4sVm27Rp0wwHBwcjMDAwUWMBAIAXt2HDBkOSMW7cOLNt6tSphiSjT58+Vn3/97//GZKMHTt2mG2zZs0yJBn79u2LNXb37t0NR0fHWHnJli1bDEnG0qVLra73ySefxBoj5li3bt2s2q9cuWI4Ojoan376aZJez8vLy/joo49itT+uXbt2hqOjY6zc5csvv7S6F4m5j/Fp2bKlUbx4cfPzwoULDUnGrl27nnqem5ub4ePj88zxH7d161ZDkrF161azbfTo0Yazs7MRFhaWqLGApMSKKQB4DsZj34A97vLly1bLtC0Wi5YtW2bVp1y5ckqbNq1V27p161SgQAHly5fPqr1p06bm8edVv359q8+urq6qXLmy1q9f/9xjxmfixInKmjWr6tSpY7a1aNFCrq6umjZtWpJfDwAAPL/Hf19Lko+PjyTp7NmzCTp/6dKlKlGihLJnz27VXrZsWaVMmTLW23ufzEmedixTpkwqUaKEVb6SFNcrUqSIZs+erbFjx+rMmTNxxrJu3TqVKVNGGTNmtGqPLy9L6H0cP368SpQoodSpU1s9Fnn69Ok440hqEydOVJEiRVS2bFmzrUOHDoqOjrZ6PBN41ShMAcBzyJYtmyTp0qVLVu1eXl7mEu3p06fHeW7mzJljtd26dUuenp6x2mPagoKCnhpPfIUyScqQIUOcbbdu3XrqmNmyZVNQUJDu3bv31H4xHjx4oD///FMXL16Ug4OD1R4Kt27d0qRJk54aJwAAeLWeLLy4uLhIkkJCQhJ0/rVr17R161Y5ODjIwcFB9vb2srOzk52dne7evRsr14grB4oRX77yeA6UFNebMmWKmjdvrqFDh8rb21tZsmRRt27drPaYSmxelpD7OGLECH322Wdq27atTpw4oYcPH8owDHXu3FkPHjyI77bEK1u2bLp06VKCc6uwsDDNnTtXBw4csPoC1d3dXffu3eNxPtgUhSkAeA5Vq1aVvb29/v7770SfmyxZslhtadOm1fXr12O1x7R5eHhIerTaSZLCw8Ot+l25ciXe68U37pOrtp5Us2ZNPXz4UGvWrHlqvxiLFy/W7du3FRISYrWHgmEYCgwM1Llz517KKi0AAPB8nnxjW2J5eHioRo0aevjwoR4+fKioqChFR0ebv/+nTJli1T+uHChGfPmKu7t7kl4vffr0mjRpkoKCgnTw4EH16NFD//vf/6z21ExoXhYjIffxzz//VNWqVdWjRw9lypTJfFvxuXPnnnluXGrWrKnQ0FBt3749Qf1nzpwpOzs7RUZGxsrTDh06pF27dunAgQPPFQvwoihMAcBzyJIlizp27Kh58+Zp27ZtLzxe1apVdeTIEZ06dcqqff78+eZxScqVK5ck6fDhw1b9nnxc8HFLly61+hwWFqaNGzeaY8bno48+UsaMGfXVV1/p33//jbPP77//rsDAQEmPloeXLl3aLJ49LmPGjCpSpMhzbe4JAABsx9nZWZIUERER61i9evW0ZcsWMxd4EU/mK1evXtWuXbus8pWkvJ6dnZ0KFSqk/v37q2vXrjpy5Ihu374t6VHetX379ljFqZi8rEqVKs91zSffkHfx4kVt3rz5ucbq0aOH3NzcNGDAAKs3OD9uxIgR5j+3iRMnqmrVqnEW63x8fJQlSxbyNNgMhSkAeE4//fST/Pz8VKNGDY0cOVKnT59WZGSkgoODtX37ds2bN09Swr5F69+/v9KlS6dmzZppz549CgsL05w5c/T111+rTp06qlatmiSpRIkSKlKkiL799lsdOXJEwcHBmjBhwlMf9YuIiNCoUaN069YtnT59Wi1btlRkZKQGDRr01JhcXV21ZMkS3bhxQ76+vlqyZIlCQkJ0//597dq1S61bt1aXLl0UHR2tCxcuaO3atapZs2a849WqVUsLFy585iOEAADg9VGgQAHZ2dlpxYoVsYpTgwcPVoYMGVSrVi2tWbNGoaGhCgoK0qZNm9SmTZtE7ZF5/fp1TZgwQcHBwTp27JiaN2+ulClTasCAAUl6vbJly2rOnDm6ePGiIiIidODAAS1btkxFihRRmjRpJEnffPONHB0d1aRJEx0+fFghISGaOnWqRo0apXbt2un9999P8Lxi1K9fX6tWrdL8+fP177//ateuXWrZsqWZ4yVWpkyZtGDBAh0+fFiVKlXS33//rbCwMN29e1dbt25VgwYNNHDgQBmGoX379mnv3r1PzdNq1qypGTNmJHgLByApUZgCgOfk5OSk5cuXa+LEidq0aZN8fX3l7OysPHnyqGvXrkqTJo1Wr1791CQgRsaMGRUQEKCCBQuqZs2a8vDw0MCBA9WzZ0/z2znpUZFrzpw5ypw5s4oXL678+fPr2rVr+vLLL+Mdu3fv3rp//74KFSokHx8f3blzRxs2bFCBAgWeGVfx4sV16NAh1atXT9988428vLzk7u6udu3ayWKxaOPGjfLy8tLkyZMVHR39zMJURESE/vzzz2deFwAAvB5y5cqlUaNGadq0aXJ2djY37JYePRa3a9cuVa9eXd26dVP69Onl4+Ojb7/9VnXq1FHFihUTfJ1Bgwbp7Nmzyps3r4oVKyYnJydt3rxZWbNmNfskxfXGjh2rRYsWqXz58nJzc1OjRo1UpUoV/f333+aXid7e3goICJCnp6cqVKig9OnTa+TIkRo8eHCsxwUTavDgwerRo4d69OihdOnSqXfv3ho7dqxy5sz5XONJUuXKlXX48GH5+vqqT58+ypgxozJkyKCPP/5YHh4eCggIUIoUKcz9o56Vp4WEhJhfrAKvksVgJ1oAeCstW7ZM9erV0z///KNy5crZOhwAAIBYpk2bpg4dOujYsWOx3k4M4N3AiikAAAAAAADYBIUpAAAAAAAA2ASFKQAAAAAAANgEe0wBAAAAAADAJlgxBQAAAAAAAJugMAUAAAAAAACbcLB1AEBcoqOjFRgYqNSpU8tisdg6HADAO8AwDIWHhytTpkyys+O7O7z+yJcAAK/Sy8qVKEzhtRQYGKgsWbLYOgwAwDvo0qVL8vLysnUYwDORLwEAbCGpcyUKU3gtpU6dWtKjv/AuLi42jgYA8C4ICwtTlixZzN9BwOuOfAkA8Cq9rFyJwhReSzHL0V1cXEi0AACvFI9E4U1BvgQAsIWkzpXYQAEAAAAAAAA2QWEKAAAAAAAANkFhCgAAAAAAADZBYQoAAAAAAAA2QWEKAAAAAAAANkFhCgAAAAAAADZBYQoAAAAAAAA2QWEKAAAAAAAANkFhCgAAAAAAADZBYQoAAAAAAAA2QWEKAAAAAAAANkFhCgAAAAAAADZBYQoAAAAAAAA2QWEKAAAAAAAANuFg6wCAp9nhnknOFoutwwAAvKbKRobbOgTA5siXAABP87rnS6yYAgAAAAAAgE1QmAIAAAAAAIBNUJgCAAAAAACATVCYAgAAAAAAgE1QmAIAAAAAAIBNUJgCAAAAAACATVCYAgAAAAAAgE1QmAIAAAAAAIBNONg6ALw+oqOjtXLlSgUEBOjevXvy9vZW48aNlT59eqt+ERERWrRokQ4cOCA3NzfVqVNHBQsWjDXe6tWrtW3bNt2/f1/58+dX8+bN5eTk9KqmAwAA8FIEBgZq2LBhKl26tNq0aWPrcAAAeKOxYgqSpKtXr6pUqVL67LPPZBiGPD09tWzZMuXKlUsLFiww+505c0YFChTQokWLlDp1ap05c0bFihXT+PHjrcbr2LGjPvjgA0VGRsrNzU2jR49W0aJFFRYW9qqnBgAAkKSmTJmiKVOmqG/fvnrw4IGtwwEA4I1mMQzDsHUQsL3y5csrNDRUW7duVerUqc32ESNGaPDgwdqzZ48KFiyooKAgPXjwQBkzZrTq8+233yosLEwODg4KCwuTm5ubpk+frtatW0uSbt68qQwZMmjOnDlq2rTpM+MJCwuTq6urVjs4y9liSfoJAwDeCmUjw5NsrJjfPaGhoXJxcUmycfF2MQxD3t7e6tWrlwYPHqyJEyeqcePGsfpt2bJFS5cuVdq0aVW7dm1t3bpVnp6eatiwodnn8uXLmj17tq5fvy5vb2+1bt1aqVKlSnAs5EsAgIRIqnzpZeVKrJiCtm3bpi1btui7776zKkpJUt++fZUhQwb98MMPkiQPDw+ropQkFSxYUPfu3VNISIgkKUWKFHJzc1NoaKjZJzQ0VIZhxDoXAADgTbJ+/XrduHFDHTp0ULt27TRp0qRYff78809VrlxZ4eHhunfvnho0aKDBgwdry5YtZp9NmzapSJEiOn78uNKlS6dly5bJx8dHN27ceJXTAQDA5thjCtq2bZskqUKFCrGOJUuWTGXKlDH7xGXKlCkqWLCgPDw8JEnJkyfXokWL1KdPH61fv16pUqXSrl279Msvv8jX1zfOMSIiIhQREWF+5pE/AADwOpo8ebJatWqlVKlS6eOPP1ahQoV0+fJleXl5SZIiIyPl7++voUOHasCAAZKk5s2by8fHxxwjKipKH374oUaMGKGPPvpIkvT555+rdu3aGjZsmH788cc4r02+BAB4G7FiCgoNDZWjo6NcXV3jPO7p6Wm1+ulxY8aM0YoVK/T7779btQcEBOjq1avy9vZW7ty5ZW9vrw0bNuj+/ftxjjN8+HC5urqaP1myZHmxSQEAACSx4OBgLVy40Cwm5c+fX2XKlNHUqVPNPidOnNDVq1fVsmVLs61AgQIqUqSI+fnAgQO6cOGCAgIC1KtXL/Xs2VM9evTQrVu3tHfv3nivT74EAHgbsWIKSpcunSIiIhQUFGSuenrcpUuX4nwE7/fff9cXX3yhv/76S2XLljXbjxw5ogEDBmj16tXy8/OTJH322WfKkSOHxo8fr969e8caa+DAgerTp4/5OSwsjGQLAAC8VqZPny7DMPTHH3/ojz/+kCQ9fPhQU6ZM0VdffSWLxWI+ivdkTvX455s3b0qS8uXLZ/XG4jx58ihdunTxXp98CQDwNqIwBVWuXFmStGzZMrVv397qWHh4uDZv3qxOnTpZtU+cOFHdu3fXrFmzrDbxlKSTJ09KkooVK2a2ubq6KleuXDpx4kScMTg6OsrR0fEFZwIAAPDyxDzGly9fPrMtX758+uabb7Ru3TpVq1ZNmTNnliRduXJFefLkMftdvnzZXDUV84VfuXLlVKZMmQRfn3wJAPA24lE+qFChQmrWrJm+/vprXbhwwWyPjo5W//79JUndu3c326dMmaLPPvtMs2bNivMtND4+PrJYLFq1apXZdvHiRR09elSFChV6iTMBAAB4OXbt2qVDhw7pu+++02effWb1U7duXXMT9Dx58ihv3rwaP368ee6mTZt0/Phx83OhQoVUqFAhDRo0yGrPqKCgIG3fvv3VTQoAgNcAK6YgSZo6dao6d+4sHx8f1axZU66urtqyZYvs7Oy0cuVK89u/PXv26KOPPlKBAgW0fv16rV+/3hxj4MCBypw5s3Lnzq3BgwerY8eOmjNnjlxcXLRixQqVLVvW3JMBAADgTTJ58mSVKFFCmTJlinWsYcOG+uCDD3T79m2lTZtW48aNU/369XXw4EGlS5dOO3bsUI4cOWRn9+g7YYvFojlz5qhevXoqUKCAypUrp6CgIJ04cUL//e9/X/XUAACwKQpTkCQ5Oztr1qxZOnXqlHbs2KHr169rxowZGjRokIoXL272S5cunX766ac4x3h8afk333yjNm3aaNeuXbp//766deumUqVKvfR5AAAAvAxly5ZVmzZt4jxWs2ZNjR49WiEhIUqbNq38/Px08uRJrV27VmnSpNHPP/+s6tWry93d3TwnX758Onr0qDZu3Kjz588rS5YsKleunFKlSvWqpgQAwGvBYhiGYesg8Hr69ddf1aNHD61du1aVKlV6pdcOCwuTq6urVjs4y9lieaXXBgC8OcpGhifZWDG/e0JDQ+Xi4pJk4+Ldc+LECWXKlEmpU6eWJG3fvl2+vr7avn17kn5RR74EAEiIpMqXXlauxIopxOvTTz9VqlSpzDfHAAAA4NnCw8NVrFgxFS9eXPfv39fKlSvVr18/Vo8DABAHClN4qnbt2tk6BAAAgDdK8eLFtW3bNm3atEmRkZEaNmyY1Zv8AADA/6EwBQAAACQxDw8PNWnSxNZhAADw2rOzdQAAAAAAAAB4N1GYAgAAAAAAgE1QmAIAAAAAAIBNUJgCAAAAAACATbD5OV5rpW4FysXFxdZhAAAAvLbIlwAAbzJWTAEAAAAAAMAmKEwBAAAAAADAJihMAQAAAAAAwCYoTAEAAAAAAMAmKEwBAAAAAADAJihMAQAAAAAAwCYcbB0A8DQ73DPJ2WKxdRgA3nJlI8NtHQIAPDfyJQDvGnK3twsrpgAAAAAAAGATFKYAAAAAAABgExSmAAAAAAAAYBMUpgAAAAAAAGATFKYAAAAAAABgExSmAAAAAAAAYBMUpgAAAAAAAGATFKZeQLVq1bRv3z5bh5FgLzPeevXqaceOHS9lbAAA8HL07dtX06ZNs3UYAADgHfbaF6YGDRqkQYMG2TSG6OhoVatWTQcOHLBqX7dunW7dupXk12vYsKEWLFjw3OcnJN74+jyvDRs26ObNm0kyFgAA74pBgwapWrVqsX569+79Sq6/Z88enT59+pVc61Xp3bu3/vzzz1d+/oteFwCAd5WDrQN4lqQqnLyI6OhorVu3TsHBwVbta9asUdGiRZP8ehs3blSlSpWe+/yExBtfHwAA8OocOHBA9+7d05AhQ6za3dzcbBPQW2DXrl1ydXV95ee/6HUBAHhXvfaFqWeJjIzUhAkTtH79etnZ2alGjRrq3Lmz7O3tzT6GYeivv/7S0qVLdf/+fdWqVUudOnWSxWKRJDVv3ly3b9+WnZ2dsmTJoiZNmqh27drm+U2aNJH0aLl7mjRplDVrVk2ZMkUjRozQqFGjlDZt2gTFcv/+fdWtW1ffffedVq5cqYMHDyp9+vTq16+fvL29EzznpIg3rj7Dhw9X69at9eeffypTpkzmeLVr19agQYNUqlQpSdLdu3c1atQo7d69W9myZVOXLl1ixWgYhmbNmqUlS5YoMjJSRYsWVa9evZQqVaoEzxMAgHdBunTpVK1atXiPd+vWTSVLllRISIi2b98uSerSpYvKli2rn3/+WZs2bZKHh4f69eunAgUKWJ1XtmxZBQcHa8uWLbKzs1PXrl1VoUKFp8azcOFCzZkzR+Hh4SpWrJh69eqlNGnSSJKmT5+uzZs3a+LEiVbnfPvtt7JYLPr666+fO15J2rZtm6ZMmaLr16/L29tbPXr0UI4cOazm5Ovrq/DwcG3cuFH29vb68MMP5efnJ+nRCrTDhw8rMDBQW7ZskSTNnTvXjD/GuXPnNG7cOJ0+fVrZsmXTxx9/rEKFCsV7/q+//qoNGzbIYrHIw8ND5cuX18cffywHB4enXrd79+5q3bq1atWqZV574MCBypEjhz7++OOnxgIAwLvitX+U71latWqlH3/8UY0aNVKdOnU0ZMiQWIWSjz76SJ999pmKFy+uDz74QDt37tTo0aPN4927d5e/v7/69u2rggULqk2bNlb7LfTs2VOS9MEHH8jf318fffSRpNiP8j0rlocPH2rdunWqX7++kiVLpg8//FBBQUGqUKGC/v333wTPOSnijavPvXv3tG7dOt29e9fqeuvXr7d6TK958+aaPXu2WrZsqYIFC6pWrVq6d++e1TmffPKJhg0bpho1aqht27batWuXypQpo8jIyATPEwAASDt27NAnn3yiEydOqGXLlnJ1dVX16tVVpUoVXbhwQZ06dVJ0dLQqVqyo8PBwq/O6dOmijRs3qlmzZsqUKZOqVatmFk7i8vPPP6tNmzYqWrSo2rZtq7Vr16pcuXLm7+/ixYtr0qRJOn78uHnOnTt39P333ytv3rwvFO///vc/NWzYUPny5dNHH30kBwcHFSlSRKdOnYo1pz179qhFixbKkSOHatasqT179kh69OVclixZVK5cOfn7+8vf31/Ozs5Wc7xz5458fX0VEhKijz76SAULFlTnzp11+fLleM9v0KCB/P39NWDAANWsWVO//vqrOnfubI4Z33lbtmzRlStXrK6/b98+nTx58pmxAADwrnijV0xt375dCxcu1MGDB+Xj4yNJyp07typVqqTevXurYMGC2r59uyZPnqxt27apTJkykh4lD48nQuXLlzf/XKNGDTk7O2vs2LFq3769JJnfLBYrVizeR+wSEkuMPn36aMCAAZIkPz8/ubm5aevWrapevXqC5p0U8cbV5/z588+89rZt27Ry5UqdOHHCXOXl4eGh5s2bm312796tqVOn6uLFi8qQIYOkR5uje3t7a86cOWrTpk2scSMiIhQREWF+DgsLe2YsAAC8DbZt2xZrxVTz5s3NFTWSVLlyZf3666+SHv1OnTt3rjw9PfXf//5XklSzZk2lSZNG//zzj9Uq6mzZsmnu3LmyWCxq0qSJbt68qW+++Ubr16+PFcfdu3f19ddfa9SoUfr0008lSbVq1VK2bNk0efJkffLJJ8qfP798fX01ZcoUff/995Kkv/76S46OjmrQoMFzxxsREaEePXpo8uTJatiwoSSpfv36CgwM1MiRIzVp0iRz7HLlyun333+X9GhfzvXr12v+/PkqVqyYChcuLFdXV+XMmTPeVWjHjx/XtWvXNGHCBCVPnlySzGKZl5dXnOf7+PiY+Z0klS5dWvnz59eYMWPk7u6eoOsmNpa4kC8BAN5Gb3RhKiAgQNmzZ7dKFCpUqCAXFxft2rVLBQsW1Pr165UlSxazKBUjderU5p/Pnj2r33//XSdPnlRYWJhu376d6I1AExJLjMdjSZkypdKlS6dr164l+FpJEe/zCggIkLe3t9Wjh3Xq1LHqs3btWjk4OOjDDz+UYRiSHj3aFx4eriNHjsQ57vDhw2PtrwEAwLsgd+7c8vf3t2rLli2b1eeYx+klyd7eXhkzZrRqc3R0jDOfqFmzprl1gfTod3bHjh3jjOPEiRMKCQlR3bp1zTYXFxdVrFjRXAUlSZ07d9bAgQM1bNgwOTg4aMqUKWrdurUcHR2fO959+/bp9u3b+umnnzRhwgQZhiHDMHT27FmlS5fOKs4nc7rs2bMnKo/KkyeP3N3d1bZtW3Xu3Flly5aNtarqSaGhoZo4caL27NmjoKAgRUdHyzAMnTlzRu7u7gm+9ovGQr4EAHgbvdGFqZCQELm4uMRqd3V1VUhIiCTp33//feoGopcuXVLx4sVVv359tWzZUm5ubtqzZ4+++OKLJI8lRsw3YjEsFku834y9rHifV2hoqFVRT5KcnJzMPRakR9/epUuXTv369Yt1ftasWeMcd+DAgerTp4/VGFmyZEmiqAEAeH09a48pKe7cISH5xJO/s1OnTq27d+8qMjIy1vkx+cqT+cyTuUzz5s3Vq1cvrVixQnnz5tW2bdv0yy+/vFC8MSt/Pvvss1gbiD85hxfJo6RH89uzZ49+/fVXDRw4UEePHlXTpk01YcIEpUyZMlb/6Oho+fn5ydHRUZ06dVKGDBlkGIbWr18fa/uDxEpsLORLAIC30RtdmMqZM6fOnz+vBw8eKFmyZJKk8PBwBQYGmhtl5syZU+fOndO9e/fk5OQUa4xVq1bJw8PDao+mmOf+Yzz+TeOLxJIUkireuPrE3J/79++bbWFhYVZLxrNnz65z584pKirK3NT93LlzevjwodknR44cunbtmkqWLBlnsS4ujo6OVt+0AgCAF/f4/kwxnzNlyhSruCM9ymWkR3lFyZIlzfYTJ07EWu3dqlUrTZkyRXnz5lXRokX13nvvvVCcMblSqlSpEvUoXFzs7J69hWrWrFk1YsQISdKFCxdUsmRJTZw4UT179ox1/pkzZ7Rr1y5duXLFfDnM4cOHE3RdJycnq7xKkm7cuJHgWJ5EvgQAeBu90Zuf16tXT3Z2dlYbmQ8ZMkTp06c3387SuHFjJUuWTAMHDjS/TQsMDNTKlSslSc7OzgoKCjI3BQ8MDNTYsWOtrmNvb6+0adM+dZl4QmJJCkkVb1x9MmTIIA8PDy1fvtxsGzlypNV5DRo00MOHDzVu3DhJjx7RGzx4sFWfpk2bKlWqVOrevbvVZueLFy+OM5EDAAAvx8KFC3Xo0CFJUlBQkMaNG6e2bdvG2TdbtmyqVKmS/vOf/+jBgweSpGXLlmnXrl1q166dVd/OnTtr+fLlmjp1aryPBiZG7ty5VaVKFfn7+1vlJgcPHtSSJUsSNdaztkjYt2+fli1bZn5Onz69UqZMaW4/8OT5MY/WnThxQtKjfZ6+/PLLBF23YMGCWrlypZmD/v3339q/f3+CYwEA4F3wRhSmYjYFffznu+++U9q0afXHH39o1KhRKlCggPLkyaPp06drxowZ5vLntGnTaunSpVq4cKGyZMmiEiVKqHz58uYy8SZNmqhw4cLKmzevypQpE2tzyxgff/yxunTpokqVKsWZgCUklsQYP358rDkvW7YsyeKNr89PP/2kwYMHq3jx4vL29taFCxfMFWAx8/ztt9/0xRdfqFChQsqZM6fu3r2rFClSmH3SpEmjlStXKiAgQJkzZ1aZMmXk6empP/74Q+nTp0/0vQAA4G0WV57TqlWrJBm7YsWKqlmzpkqWLKncuXPL09NTAwcOjLf/xIkTdebMGWXLlk1FixZVixYtNGrUKL3//vtW/YoVK6ZChQopPDxcH3zwQZLEOmvWLGXIkEG5cuVSqVKllCtXLnXo0EEZM2ZM1Dgffvihpk+fLl9fX1WrVk3BwcFWxzNkyKDff/9dGTNmVLly5ZQtWzZ5e3urU6dOcZ7v5OSkAQMGqHbt2ipbtqyyZctmtYXB0647ePBg7du3z1xZ9s0331jlbc+KBQCAd4HFeM2/kjl48GCsJc+S5Onpaf5iv3fvng4cOCA7Ozu99957cS5Pj4qK0sGDBxUdHa1ChQpZ9TEMQ0eOHFFISIjy588vi8WivXv3xlpKfu7cOV28eFHJkydXmTJltHbtWhUtWlRp06Y1+zwtlqioKG3YsCHWI25btmxRrly5zMRr06ZN5jeVj8ufP78yZ86cpPE+2Ud69I3qqVOnlD17dmXMmFEbNmxQoUKF5OHhYZ53+/ZtHT16VNmyZVOWLFm0ceNG+fj4WPUxDEPHjh1TcHCw8ubNa3XsWcLCwuTq6qrVDs5yTsCjiQDwIspGhj+7E956Mb97QkNDE/wo+ouKL89JkSKFypUrJ0nauXOnMmTIYLUhesyXP4/vL7R161blyJHDfNysePHiatq0qXr16qVjx47Jzs5OhQsXtnqcf8+ePXJzc1OuXLnMtujoaB06dEjh4eEqWLCg0qRJE2fsDRo0UMqUKTVr1iyr9ueNN8alS5d0/vx5ZcmSRdmzZ3/m2EeOHJHFYlGBAgXMtps3b+rkyZO6d++eKlSoEGdueO3aNZ05c0aZM2eOdZ24zr9w4YIuXLigrFmzKlu2bFq3bp2KFStmdX/iOu/u3bs6cuSIeZ8PHjwoZ2dn5c6dO0GxPA35EoB3FbmbbbysXOm1L0zh3USiBeBVIrmBZJvC1MsUU5h68o1/SeHs2bPKmzevNm7cKF9f3yQfHwlDvgTgXUXuZhsvK1d6ozc/BwAAwKtVt25d7dy5U02bNqUoBQAAXhiFKQAAgLfQr7/+qgwZMiT5uP369ZOLi4uKFi2a5GMDAIB3D4UpAACAt1DJkiVfyriVKlV6KeMCAIB30xvxVj4AAAAAAAC8fShMAQAAAAAAwCYoTAEAAAAAAMAmKEwBAAAAAADAJtj8HK+1UrcC5eLiYuswAAAAXlvkSwCANxkrpgAAAAAAAGATFKYAAAAAAABgExSmAAAAAAAAYBMUpgAAAAAAAGATFKYAAAAAAABgExSmAAAAAAAAYBMOtg4AeJod7pnkbLHYOgzgnVQ2MtzWIQAAEoB8CcDrinwSCcGKKQAAAAAAANgEhSkAAAAAAADYBIUpAAAAAAAA2ASFKQAAAAAAANgEhSkAAAAAAADYBIUpAAAAAAAA2ASFKQAAAAAAANiEg60DwOvn3Llz2rFjh3x9fZUlSxZbhwMAAPDauH//vnbu3KmgoCB5enqqRIkSSpYsWax+ISEh2rNnjySpSJEi8vDwiNUnIiJCu3bt0q1bt+Tl5aWiRYvKYrG89DkAAPA6oTCFWL7++mv973//U9euXTV+/HhbhwMAAPBamDhxogYMGKBs2bIpR44cOn78uEJCQjRhwgTVr1/f7Ne7d2/NmTNH+fPn14MHD7Rnzx6NHDlS3bp1M/vs3LlTDRo0kLu7u7y9vbVr1y6lS5dOq1evVvr06W0xPQAAbIJH+WAlJCRE8+fPV//+/TVr1izdvXs3zn7nzp3T4sWLtX//fkVHR2v+/PkKDAy06nP37l2tXbtWK1as0OXLl19F+AAAAC/FjBkz1KVLF/3yyy/at2+fFixYoKNHj2rAgAFq3LixNm3aZPbNmzevzpw5o7Vr12rTpk0aP368evTooePHj5t9+vfvr6JFi+rw4cNatGiRTpw4oaCgII0ZM8YW0wMAwGYoTMHKjBkzlD17dg0fPlypU6fW3LlzY/X55ZdflC9fPv3444/q3Lmz/Pz81Lp1a+3du9fss3btWmXPnl2DBg3SL7/8okKFCmno0KGvcioAAABJwjAMDRw4UE2bNlWrVq2sjvXs2VNly5bVwIEDzbauXbsqRYoU5ucmTZooOjpaBw8eNNsiIiKUPXt283OqVKmUPn16RUZGvryJAADwGqIwBSuTJ0/WRx99JHt7e3Xs2FGTJ0+2On716lX169dPv/32mzZs2KDdu3erUKFCioiIMPsEBQWpSZMm+v3337V161YtX75cO3bs0IgRI7R9+/Y4rxsREaGwsDCrHwAAgNfB0aNHdfnyZTVp0iTO402aNFFAQIBCQ0PjPL527VpJUsGCBc22UaNGacWKFfr66681bdo0dejQQRaLRf379483DvIlAMDbiMIUTHv27NHRo0f14YcfSpI6d+6sbdu26eTJk2afZcuWycXFxewjKVYCtWDBAtnb2+vhw4eaO3eu5s6dq/3798vLy0sbN26M89rDhw+Xq6ur+cOm6wAA4HVx7do1SVLWrFnjPJ41a1YZhqEbN27EOhYYGKhPPvlE7dq1sypMZc2aVQUKFNC8efO0YMECrVu3TqVKlZKbm1u8cZAvAQDeRmx+DtPkyZOVL18+rVmzxmzLlSuXJk+erJEjR0qSLl26pKxZs1q9MSZz5sxycPi/v0rnz5+XxWLRvHnzrMZ/7733lDlz5jivPXDgQPXp08f8HBYWRrIFAABeC6lSpZIk3bp1K87jQUFBkqS0adNatd+8eVN+fn7Knz+/fvvtN6tj9evXV4ECBXT06FFZLBb9+++/KlGihPr27Rvvy2fIlwAAbyMKU5Ak3bt3TzNnzlT58uW1aNEisz1Xrlz6448/9N1338nBwUFp06ZVSEiI1bl37tzRw4cPzc8uLi6yt7fX7NmzE3x9R0dHOTo6vug0AAAAklzhwoWVIkUKbd++XXXr1o11fPv27cqePbvc3d3NtqCgIFWpUkUZMmTQkiVLrPacCg4O1sGDB/XNN9+YX/Y5Ozurdu3aWrZsWbxxkC8BAN5GPMoHSdLcuXPl6OioxYsXa/bs2ebP4sWLFRERoaVLl0qSfH19debMGR05csQ8d/HixVZj1ahRQzdv3tSCBQus2u/fv6/bt2+//MkAAAAkIScnJ3Xp0kU///yzLl68aHXs4MGDmj59utVKppiiVLp06bRs2TKlTJnS6hw3NzelSpVKx44ds2o/duyYvLy8Xt5EAAB4DbFiCpIePcZXr1492dlZ1yqTJUum2rVra/LkyWrUqJFKlCihxo0bq06dOurTp4/Cw8P122+/yd7e3vzG7/3339cXX3yhDz74QN26dVOBAgV09uxZzZs3T7Nnz461zB0AAOB1N2LECJ0+fVolSpRQt27dlCNHDh07dkzjx49X586d1b17d0lSdHS0/Pz8dPnyZfXq1UtLliwxxyhatKjy5Mkji8Wir7/+WoMGDVJ4eLjy5cunjRs3avXq1fr7779tNUUAAGyCwhQUHh6uTJkyqUOHDnEe/+ijjzRp0iRFRETI0dFRM2fO1IQJE7R3715ly5ZNf//9twoWLKjUqVOb53z33XeqUaOGFi1apC1btihv3rxav359vHtMAQAAvM5SpEihZcuWadWqVVq1apVOnDihpUuXqk6dOlZ7QkVFRSlv3rzKmzevVq9ebTWGq6ur8uTJI0n6/PPPVapUKS1fvlybN29WtmzZdOTIEfM4AADvCothGIatg8Cb5fbt21arnhYtWqTmzZvr+vXrSpMmTZJcIywsTK6urlrt4CznxzZaB/DqlI0Mt3UIwCsV87snNDRULi4utg4Hb4A5c+aoZcuWWrhwoRo0aPDKr0++BOB1Rz75dnlZuRIrppBoY8aM0fXr11W6dGlduHBBP/30k/r165dkRSkAAIA3QfPmzXX27FmtXLlStWrVUvLkyW0dEgAAbxwKU0i0b7/9Vv/73/+0bds2pUqVSnPnzlWNGjVsHRYAAMAr5+/vb+sQAAB4o1GYQqLZ29urXbt2ateuna1DAQAAAAAAbzC7Z3cBAAAAAAAAkh6FKQAAAAAAANgEhSkAAAAAAADYBIUpAAAAAAAA2ASbn+O1VupWoFxcXGwdBgAAwGuLfAkA8CZjxRQAAAAAAABsgsIUAAAAAAAAbILCFAAAAAAAAGyCwhQAAAAAAABsgsIUAAAAAAAAbILCFAAAAAAAAGzCwdYBAE+zwz2TnC0WW4cBvNHKRobbOgQAwEtEvgTAFsgxkVRYMQUAAAAAAACboDAFAAAAAAAAm6AwBQAAAAAAAJugMAUAAAAAAACboDAFAAAAAAAAm6AwBQAAAAAAAJugMAUAAAAAAACboDD1/02bNk3r16+3dRgv1eDBg3X27NmXMvawYcN08uTJlzI2AAB4M71p+dXLjHfmzJlavXr1SxkbAIA32SstTAUEBGjEiBGv8pJxmjJlijZu3GjV9rISkdmzZ2v27NlJPu7zGDJkiFmYio6O1uDBg3X+/PkkGZvCFAAAtkF+9XwSEm9cfZ4XhSkAAOJGYer/a9++vapUqZLk13qdClODBg1Szpw5JT0qTA0ZMiTJClMAAMA2yK+eT0LiTcrCFAAAiJuDrQN40qVLl7RkyRKFh4erWLFi8vPzi9UnNDRUS5YsUWBgoIoVK6Zq1aqZx+bNm6fDhw/LYrHIw8NDvr6+eu+998zjs2fP1sWLF7Vx40YNHjxYktSvX7/niuX3339XgQIF5OLion/++Uf29vaqX7++MmXKlKg579mzRxs2bJCdnZ38/PxUqFAh81hkZKSGDRumzp076/jx4zp48KDSp0+vZs2aydHR0WqcNWvWaO/evcqWLZvq1KmjcePGqWnTpsqTJ0+sa37//feSHn0zuHHjRnl4eKhNmzb68ccf1aNHD6VNm9bsO3ToUDVv3txqnHXr1mn37t3mteJy9uxZrVixQpGRkSpatKgqVaqUqPsCAACSxruWXyVFvPH1mTRpkipUqKCiRYuafWfMmKH06dOrevXqZtuxY8e0fPlypUmTxupePi4kJESLFi3S9evX5e3trXr16il58uQJnicAAG+D12qPqbVr1ypPnjxavXq1rly5ojZt2qhZs2ZWfXbu3Clvb2+NHz9eN27c0MiRI9WtW7dYYxmGof3796tixYr64YcfnnntJ5duJySW33//XV26dFH79u117tw5zZ07V4UKFdKVK1cSPOehQ4eqfPnyOnHihA4dOqTixYtr3Lhx5vHIyEgNGTJEtWvX1pgxY3T16lUNHTpUlStXlmEYZr++ffuqcePGOn/+vJYuXaoSJUpo8ODBVo/XPf4oX1xCQkI0ZMgQ3b59O1aMj48zcOBANWzY0OpakZGRVudMnjxZRYsW1b59+3TlyhV16NBBbdu2TfB9AQAASeNdzK+SKt64/Pjjj9q7d69V24wZM6we01u+fLmKFCmiHTt2aO/evSpfvrwOHjxodc6ePXuUN29eLVq0SLdu3dKYMWNUrFgxhYSEJHyCAAC8BV6bFVMPHz7UJ598ok8++URjx46VJHXv3l0FCxbU4sWL1aBBA0VHR+uDDz5QnTp1NHXqVFksFkmy+kXftGlTNW3a1PzcqlUr1a1bVx9//LGcnZ3VsmVL/fzzz6pUqZL57dfzxBIjWbJk2rlzpxwcHGQYhgoUKKBp06bpyy+/fOacT58+rSFDhmj+/PmqX7++JKl8+fLq3r27mjZtqowZM5p9y5cvr19++UWS1Lt3b2XJkkXbtm2Tr6+vTpw4oR9//FF///23+Y3c6NGj1b9//3iv/fnnn+vLL79U+/btzZVMCXms79SpUxo1apRWrVplXmvkyJHy9/c3+5w/f16ffvqptm3bpmLFikmSvvjiC+XJk0erVq1SzZo1Y40bERGhiIgI83NYWNgzYwEAAE/3LuZXSRVvQvrEN8+ePXvq888/19ChQyVJbdu2VZkyZcw+0dHRat26tfz9/dW7d29Jj4poVapU0YgRI+J9NJN8CQDwNnptClOnTp3S6dOn9fHHH5ttefLkUeXKlbVy5Uo1aNBABw4c0JkzZ7Ro0SIzaZKkwoULW40VEBCgPXv2KCgoSA8ePNC9e/d0+vRpFSlSJMliiVG3bl05ODy6jRaLRUWKFEnwvk1r1qyRh4eHWZSSpA8//FCfffaZNm3apJYtW5rtjRo1Mv+cKVMmpUuXTufPn5evr6/Wrl2rzJkzWy0T79Chw1MLU89rzZo1ypgxo9W1OnXqZFWYWrJkiZycnLRy5UqtWLFC0qNky8XFRTt27IizMDV8+HANGTIkyeMFAOBd9i7mV0kV7/M6efKkzpw5ow4dOphtpUuXVsGCBc3Phw8f1okTJ3T58mUNHTpUhmGYK+F37NgR79jkSwCAt9Fr8yhfYGCgJFmtEpIeFWFilm7fuHFDkuTl5RXvOJ06dVKDBg20b98+RUZGys7u0RSDg4OTNJYYqVKlsvrs4OCgBw8eJPg6np6eVm329vZKnz59oq5z9epVZciQweq4u7u7kiVLlqA4EuPq1auxYvbw8DCTx5g+Tk5OevjwoaKiohQVFaXo6Gh9+OGHKlWqVJzjDhw4UKGhoebPpUuXkjx2AADeNe9ifpVU8T6vq1evSlKsfOnxecf0SZ48uZkvRUdHq2LFimrRokW8Y5MvAQDeRq/NiqmYDS2vXr0qV1dXs/3q1avKkSOHJCl9+vSSpMuXL8vNzS3WGFeuXNGUKVN08OBBcwPxixcv6j//+Y9Vv8e/DXzeWJJCpkyZdO3aNau2qKgo3bhxQ5kzZ07wOBkzZtT169et2m7duvXUBC6uexBTyHr48KHZFhERYTVOxowZY8UcFBRkdU769Ol19+5dff3117K3t0/QHBwdHWNt5g4AAF7Mu5hfJVW88fVJliyZVd4jSeHh4eafYwpQ165dU65cucz2q1evmqu1Yu55kyZNVLx48YRMSxL5EgDg7fTarJjKnTu3vL29NWnSJLPt9OnT2rBhg2rVqiVJKlKkiHLmzBlr88qYPRDu3r0rSUqRIoV5bPz48bGu5erqapVAPE8sSaFatWoKCgrSsmXLzLYZM2bIYrGoQoUKiRrnypUrVpt1/vHHH089x97eXs7Ozlb3wdPTU46Ojtq/f7/ZtnjxYkVHR5uf/fz8dPXqVa1du9Zsmzx5stXY9evX1927d602cZcevYXn8uXLCZ4XAAB4Me9ifpVU8cbXJ3v27Fa50oULF7Rv3z7zc548eZQjRw5NnTrVbAsICNCRI0fMz0WKFJG3t7eGDx9ulWfdvXtXBw4ceMYMAQB4u7zyFVP379+PtYFkqlSp1K9fP/3666+qX7++zpw5oyxZsmjOnDlq0KCBueeAnZ2d/ve//6lOnTo6ceKESpUqpUOHDil37tz65Zdf5O3trcqVK6t27dpq1KiRjh07pmPHjsWKoUqVKho+fLgcHBzk5OQU6/XADg4Oz4wlMY4fPx5rzj4+PmratKkGDRqkFi1aqHXr1nr48KH+97//afTo0Yl6JXLevHnVs2dPNWzYUG3atFF4eLh27dqlZMmSmUvX41KlShV9/fXX2rFjhzw9PfXZZ5+pV69e+uSTTxQQEKCQkBDt37/fatVT7ty51bdvXzVq1Eht2rRRWFiYdu/ebfXYYK5cufTbb7/pk08+0Zo1a1SgQAGdPXtWR44c0ZIlSxJ+4wAAQIKQXz3i4+OjJk2aJEm88fXp2bOnmjRponv37snV1VWrV6+Wu7u71Tx//PFHNW3aVCdPnlT69Om1ZMkSZcmSxexjZ2en2bNnq06dOipevLgqV66soKAgbdu2TcOHD3/p+2ABAPA6eaWFqdKlS1ttkv0kPz8/nTx5UkuWLFF4eLj+/PNPVa9ePdYYp0+f1qJFi3Tz5k3VqVNHVapUkfRoufXff/+tefPm6dy5c2rRooUaNGigMWPGKHv27OYYvXv3Vr58+XTo0CHdu3dPktS+fXtlzZo1UbF8/PHHKlCggFVb48aNrYpBLVu21PHjx+Od81dffaVatWppw4YNsrOz0+7du81l59KjvQcGDRoUa9+HPn366L333jM/jx07VjVr1tTevXuVLVs2ff/998qUKZPSpElj9hk0aJBy5sxpfv7rr780b948Xbx40WwbMWKEKlWqpEOHDsnX11fjx4/XmDFjlCdPHrPPyJEj5efnp927dytbtmyaMGGCxo0bZ9WnY8eO8vPz08qVKxUcHCxfX1/VqlWL5ecAACQx8itrSRlvXH3q1q2rnTt3av369XJzc5O/v7/Wr19vPp4nPVo9vn//fi1fvlxp0qRR//79tXXrVnl4eJh9ihUrptOnT2vp0qU6f/683n//fX3//fex9g0FAOBtZzFiXgGCN9rJkyetCkMTJkxQ//79de3aNTk7O9swsucTFhb26FtIB2c5J2APCADxKxv59EdVADwS87snNDRULi4utg4HeCbyJQC2RI757nlZudJrs/k5XsySJUu0aNEilS5dWhcuXNCyZcs0bty4N7IoBQAAAAAA3g0Upt4S/fr1U4UKFbRt2zbly5dPw4YNU+7cuW0dFgAAAAAAQLwoTL1FSpYsqZIlS9o6DAAAAAAAgASJ/5VtAAAAAAAAwEtEYQoAAAAAAAA2QWEKAAAAAAAANkFhCgAAAAAAADbB5ud4rZW6FSgXFxdbhwEAAPDaIl8CALzJWDEFAAAAAAAAm6AwBQAAAAAAAJugMAUAAAAAAACboDAFAAAAAAAAm6AwBQAAAAAAAJugMAUAAAAAAACbcLB1AMDT7HDPJGeLxdZhAK+dspHhtg4BAPCaIF8C8KLILWFLrJgCAAAAAACATVCYAgAAAAAAgE1QmAIAAAAAAIBNUJgCAAAAAACATVCYAgAAAAAAgE1QmAIAAAAAAIBNUJgCAAAAAACATbzVhSkvLy9t2bLF1mEAAAC88caOHavGjRvbOoyXqlmzZho1atRLGbt9+/YaOnToSxkbAIA32UstTBUsWFBeXl6xfl7WL/wnXblyRffv338l13oVoqKi5OXlpe3bt7/S81/0ugAAIGGOHj0qLy8vHT161KZxjBo1Ss2aNbNqCwsL040bN5L8WpMnT1bBggWTfNzncfPmTYWGhpqfGzdurLFjxybJ2EFBQQoJCUmSsQAAeJs4vMzBr1y5ot69e6tTp05W7S4uLi/zsm8twzB05coVRUREvNLzX/S6AAAgYSIjI3XlyhVFRkbaNI7Q0FDdvHnTqq1Pnz7q1q1bkl8rPDxcV65cSfJxn8fcuXPl6Ohofr5x44bCwsJsGBEAAG+/l/4on6ura6wVUzGFqbt378rLy0tTpkxRs2bNlD9/flWsWFFbtmzRvn37VLduXeXOnVt169bV6dOnzTFjzps9e7aaNm2qfPnyqXLlytq2bdtTYwkKCtKnn36qfPnyqUCBAurdu7dVstG2bVt9/vnnVuc8ePBAhQoV0qxZs547Xkm6f/++vvnmG7333nsqUKCA2rRpo/Pnz8ea09y5c9WyZUsVKFBAlSpV0saNG80+hQsXlvRombmXl5cqVaoU5zynTJkiX19f5c2bV02bNtWRI0eeen7RokXl5eWlrFmzytfXV2PHjlVUVNRTr3v58mV5eXlZzUGSvL29tW7dumfGAgAAns+CBQtUoUIF5cyZU1WqVNHff/8dq8+2bdvUoEED5c6dW9WqVdOGDRvMYzNmzDBzsvz586t58+Y6duyYeXzy5Mn673//q+3bt5v9lixZokmTJqlLly6JimXo0KFq3bq1xo0bp/Lly6tgwYLq1auX7t69m+D5RkVFacSIEXr//ffl7e2tFi1a6OTJk1Z96tSpo5EjR6p37956//33Vbx4cf3yyy9WfSIiItSvXz/lzZtXZcqU0ffff6927dpZPV736aef6rfffpMkde7cWbt27dLYsWPN+3Dz5k2VKVNGs2bNshq7bdu2VuM8ePBA/v7+ypcvn3mt6OjoWHObNm2aypUrp9y5c6tWrVratGlTgu8LAABvC5vuMRUdHa0rV67oq6++UuvWrbVgwQJlzpxZ9erVU9u2bdW1a1ctWbJEyZIlU6tWrWKd9/HHH6tevXqaP3++3n//ffn5+eny5ctxXsswDNWrV08HDhzQH3/8oYkTJ2rDhg1q3ry52ad27dqaOHGi1eN/y5Yt0+nTp1WrVq3njtcwDDVo0EA7d+7U+PHjNX/+fGXIkEFly5Y1l4vHjN2jRw81a9ZMCxcuVMmSJVW/fn0FBQVJkpns/frrrwoICNBff/0Va56rV69Wz5491a9fPy1fvlxt27aVv7//U89fvny5AgIC9M8//+jLL7/Uzz//rO+//94cM67zHj58qCtXrujhw4dW1798+bLu3bv3zFgAAEDirVy5Uq1atVKzZs20bNkyVa9eXXXr1rX6cm79+vWqXLmyfHx8NH/+fH3xxRcaOnSo+aVTo0aNFBAQoICAAM2fP19eXl6qWLGimZO0aNFCHTp0UNGiRc1+fn5+sR7lS0gsISEhmj17tvbs2aNffvlF48eP14IFCxK119KXX36p//73v/r22281f/58OTg4qHz58lZfLl6/fl1fffWVsmTJolmzZql3797q2bOnVq9ebfbp27ev5s6dq3Hjxmn8+PHaunWr/ve//1k9Xvf4o3wjR45UkSJF1LlzZ/M+uLu76+rVq/r333+tYrx586bVOAMGDNCsWbP0008/6ddff9U///wTq2g3bNgwfffdd/L399fy5cvVtGlT1apVSwEBAQm+NwAAvA1e6qN8kjR48GCNHj3aqu2PP/5Q1apVrfo0bNhQkvT111+bv8jr1q0rSfriiy9UsmRJhYaGytXV1Tyve/fu+vDDDyU92pBzw4YN+vHHH2NdT3pUXNm9e7fOnTsnLy8vSY++MSxUqJB27NihUqVKqXHjxurevbsWLFigDz74QNKjFT+NGzeWm5ub7ty581zxrl+/Xlu3btWNGzeUMmVKSdKYMWO0cuVK/fXXX/r444+t7kWTJk0kPUpYxo0bpx07dqhOnTrKmDGjJCldunTmHJ508OBBFSxYUI0aNZL0aAVTvXr1JCne82PaJSlbtmwaOnSohgwZooEDB8Z73pMrpRIby5MiIiKsHhVk2TwAALENGTJEHTp0UPfu3SVJBQoU0J49e/Sf//xHK1eulPSokNO8eXN999135nmVK1eWxWKRJDk7O8vZ2dk8NnbsWC1btkwrVqxQq1atlCpVKqVOnVqOjo7x5hsJjUWSMmXKpEmTJsnB4VHa2aVLFy1atEjDhg175nxDQ0P1448/atq0aWYOMXXqVOXIkUPjx4/XgAEDzL5NmzZVnz59JEn58uXTtGnTtGbNGlWvXl3BwcGaOHGi/vrrL1WvXl2SNH369KfOz93dXcmTJ5eLi8tT+z0pJCREv/76q2bNmqUaNWqY18qSJYvZJzw8XEOHDtWqVatUoUIFSVKePHm0Z88e/fe//1Xp0qXjHJt8CQDwNnrpK6Z69+5tfssU8+Pr62vV5/ENL9OlSxdvW8zKoRjlypWz+ly+fHkdPHgwzjgOHDigHDlyWCUWPj4+8vDw0IEDByRJjo6OatOmjaZOnSpJunbtmlatWhVrj6zExrtt2zZFRkaqQIECypYtm7Jly6asWbPq3LlzsR758/HxMf/s4OAgd3f3WPN+mlq1aunIkSNq3Lixpk+frsDAQNnZPf0f86ZNm9SwYUMVKFBAXl5e6tmzZ4IKT0kZy/Dhw+Xq6mr+PJ68AQCARyuwDx06ZBYyYlSqVMnMZaKiorRz506zIBIjpiglSXfu3NHXX3+t0qVLK3v27PLy8tKFCxcS9bs/IbHEyJcvn1mUkqQMGTIkOLc5fvy4IiIirK6TPHlylS1bNtZ1Hs+hnrzOiRMnFBkZaZWDuri4xDonKZw4cUIRERFWeaqbm5sKFSpkft6/f7/u3bunNm3aKHv27GZuOHPmzFi54ePIlwAAb6OXvmIqZo+pp7G3t09Qm2EYVp+TJ08e63N8G3RHRERYbWYZw9HR0eqcjz76SIULF9aFCxf0119/KUuWLKpcufILxXv//n1lzZrVar+oGKlTp07wOAlRsGBBHT9+XLNnz9Zff/2lLl26qEWLFpoyZUqc/Y8ePaqaNWtq0KBB+vbbb+Xm5qaNGzeaK9FexNNieTxBlqSBAwea33JKj74BJNkCAOD/REdH68GDB7HymcdzmejoaEVHR8eZ88To2rWrTpw4oe+++045c+ZUihQpVLdu3US95CQhscR4kdwmZqy4rhOzdUBCrhOzmfyTuePT7tPziok5rjw1Rsy2EYsXLza/0Iyr35PIlwAAbyOb7jH1og4dOmT1+cCBA8qdO3ecffPmzauzZ88qPDzcbLt+/bquXr2qPHnymG0FCxZUyZIlNW3aNE2dOlXt27ePVURJrIIFC+r8+fOys7OLtRH8448mPoudnZ0sFsszk7nMmTOrb9++WrZsmfbs2aNp06YpICAgzvM3btyonDlzyt/fX4ULF1bWrFkVGBj4zOvGbGD/+P28ceNGrGQ0vlie5OjoKBcXF6sfAADwf+zt7ZUrV65YK4X27dtn5jLJkiVT7ty5tWvXrnjHWb16tT7//HNVr15d3t7e8vDw0IULF2Jd62n5RkJiSQoxYz15nf379yfqOt7e3pKkw4cPm20PHz602vQ9LnHdBxcXF6v8R5LV/YvJRR/PUx88eGB1rQIFCshisejUqVOxcsP06dPHGw/5EgDgbfRGF6bGjBmjEydOSJLmzJmjdevWqWvXrnH2bdCggTw8PNSvXz9FRkbq3r176tGjh/Lly6dq1apZ9e3cubNGjx6tkydPqn379i8cZ+PGjZUjRw61bdvWLPrcvn1bI0eO1I4dOxI8jp2dnTw9Pc05x+W3337TggULzALRxYsXZbFY5OHhEef5Mcv3jx49KknauXNnrD264jovbdq0ypkzpyZPnizDMHTv3j2rb/CeFQsAAEi87t2769dff9WePXskSZs3b9Yff/yhHj16mH369OmjX3/9VcuXL5dhGAoPD5e/v7+5+bmXl5dWr16tqKgo3b9/Xz169LDauFt69MXS+fPnn7qKKiGxvChPT081a9ZM/v7+unHjhqKjo/XDDz/o9OnTsd4Q+DSZMmVS3bp1NXDgQAUHBysqKkpDhw7V9evXn3pe5syZY+VdxYsX1+zZs3Xnzh0ZhqGff/5Zx48fN49nzJhRDRo0kL+/v27fvq2oqCgNGjTIauP4zJkz64MPPlD//v3NIuL9+/c1d+7ceFe5AwDwtnrphanBgwfH+iYoqRKW1q1bq2rVqnJxcVHnzp01btw4FS9ePM6+KVKk0MKFC7Vz5065uLjIzc1NFy5c0Lx582It/W7ZsqUsFouqVq2qrFmzvnCcKVKk0Lp16+Ts7Kzs2bMrTZo0yps3r0JDQ632pkqIQYMGqV+/fvL09FSlSpViHa9atapmzZolDw8Pubu7q3379vrtt9/Mb++ePL9evXr64IMP9N577ylNmjRq3LixWrRokaDrTpkyRYsWLZKrq6uyZs2qfPnyWS2Jf1YsAAAgbjVq1IiVP505c0Zdu3ZVp06dVKFCBbm4uKhu3boaMGCA1duAu3btqm+//Vbt27dX6tSpzdwjJt/59ddftXbtWrm5uSlt2rS6ffu2ChcubHX9Zs2aKV26dHJ3d5eXl5eWLFkSK8aExJJQYWFhseZbtGhRSdL48eOVMWNGeXl5ycXFRT/++KP++usvZc+ePVHXmDBhgqKjo5UuXTp5eHjo4MGDKlGixFMfnevRo4f++ecf8z7cvHlT3377rflFm4eHh7Zt2xZrs/Lx48fLzs5O6dOnl4eHh44fPx5rj9VJkyapUaNGqlq1qlxdXZU+fXotWLAg1hYSAAC87SxGYjYwSqTAwEBFR0fHand2dlaaNGlkGIauXLmi9OnTm0lBdHS0AgMDlSFDBiVLlkzSo408r169qowZM8re3l537txR6tSptX37dpUuXVohISFydnY2+8e4fPmy0qVLF2v/gJCQEFkslngfo7tz544yZcqkiRMnWhVpnjfex0VEROjOnTtyd3e3ao9rbOnRBuwuLi7m2/xirnnz5k1Jjzb2jMvDhw8VHh6uNGnSxDoW1/kREREKDw+Xh4eHIiIidPPmzVh7g8V33eDgYPM6V65ckbu7u1KkSJGgWOITFhYmV1dXrXZwlvMLPkoJvI3KRoY/uxOARIn53RMaGmqzR6QePHgQ7yoeT09PcxPxBw8eKDg4WGnTprXaWPxx0dHRunXrljw8POLcluDWrVtydnZWihQpdPPmTfMxscfduXNHoaGhSps2rR48eKCIiIhYeyI9LZbQ0FA9ePDAarX03bt3FRYWJk9PT/MaT67Ykh49Rvf4m4Pv3r2ru3fvxrny+saNG3JycrLauzM4OFiSYuUfwcHBSpUqlRwcHJQzZ071799fn376qSTFex9u3bqle/fuKVOmTOaLXMLDw+Xk5CQHBwcFBQUpWbJksXLLmGslS5ZMt27dkoODQ6w+UVFRun37ttzd3Z/5wponkS8BSCrklkiIl5UrvdTC1MvyZGEqqY0cOVI///yzzp49G6vYhVeDRAt4OpIHIOm9DoUpvDxr166VYRiqWrWqoqOjNXToUI0cOdLc5+lNRL4EIKmQWyIhXlau9NLfyvcmOXjwoKpXr65///1Xs2bNoigFAADwlihUqJA6dOigZs2a6f79+8qRI4eWL1/+xhalAAB4W7yRhSlnZ2ddunTpqW8teR758+fXnj17lC5duqfuNwAAAIA3S4YMGbRixQpFREQoKirKapsEAABgO29kYcpisbyUb7eSJUumzJkzJ/m4AAAAeD08ufcoAACwrZf+Vj4AAAAAAAAgLhSmAAAAAAAAYBMUpgAAAAAAAGATb+QeU3h3lLoVyCu7AQAAnoJ8CQDwJmPFFAAAAAAAAGyCwhQAAAAAAABsgsIUAAAAAAAAbILCFAAAAAAAAGyCwhQAAAAAAABsgsIUAAAAAAAAbMLB1gEAT7PDPZOcLRZbh4E3VNnIcFuHAADAS0e+hHcVuR7wdmDFFAAAAAAAAGyCwhQAAAAAAABsgsIUAAAAAAAAbILCFAAAAAAAAGyCwhQAAAAAAABsgsIUAAAAAAAAbILCFAAAAAAAAGyCwhQAAAAAAABswsHWAeD1smPHDgUEBOjevXvy9vZWzZo1lSpVKqs+0dHR2rBhgw4cOCA3Nzf5+fkpS5YscY63bt067d27V5kzZ1bjxo2VIkWKVzENAACAl+Ls2bNau3atgoKC5OnpqZo1aypTpkyx+h04cEBbtmyRJJUpU0ZFixaNc7wjR45o3bp1Sp48uRo2bChPT8+XGj8AAK8bVkxBkhQcHKzq1aurbt26OnbsmEJCQjRmzBjlypVL69evN/tdvHhRhQsX1ogRI3TlyhUtX75cefPm1axZs6zGu3Pnjvz8/NS+fXsFBgZqw4YNKleunCIjI1/11AAAAF6YYRjq16+fChYsqA0bNujOnTtatmyZvL29NWbMGKu+tWvXVvv27XXs2DEdPHhQFStWVI8ePWKNOWDAAJUuXVoHDx7UqVOnVL16dR06dOhVTQkAgNcCK6YgSWrdurXOnTuno0ePKl26dGa7v7+/GjRooP379ytXrlxKliyZ5s2bp3z58pl9vvnmG3Xp0kUtWrSQnd2jWmffvn115swZHTx4UGnSpJEkHT9+XBaL5dVODAAAIAl8//33GjdunDZv3qxSpUqZ7StWrFCdOnXk5eWlFi1aSJL69++vypUrm31atGihqlWrql27dipevLgkadasWRo7dqy2bdumEiVKSJJCQ0MVGhr6CmcFAIDtsWIK2r17t1auXKmhQ4daFaUkafDgwUqdOrXGjh0rScqYMaNVUUqSSpUqpfDwcN26dUvSo6Rq2rRp6tu3r1mUkqR8+fIpWbJkL3k2AAAASevBgwcaOXKkOnbsaFWUkh6tjmrYsKG+/fZbs+3xopQk85xz586ZbT/++KMaNmxoFqUkydXVVVmzZn0ZUwAA4LXFiinon3/+kSRVr1491rEUKVKoXLly2rRpU7znz549W7ly5TKLWvv27VNkZKTKlSun2bNn6+LFi8qVK5fq168fb2EqIiJCERER5uewsLAXmRIAAECSOXz4sLntQVz8/Py0aNEiBQUFycPDI9bx2bNny97eXiVLlpQkPXz4UHv37lXbtm21adMmBQQEyNPTU7Vr1471JeHjyJcAAG8jVkxBt2/flqOjo9XqpsdlypRJwcHBcR6bMmWKZs2apV9//dVsCwkJkSR16dJFf/31l27cuKGvv/5ahQsXjnec4cOHy9XV1fyJbzN1AACAV+327duSHq0cj0vM5udx5TkHDx5Ur1695O/vr2zZskl6tBfnw4cPNWPGDH3++ee6deuWpk+frjx58iggICDeOMiXAABvIwpTUNq0aRURERFv0SgwMFAZMmSI1T537lx17dpVkyZNsvoG0dnZWdKjZesLFy7U6NGjtWvXLoWGhmr06NFxXmPgwIHmvgqhoaG6dOlSEswMAADgxaVNm1aSdPXq1TiPBwYGSlKsfOnYsWPy8/NT8+bN9Z///MdsT5kypSwWiyIjI7V9+3Z9//33Wrt2rSpUqKC+ffvGGwf5EgDgbURhCipfvrwkafXq1bGO3b9/X1u2bFGFChWs2ufNm6c2bdpowoQJat++vdWx/PnzS5J8fX3NNmdnZ7333ns6ceJEnDE4OjrKxcXF6gcAAOB14OPjozRp0sSZK0nSmjVrVKRIEav85cSJE6pSpYrq1KmjiRMnWr0AJnny5MqZM6dKly5tvjhGepQ7xZcrSeRLAIC3E4UpqHjx4qpZs6a++uorBQUFWR0bPHiw/v33X6tXHC9YsECtW7fWr7/+qo4dO8Yaz8vLS2XKlNHWrVvNtn///Vf79+9XwYIFX95EAAAAXoJkyZJpwIABmjJlinbu3Gl1bOXKlVq8eLG++uors+3kyZOqXLmyatWqpUmTJlkVn2I0b95c27dvV3R0tNm2detWciUAwDvHYhiGYesgYHvBwcFq3ry59u/fr6ZNm8rV1VWbNm3SpUuXNH36dPPtMgcPHlTx4sVVoEABNWzY0GqMrl27ytPT0+xXpUoVVahQQbly5dLy5cuVPHlybd68OUHf7oWFhcnV1VWrHZzl/Ng3jEBilI0Mt3UIAN4gMb97QkNDWYmCWKKjo9W/f3/9+uuvatSokXLkyKFjx45pzZo1Gjp0qHr27Cnp0cbm2bNn1927d/XZZ59ZFaVq1qyp0qVLS3r0961y5cpycHBQxYoVtXfvXu3fv19r1qzR+++/n6CYyJfwriPXA16tl5Ur8VY+SJLSpEmjNWvWKCAgQAEBAbp+/bp2796t0aNHW73yOGXKlPriiy+eOV7hwoV1/PhxLVy4UMHBwRo+fLjq1q0re3v7lzkNAACAl8LOzk5jxozRp59+qjVr1ujmzZvauXOnqlWrZhalYnTu3PmZ47m4uGj79u1atGiRzp49qw8//FDz5s2Tm5vbS5oBAACvJ1ZMIV7Dhw/XkCFDtGXLFhUvXvyVXptvAJEU+BYNQGKwYgqJtWXLFlWuXFljxoyx2vbgVSFfwruOXA94tVgxhVfO399fjo6OOnTo0CsvTAEAALzuypUrp1mzZunkyZOKjIxU8uTJbR0SAABvHApTiJfFYlGfPn1sHQYAAMBrq2nTprYOAQCANxpv5QMAAAAAAIBNUJgCAAAAAACATVCYAgAAAAAAgE1QmAIAAAAAAIBNUJgCAAAAAACATfBWPrzWSt0KlIuLi63DAAAAeG2RLwEA3mSsmAIAAAAAAIBNUJgCAAAAAACATVCYAgAAAAAAgE1QmAIAAAAAAIBNUJgCAAAAAACATVCYAgAAAAAAgE042DoA4Gl2uGeSs8Vi6zCeqmxkuK1DAAAA77A3IV96HZHDAcDrgRVTAAAAAAAAsAkKUwAAAAAAALAJClMAAAAAAACwCQpTAAAAAAAAsAkKUwAAAAAAALAJClMAAAAAAACwCQpTAAAAAAAAsAkKUwAAAAAAALAJClOI5d9//9WwYcPk6+urvHnzqk6dOpo1a5aio6NtHRoAAMBrYebMmfLz81PevHlVqVIljRw5UuHh4bYOCwCAN46DrQPA66d169Y6ceKExowZozx58ujs2bOaPn26QkJC9Mknn9g6PAAAAJv68ccf9dVXX2ncuHHy9fXVzZs3tXLlSvXs2VNTpkyxdXgAALxRWDEFK6GhoVqyZIm+++471a5dW97e3qpevbqmT5+ujz/+2KrvtGnTVK5cOeXOnVu1atXSpk2bzGNLly5Vzpw5dfLkSbNt6tSpyp8/v65evfrK5gMAAJDU/vjjD3Xs2FEdOnRQnjx55Ovrq6FDh2rixIlW/bZt26a6devK29tb5cuX12+//SbDMCRJly5dUp48eTRr1iyz/5EjR5Q9e3atXr36lc4HAABbojAFKw4ODrJYLDp8+HCsY/b29uafhw0bpu+++07+/v5avny5mjZtqlq1aikgIECSVK9ePRUvXlytW7fWgwcPdOrUKXXv3l39+/dXxowZX9l8AAAAkpqjo6NOnDihhw8fWrU/nitt2rRJtWvXVv369bV8+XJ9+eWXGjFihEaNGiVJypIli3r16qWuXbvq/PnzioiIUKtWrVShQgVVr179lc4HAABbshgxX9sA/9+3336rwYMHq3DhwqpYsaJ8fX1VvXp1ubm5SZLCw8OVIUMGrVq1ShUqVDDP+/TTTxUSEqKZM2dKkoKDg1WkSBE1a9ZMmzdvVrZs2TRv3rw4rxkREaGIiAjzc1hYmLJkyaLVDs5ytlhe3mSTQNlI9pMAgLdBWFiYXF1dFRoaKhcXF1uHg9fYqlWr1LRpU6VJk0bVqlVT6dKlVbNmTWXLls3sU758efn5+embb74x2+bMmaNevXopMDDQbKtXr55u376t999/XytWrND+/fvj/fv3JudLryNyOABInJeVK1GYQpwOHz6sxYsXa/v27frnn39kb2+vmTNnqmbNmvrnn39UoUIFZcmSRXZ2djIMQ4ZhKCwsTHny5NHOnTvNcTZt2qTKlSsrU6ZMOnjwoNKmTRvn9QYPHqwhQ4bEan8TEi2SGgB4O1CYQmJcu3ZNCxcu1D///KPNmzfr2rVr+vbbb/XFF18oOjpajo6OSpMmjVKmTGnmShEREbpx44bu3LkjZ2dnSdLNmzdVqFAhBQUFafPmzSpbtmy813yT86XXETkcACQOhSnYTHh4uBo0aKAzZ87owoULWrNmjapXr669e/cqXbp0Vn2TJ0+u9OnTm59nz56tVq1aycvLSwcOHIi3MPUmfwNIUgMAbwcKU3he0dHRGjhwoEaNGqXTp08rS5YsSp48uX755RfVr18/Vv/MmTPL8v/zm9OnT+u9997T/fv3n1mYepPzpdcRORwAJM7LypXYYwrPlDp1atWrV09Xr15VdHS0ChQoIIvFolOnTsnLy8vq5/Gi1MWLF/XJJ59o7NixSp8+vbp06RLvNRwdHeXi4mL1AwAA8Caws7NTy5YtZRiGrly5omTJkil37tw6fPhwrFzJy8vLLEo9fPhQrVu3Vq1atdS1a1e1bdtW4eHxF0vIlwAAbyMKU7ASFBSkjh07av/+/YqKipIknTlzRlOnTlWlSpVkZ2enzJkz64MPPlD//v21a9cuSdL9+/c1d+5c8xXJ0dHRatu2rcqUKaPevXtr5syZWrFihaZNm2arqQEAACSJTz/9VKtWrdK9e/ckSSEhIRo9erTc3d1VpEgRSdLnn3+uSZMmadasWYqOjlZUVJQCAgL01VdfmeMMHjxYV65c0W+//abRo0fLyclJ3bt3t8mcAACwFQpTsJImTRpVqFBBH330kVxcXOTu7q7ChQurUKFCmjFjhtlv0qRJatSokapWrSpXV1elT59eCxYsUOXKlSVJI0aM0LFjxzR16lRJUt68eTV27Fj16NFDZ8+etcncAAAAkkLz5s31888/K3369EqXLp0yZMigK1euaOXKleYqps6dO+u///2vPv/8c6VKlUpubm7q16+f+ca9LVu2aOTIkfrjjz+UNm1apUiRQjNnztTs2bM1d+5cW04PAIBXij2mEK/IyEjduXMn3n2hJCkqKkq3b9+Wu7u77Oz+r8557do1c9PPx125ckUuLi5KnTr1U68d8+zqm7BnAvsTAMDbgT2mkFiGYejWrVtKmzatVR70pFu3bilVqlRydHQ024KDgxURESFPT0+rvjdv3pTFYpGHh8czr/8m5UuvI3I4AEicl5UrOSTZSHjrJE+e/KlFKUmyt7ePtQG6pFhJVozMmTMnSWwAAAC2ltACkru7e6y2J7+8ixFXXgUAwNuMR/kAAAAAAABgExSmAAAAAAAAYBMUpgAAAAAAAGATFKYAAAAAAABgExSmAAAAAAAAYBMUpgAAAAAAAGATFKYAAAAAAABgEw62DgB4mlK3AuXi4mLrMAAAAF5b5EsAgDcZK6YAAAAAAABgExSmAAAAAAAAYBMUpgAAAAAAAGATFKYAAAAAAABgExSmAAAAAAAAYBMUpgAAAAAAAGATDrYOAHiaHe6Z5Gyx2DqMOJWNDLd1CAAAAK91vpQUyLkA4O3GiikAAAAAAADYBIUpAAAAAAAA2ASFKQAAAAAAANgEhSkAAAAAAADYBIUpAAAAAAAA2ASFKQAAAAAAANgEhSkAAAC8cvfv31dkZKStwwAAADZGYQpWIiIitGfPHm3ZskVXr16Nt194eLh27typkydPKjo6Ot5+hw4d0tq1axUVFfUywgUAAC9BfL+37927l2TFpKZNm+qLL75IkrEe96IxPu/5SXlvAAB4l1CYgiQpOjpa3377rTw8PPThhx/K399fefLkUY0aNXT58mWzX3BwsDp16qTs2bOre/fuqlKlinx8fLRv3z6r8ebOnauSJUuqWrVq8vPz07179171lAAAQCJcvXpV7dq1k7u7u1KmTKlcuXJp0KBBCg4ONvv4+flp2LBhNozy2V40xuc9/024NwAAvI4oTEGS1L9/f40dO1YrVqzQ4cOHtWXLFl2+fFkODg6qUKGCQkNDJUk3b96Ur6+vrl+/rh07duj8+fMqXLiwmjRpYjXexYsXNW7cOP3222+2mA4AAEikhg0b6uLFi9q1a5ciIiK0YcMGJU+eXEuXLpX06NG76OhoRUZG6s6dO7pz544Mw9Ddu3f18OFDq7Hie0zvyX7xia/fvXv39ODBA0mPvlQzDCPWdeOKMT5PrgyL7/yIiAjduXNH//77b5zjJcW9YXU5AOBdRWEKOn/+vH766Sd99dVXKl++vNnu6uqqyZMn6/r16/rxxx8lSXny5FHHjh3l4OAgSXJwcFDbtm117tw53bhxwzy3b9++KlWq1CudBwAAeD5BQUHauXOn+vbtq5w5c0qSsmbNqi+//FLt2rWTJHXo0EE7duzQ6NGj5enpKU9PT928eVMFChTQtGnTrMZr2LCh1WN6wcHBatiwoVKkSCEPDw917NhR//77r9U5Dx8+lL+/vzw8PJQyZUrlyJFDEydOtOpTvnx59ejRQ7Vq1ZKbm5tcXFzUvXt3s1gUX4xPmjFjhnLmzCknJydlyZJFI0eOVFRUVLznd+nSRZ6ensqQIYNSpUql6tWr69SpU+Z4L3Jv4osFAIB3BYUp6O+//1ZUVJRat24d65inp6eqVq2qFStWxHt+QECA3NzclC5duueOISIiQmFhYVY/AADg1XBxcVGKFCm0dOnSePdJmjVrlsqUKaMvvvjCXBWUPn36BI3frVs3Xbp0SWfOnNGFCxfk7OysjRs3WvXp3r27/vnnHwUEBCgiIkIzZszQgAEDtHjxYqt+f/zxh7p3767g4GBt3LhRU6ZM0Zw5cxIcY2BgoNq3b6/Ro0crIiJCu3fvVlhYmC5duhTv+dOmTTM/BwYGKlu2bGrRooVZEHvee/O0WOJCvgQAeBtRmIKuXLmi5MmTy9PTM87j2bNnt9pn6nE7d+7UqFGjNHjwYFkslueOYfjw4XJ1dTV/smTJ8txjAQCAxEmePLkmTpyoWbNmycPDQzVq1NC3336ro0ePvvDYN27c0F9//aVRo0YpW7ZscnZ21ujRo+Xu7m72uXnzpiZOnKgxY8bIy8tL9+/f1/vvv6+2bdvqzz//tBqvTZs2ql27tuzt7VWsWDFVqVJF27dvT3A8169fV3R0tCpUqCCLxaIMGTLou+++U/bs2RN0voODg77++mvt27dPFy9eTPB1kyIW8iUAwNuIwhTk5OSkyMjIeDcoDwkJkZubW6z2I0eOqE6dOmrfvr169uz5QjEMHDhQoaGh5k983xQCAICXo02bNrp8+bKmTp2qQoUK6a+//lLhwoU1adKkFxo35g2+RYsWNdscHR1VsGBB8/OhQ4cUFRWlatWqycPDQ+nSpVP69Ok1efJkq60CJClHjhxWn11dXRUSEpLgeAoXLqzq1avLx8dH3bp10/z582M9VvikrVu3qly5ckqZMqXc3d1VoEABSXrhfCWxsZAvAQDeRhSmYCaKu3fvjnXMMAzt2bNH7733nlX70aNHVaVKFTVq1Ejjx49/4RgcHR3l4uJi9QMAAF4tV1dXNWnSRKNHj9bhw4fVtGlT+fv7P/WcuFZMR0dHm3+2t7eXFHtz77j2UTp27Jj5KFzMzz///PPM6yWGvb29Vq5cqUWLFiljxowaMWKEvL29rfaMety9e/dUr149Va1aVYGBgbp//74CAwMlPXsz94Tcm8TEQr4EAHgbUZiCqlWrprx582rQoP/X3n1HRXG9fQD/0qVIV8ACCghKEMUGioiKvUTFqDFGo7FrUBNNRE1iEmOMUdM0iUasib1h7xUs2EBsIKIgCIigFJUicN8/eJmfyy6wNBfN93OO57gzc+c+z50tlzszd+bKdRI3btyI27dvY9KkSdKy27dvo3Pnznj33XexYsWKCncQiYiIqPpRU1NDy5Yt8eLFC2kwRVtbW66vYGJigpSUFOl1fn4+7ty5I712dHSEpqYmLly4IC179uwZbty4Ib1u1qwZtLS0sH///grHrSjGotTU1ODu7o4vv/wSFy9ehImJCTZv3qywfFRUFJ4+fQpfX1/pCnJFtw6Wp21Ki4WIiOi/QFPVAZDqaWhoYPv27ejWrRu8vb0xduxYGBkZ4fTp0/jzzz+xZMkSeHh4ACi4ZL1z586oU6cOBg8ejOPHj0v7cXd3h4GBAQAgIiICsbGxCAsLAwCcOnUKNWrUQPPmzWFubv76kyQiIqJiPXr0CD4+PvD19UXLli1hbGyMS5cu4ddff4WPjw/U1QvOZTZs2BDBwcF4/PgxdHV1oa+vj06dOsHf3x+9evWCkZERFi9ejJiYGGnfpqamGDVqFD777DPpFr3CW9IKmZmZYfr06Zg1axb09fXRpUsXJCcnIyAgADo6Ovjiiy+UzkVRjK+eRDt06BC2b9+OsWPHws7ODteuXUNcXByaNGmisLy1tTVq1qyJ33//HZ988oncCbuS6i2tbUqLhYiI6L+AA1MEAHB2dsatW7fg7++PXbt24dGjRwgKCsJff/2FCRMmSNs9evRImhNi4cKFMvtYsWKFNDB19OhRBAQEAAC8vb3x66+/AgDmz5/PgSkiIqJqxsLCAkuWLMEff/yBr7/+GqmpqahXrx4mTJiAGTNmSNv5+flh/PjxcHJyQmZmJu7du4e5c+ciJSUFPXr0gLGxMYYMGYI+ffpAR0dHKvfrr79i2rRp6Nu3L0xMTODj44N+/frJbLNgwQLY2NhI29avXx8DBgyQGQTS09ODtra2TOw1atSApub/urSKYnz1CXldu3ZFfHw8pk2bhrt376Ju3bpYsGAB3nvvvWLL79ixA35+fli2bBmsra0xd+5cTJo0SbpNsbxtU1osRERE/wVqovA5t0RFDBs2DCdPnsTVq1eLfWJfVUlPT4eRkRGOaOpDv5reKtguJ0PVIRARUSUq/O1JS0vj3D30RngT+kuVgX0uIqLqoar6Spxjioq1YsUKuLu7Y+XKlaoOhYiIiIiIiIjeQryVj4plYGCAnTt3qjoMIiIiIiIiInpL8YopIiIiIiIiIiJSCQ5MERERERERERGRSnBgioiIiIiIiIiIVIIDU0REREREREREpBIcmCIiIiIiIiIiIpXgU/moWnNLiYehoaGqwyAiIiKqtthfIiKiNxmvmCIiIiIiIiIiIpXgwBQREREREREREakEB6aIiIiIiIiIiEglODBFREREREREREQqwYEpIiIiIiIiIiJSCQ5MERERERERERGRSmiqOgCikgSb1YG+mpqqw0C7nAxVh0BERESkUHXpLxWH/SgiIioJr5giIiIiIiIiIiKV4MAUERERERERERGpBAemiIiIiIiIiIhIJTgwRUREREREREREKsGBKSIiIiIiIiIiUgkOTBERERERERERkUpwYIqIiIiIiIiIiFSCA1OVIDExERs3bsSyZcvw+PHj11bvsmXLEBcXV+I2f//9N6Kjo19PQEREREQKZGVlISAgAH/++SdCQ0NfW727du1CcHBwidvs378fZ8+efU0RERERUVGa5Sn0/PlzBAYGIi4uDubm5mjWrBkaNmxY2bEpZf/+/TAyMkL79u1VUn9cXBycnZ3RqVMn1K1bF5mZma+tbl9fXzRu3Bj16tUrdpvPPvsMmzdvRoMGDV5bXERERP91eXl5CA4ORkREBGrUqIHGjRvD1dVVJbGEhoYiIiICQ4YMUUn9+fn56NChA9TU1NCqVSs4Ojq+trp/++03tG/fHm5ubsVu89dff6Fx48bw8PB4bXERERHR/5R5YGr16tWYMWMGbG1t0bx5czx9+hRffPEFXFxcsG3bNqipqVVFnMX666+/YG9vr7KBqQMHDsDCwgK7du167XVPnjwZ9evXf+31EhERUfGCg4MxfPhw5OTkoEOHDhBC4Oeff4aamhq2bNny2k/mHTt2DP/++6/KBqbu3r2LS5cuIT4+HlZWVq+17gEDBsDOzu611klERERlU6aBqW3btmH06NH4999/MWzYMGm5EAIBAQEQQkgDU+np6Th69CiePHkCFxcXuTNVW7ZsgaOjI5o3by4tO3LkCLS0tNCpUycAwO7du1G7dm3Y2Njg3LlzyMvLg7e3N8zNzQEAx48fR3R0NJ4/f45ly5YBAD788EOcPn0atWvXRt26dXHq1CkYGhqibt26uHv3LoYOHSoTx9mzZxEfH49BgwYpzLmkPA4dOoT9+/cjKysLy5Ytg7GxMT788EO5fRTm8Wo8/fv3BwA8fvwYx48fR05ODlxdXdG0aVOZsllZWTh69CiSk5PRtGlTtGrVSlrXuHFj6Ovry2wfGhqKK1euwMbGBp6engpzKqnOly9fYsWKFRg0aBASEhIQFhaG2rVro2vXrtDQ0JDZT0pKCk6cOIGsrCx07NhRbpCstNyIiIjeNlFRUejatSs+/PBD/P7779DU/F9X6+rVqxBCSK/z8vKkvky9evXQtWtXaGlpSeuDgoKQlpaG3r17S8tu3ryJkJAQqb9x+fJlREdHo0ePHggMDMTjx4/h7u4OBwcHAEB4eDjOnj2L5ORkqa/UqVMnZGZmIjo6Gt26dcORI0eQlJSE/v37Y+fOnRgzZgxq1Kgh1RkdHY39+/dj3LhxMvEpk8f169exbt06AMCmTZugra2NUaNGyfVfCvN4NZ5JkyYBKOgLHTt2DI8ePYK9vb109dWrzpw5g7t378LGxgYdOnSQ6re2tkatWrVktn348CGOHj0KExOTYk9sllbn1q1bYW9vD3Nzc5w/fx4aGhro0qULjI2NZfaTnZ2NEydOID4+Hi1atJC7ak6Z3IiIiN52ZRqYmj17Nvr37y8zKAUAampqGDBggPQ6LCwMXbt2hY2NDRo1agQ/Pz9069YNmzZtkrZZsGABPvzwQ5mBqdWrV8PAwEAamFq6dCmePHmC58+fw93dHeHh4ZgyZQouXboEa2trJCQk4Pnz59DU1ER4eDiAgg7A0qVLkZKSgvT0dHh6esLV1RWWlpb48MMP4enpKXPr28SJE9G3b1+FA1Ol5fHw4UM8fvwYmZmZCA8Pl+v4FFIUDwDs2LEDo0ePhpeXF4yNjTFjxgwMGzYMv/zyCwAgISEB7u7usLKygrOzM/7++280atQI69evByB/K9+iRYvw1VdfoW/fvkhPT4efnx9evnwpE0tpdWZnZ8PX1xdbtmxBdnY23nnnHRw9ehRNmzbFwYMHpf1s27YNo0aNQsuWLWFtbY358+dj4cKF6Nevn1L1EBERvY1+/PFH1KxZE4sXL5YZlAKAFi1aSP/PyMiAt7c3Hj9+DE9PT5w/fx41atTAyZMnpRNw27dvx927d2UGpgIDA7F48WJpYOrQoUNYsWIF5s6di6ZNmyIzMxNjx47F9u3b0bdvX6SlpeHx48fIzs6W+krNmzfHqVOnsHz5chgYGMDFxQW1a9fGxx9/jK+++gqmpqb44IMPpDqXLFmCq1evYvLkyXL5lpbH06dPERMTAwCIiIiAlpaWXN+kMI+i8QAFA2s9e/ZE3bp14eDggCVLlsDMzAyHDx+Gnp4e8vPz0bt3b0RERKBTp07YunUrZs2ahSNHjsDY2FjuVr7Tp0+jV69eaNOmDWrXro2pU6dCU1MTjRs3lmIprU4A+Omnn6Curo709HS4u7sjLCwMn376KUJCQqTjd+vWLfTu3Ru6urpo3bo1/vjjD/To0QM//PCD0vUQERH9JwglxcTECABi5cqVpW7btm1bMWjQIJGfny+EECIiIkJoa2uLrVu3Sts0a9ZMLFq0SKbckCFDxOjRo6XX3t7eol69euLp06dCCCHy8/NF8+bNhZ+fn7RN7969xdSpU2X24+3tLczNzcXjx49lljs7O4t58+ZJry9evCgAiMjIyHLnsWDBAtGyZcsS20NRPPHx8UJPT0+cPHlSWhYXFycMDQ2lZb/99ptwcXGR6hdCiFOnTkn/ByCOHj0qhCg4Ptra2mLnzp3S+tmzZwsAYu/evUrXmZGRIQCI4cOHS9tERUUJNTU1ceHCBSGEEA8ePBA6Ojri559/lrZ58eKFuHTpktL1FJWVlSXS0tKkf7GxsQKAOKKpL85qGaj8HxERvf3S0tIEAJGWllbufTRs2FAMGzas1O1mzZolbG1tRWpqqhBCiGfPnol33nlHTJo0Sdpm6tSponfv3jLl/vrrL2FnZye9njdvnlBXVxfBwcEy5dzd3aXXixYtEs2aNZPZz7x58wQAceLECZnln3zyifD29pZeZ2VlCVNTU+Hv71/uPM6fPy8AiIyMjGLbo7h4mjdvLr755hvp9cuXL4Wbm5uYO3euEEKIa9euCTU1NZGcnCxtc/36dem1l5eXmDNnjhBCiLy8POHk5CR8fX2lbY8cOSIAiOnTpytdpxBCtGzZUjg4OIjnz59L29jZ2Ykff/xRCFHQZ3V2dhY+Pj7i5cuXUrkzZ86UqZ6iqnt/if0oIqK3W2X0lRRR+oqpwqfNlTTRNgAkJSXh/PnzCAwMlC5FdnBwQO/evREQEFDsLXPF6dOnj3RZtJqaGtq0aYOoqKhSy/Xt21c6Y1VozJgxWLp0KebMmQM1NTWsXr0aXl5esLe3r/I8isaza9cuaGtrIzw8HLdv3wZQcEtkrVq1EBQUhI4dO8LMzAwJCQm4dOkS2rRpAwDw8vJSuP+DBw/CzMxM5so1X19f6aycsnUWevWqOFtbW9SqVQtRUVFwc3NDQEAA9PT0MGXKFGkbXV1d6TbDstRTaMGCBfj222+VaksiIqLq6vHjx6X2lQAgICAAo0aNgpGREQBAX18f48aNw8KFC/HHH3+UqU4HBwepnwAAbm5u2LhxY6nlGjZsKF2lXmjs2LFo3rw5YmJiYGNjg4CAAGRnZxc7P1Vl5lE0nlu3biE0NBQ+Pj5Yvnw5hBAQQsDU1BRBQUEAAGNjY6ipqeHgwYN4//33oampCWdnZ4X7j4yMxK1bt7Bz505pWdeuXWUmY1emzkI+Pj7SlU2amppo2bKl1Ee9ceMGbty4gfXr18tcOVc4zUJZ6nkV+0tERPQ2UnpgqrDD8ejRoxK3i42NBQC5+YZsbGxw5cqVssYn1VtIS0sLOTk5pZYrvAT8VcOHD8fMmTNx5swZuLm5YfPmzfjtt98Ulq/sPIrGExsbCy0tLdy4cUNmeY8ePdCkSRMAwPvvv48bN26gb9++0NTURJcuXTBt2jSFT/WJi4tD3bp1ZZZZWlrKdIaUqbNQSe0eHx8Pa2truTmnylNPoVmzZuGzzz6TXqenp3NidyIieuMYGRmV2lcCCn4rFfUxEhISkJubK3cbYGl1vqoifSUXFxe0bNkSa9euxdy5c7FmzRoMHjwYBgYGVZ6Hor4SUNDHebVNbW1tpcE/a2trrFu3Dt9//z0mT56MDh06YNSoUfDx8ZHbf1xcHAD5k6yvxq9MnYVK6ysBKHai+7LU8yr2l4iI6G2kdG/Bzs4OFhYWOHPmDD766KNit7O0tARQcMbQxsZGWp6UlCStAwANDQ3k5eXJlH3x4kWxHZ/KYGpqCh8fH6xevRoPHz5Efn4+3nvvPYXbKptHRWLJzc3F0qVLi53kUkNDAwsWLMAPP/yAmzdvYvny5fDw8MDdu3dRp04dmW1r166N5ORkmWWpqanIzc0tU53KMDMzQ2JiYoVyK0pHRwc6OjrljomIiKg6aNeunfTAluJO4AAF/YzCq9ELJSUlwdzcXBrMKa6vVNXGjh2LBQsWYNSoUTh69ChOnz5d7LbK5FFepqamAIBJkyahWbNmxW734Ycf4sMPP0RcXBx27dqFoUOHYsuWLdKDZgoVDnwlJyfLTL7+avzK1lkaMzMzAEBiYqLchOgVqYf9JSIiehupK7uhmpoavvjiC6xbtw4XLlyQW3/z5k0IIVCnTh04Ojpiw4YN0ronT55g//796Ny5s7TM2toat27dkl6npaXh/PnzZU7A0NCwTJ20MWPGYPv27Vi6dCmGDBlS7OSSyuZRXr1790ZqairWrl0rs/zp06fSmbM7d+4gLy8PampqcHZ2xvz585GZmYl79+7J7a9Tp06IiYmROTavxq5sncro1asXkpKSEBAQILO8cHLTyqqHiIjoTTNjxgzcu3cPP//8s9y65ORk6cRO586dsXnzZmngSQiBDRs2yPWVwsPDkZ+fLy07dOhQmWMqa19p6NChePz4MT7++GPY29sX++Q6ZfMoL1dXV9SrV0/uwSm5ubm4e/cugIIHxaSlpQEouBLK19cXTk5OcldtA4CjoyPq1Kkj8zCeGzduICwsrEx1KqN58+aoW7culi9fLrM8Ojq6UushIiJ6G5TpVNann36Khw8folOnTnjvvffQvHlzPH36FKGhoYiLi8PVq1ehrq6OZcuWoU+fPkhJSYGDgwM2bNiAxo0bY8yYMdK+xo0bh379+qFmzZqwsrLC1q1boa2tXeYE2rVrh++++w62trYwMDCQnlJTnE6dOsHKygoXLlwo8QlxampqSuVRXk5OTli0aBHGjRuHU6dOwcnJCffu3cOJEyewa9cuWFhY4PTp0xg4cCB69OiB2rVrY8+ePWjatClatmwptz9nZ2eMGTMGffv2xYQJE5Ceno7du3fLnK1Upk5lY583bx6GDBmCUaNGoX79+jhy5AhGjBiB0aNHV1o9REREb5o2bdpg/fr1GDduHA4fPgxvb2/k5+cjPDwc586dw549e2BpaYlvvvkGrVu3RufOndGjRw+cOHECt27dwurVq6V9DRkyBN988w369+8PDw8PnDhxApGRkSVeiaVI27ZtERUVhZkzZ6J+/fpy80oVVbNmTQwePBhr1qzBggULStxWmTzKS1NTE//88w/69euH2NhYdOnSBcnJyTh8+DC++OIL2NvbIyEhAYMHD0bXrl1hZ2eHa9eu4d69exg4cKDc/rS1tfHTTz9h5MiRePDgAWrXro1169bJ3EKoTJ3Kxr5mzRr0798f9+/fh7u7O65fvw5tbW2sXbu20uohIiJ6Gyh9xRRQMFizZMkSXLt2DS1atEB8fDz09fXxySefICQkBOrqBbvr0qULwsLC4ODggKdPn8LPzw9nzpyRGSTp2bMnTp8+DRMTE2hoaGDDhg34/vvvZc6w9e/fH+3atZOJwcvLS+axyZMmTcLvv/+Op0+fIjw8HNnZ2QrLvZpD9+7d4eTkBHd39xLzVSaPli1bFjshaEl5AMD06dMREhICR0dHJCcno3Xr1rh06ZI0aWfh454Lb9MbPXo0Lly4AF1dXQDA5MmTZeYVWLFiBZYuXYrMzEzY29vj4sWLmDJlisz8BqXVqa2tjcmTJ8vdrjhq1CiZ+aHmzJmDoKAgWFhYICsrC/Pnz8fo0aOVroeIiOhtNXToUERHR2Po0KF4+vQpsrOz0atXL1y/fh3vvPMOAMDKygrXr1/HwIED8fjxY/Ts2RM3btyAra2ttB9LS0uEhobC3d0dz58/x/Tp0/HPP/9g+PDh0jatW7eWeyCLvb09xo0bJ71u2rQpTp06BW1tbURERODp06cKy72qZ8+e0NDQKHH6hrLkMXny5BJPQBYXT8eOHXHnzh30798fycnJqF+/PgICAjBixAgAQIsWLXDx4kU0b94cjx49QsuWLREeHi71WQYMGCDT3xs2bBhOnTqFmjVrokaNGtizZw+++eYbmavCSqsTKBg0LHqisFu3bujSpYv0umvXrggPD0f79u3x9OlT9O/fX2bATpl6iIiI/gvUhBBC1UG8Trm5ubC1tcX06dMxdepUVYdDxUhPT4eRkRGOaOpDvwLzYVWWdjkZqg6BiIiqWOFvT1paGgwNDVUdjkp98MEHyMzMxK5du1QdCpWguvWXisN+FBHR26Gq+koVm5XyDbN27VocOHAA+fn5Mlf3EBERERFw9OhRnDhxAjt27MDZs2dVHQ4RERH9B5TpVr43XVRUFBo1aoRTp05V6dP/iIiIiN5ECQkJyMnJwcGDB9GqVStVh0NERET/Af+pK6bmzZun6hCIiIiIqi3Ob0RERESv23/qiikiIiIiIiIiIqo+ODBFREREREREREQqwYEpIiIiIiIiIiJSCQ5MERERERERERGRSvynJj+nN49bSjwMDQ1VHQYRERFRtcX+EhERvcl4xRQREREREREREakEB6aIiIiIiIiIiEglODBFREREREREREQqwYEpIiIiIiIiIiJSCQ5MERERERERERGRSnBgioiIiIiIiIiIVEJT1QEQlSTYrA701dRUVn+7nAyV1U1ERESkDFX3lwqx30REROXBK6aIiIiIiIiIiEglODBFREREREREREQqwYEpIiIiIiIiIiJSCQ5MERERERERERGRSnBgioiIiIiIiIiIVIIDU0REREREREREpBIcmCIiIiIiIiIiIpXQVHUAVD3l5+dDXb3kcUshBNTU1BQuz87OVlhGS0sLGhoalRIjERERkapUpK9UdJvs7Gxoa2uXuj8iIqK3EX/9SHLs2DF06tQJurq60NHRgb29PebNm4ecnBxpm9zcXKxZswatW7eGrq4uateujdGjRyM5OVnaJjIyEsbGxjL/DA0Noauri40bN6oiNSIiIqIKi4yMxNChQ2FqagotLS3Url0bY8eORWJiosx2J0+eRM+ePaU+UPfu3XHt2jWZbZ48eYIlS5bAwcEBurq6OHDgwOtMhYiIqNrgwBQBANavX4+ePXuiW7duSEhIQFZWFvz9/bF+/Xq8++67yMvLAwBcu3YNgYGBWL58OdLT03H27FmEhoZi0KBB0r4cHByQlZUl8+/bb7+Frq4u+vbtq6oUiYiIiMrtxo0baNOmDQAgJCQEeXl5CAoKQlxcHNq0aYOEhARp28WLF2PatGmIi4tDTEwMLC0t0bVrV5kTeatWrUJ8fDw2bNjw2nMhIiKqTtSEEELVQZBqpaamwtraGh999BGWLl0qsy4sLAzNmzfHihUrMHbsWIXld+3aBR8fH6SkpMDU1FThNo0aNULbtm2xfv16pWJKT0+HkZERjmjqQ7+US+CrUrucDJXVTUREr1fhb09aWhoMDQ1VHQ5VM56enkhPT0dISIjMLXdZWVmwt7eHl5dXsYNMT58+hampKXbu3IkBAwbIrEtNTYWJiQn27t2LPn36lCmm6tJfKsR+ExHR262q+kq8Yoqwf/9+ZGRkwNfXV26di4sLOnbsiM2bNxdbPiEhAdra2jAwMFC4/vTp07h7926xA1tERERE1dnDhw8RFBSEyZMny80DVaNGDYwbNw67du2Smf7gVYVXU5mYmFR5rERERG8aDkwR7t69C01NTdja2ipc7+DggMjISIXrEhMTMW/ePHz88cfQ1tZWuM2qVavg6OgIT0/PYmPIzs5Genq6zD8iIiKi6uDu3bsAAEdHR4XrHRwckJmZibi4OLl1eXl5mDJlCpydndG+ffsKxcH+EhERvY04MEUACp4IU9xdnfn5+QoHnVJTU9G7d280bNgQS5YsUVg2LS0N27dvL/VqqQULFsDIyEj6V79+/bInQURERFSF8vPzS1xetL8khMD48eMRFhaGHTt2QFOzYg/EZn+JiIjeRhyYIjRu3Bh5eXnFXhV1+/ZtODg4yCxLS0tD9+7doampiUOHDkFPT09h2U2bNiEvLw8fffRRiTHMmjULaWlp0r/Y2NjyJUNERERUyQqvlLp9+7bC9bdv34aBgQGsrKykZUIITJo0Cbt378aJEyfk+lLlwf4SERG9jTgwRejVqxdMTEzw888/y627dOkSgoKCMGLECGlZeno6unfvDgA4cuRIiZOerVq1CgMGDIC5uXmJMejo6MDQ0FDmHxEREVF1YGlpCW9vb/z+++/Izc2VWff8+XP8/fffGDZsGDQ0NKTln3zyCbZv347jx4/D2dm5UuJgf4mIiN5GHJgi1KxZE/7+/li/fj38/PwQExOD1NRU7N69Gz4+Phg8eDCGDBkCAHj27Bl69uyJnJwc7NmzBzo6OsjKykJWVpbcrYBhYWG4fPkyJz0nIiKiN97y5cvx9OlT9O/fH2FhYcjIyMDFixfRs2dPGBsbY8GCBdK2U6ZMwaZNm3DgwAE4ODhIfaW8vDxpm/z8fGRlZSE7OxsA8PLlS2RlZckNfBEREb3tODBFAAAfHx+cOXMG4eHhcHNzg4WFBfr3749Ro0Zh06ZNUPv/RxCfP38eV65cwa1bt2BjYwNjY2Pp3/Xr12X2+c8//8DR0RGdO3dWRUpERERElcbe3h5XrlxBvXr18O6778LS0hJubm4wMDDA5cuXpSfuZWdn4++//8aLFy/g6ekp01f666+/pP3t3r0bxsbGsLGxgY6ODoYOHQpjY2N89dVXqkqRiIhIJdREcTNe039aTk4OPDw8oKamhqCgoGKfuFdV0tPTYWRkhCOa+tD//0ExVWiXk6GyuomI6PUq/O1JS0vjLVKkFF9fX/z777+4evUqGjZs+Nrrry79pULsNxERvd2qqq/EK6ZIIW1tbWzZsgXR0dGYOXOmqsMhIiIiqnaWLFmCxo0bY9iwYcjJyVF1OERERG+kij2zlt5qtra2SEpKUnUYRERERNWStrY2zp8/r+owiIiI3mi8YoqIiIiIiIiIiFSCA1NERERERERERKQSHJgiIiIiIiIiIiKV4MAUERERERERERGpBAemiIiIiIiIiIhIJTgwRUREREREREREKqGp6gCISuKWEg9DQ0NVh0FERERUbbG/REREbzJeMUVERERERERERCrBgSkiIiIiIiIiIlIJDkwREREREREREZFKcGCKiIiIiIiIiIhUggNTRERERERERESkEhyYIiIiIiIiIiIileDAFBERERERERERqQQHpoiIiIiIiIiISCU4MEVERERERERERCrBgSkiIiIiIiIiIlIJDkwREREREREREZFKcGCKiIiIiIiIiIhUggNTRERERERERESkEhyYIiIiIiIiIiIildBUdQBEigghAADp6ekqjoSIiP4rCn9zCn+DiKo79peIiOh1qqq+EgemqFpKSUkBANSvX1/FkRAR0X9NRkYGjIyMVB0GUanYXyIiIlVISUmp1L4SB6aoWjI1NQUAPHjw4K344yA9PR3169dHbGwsDA0NVR1OhTGf6o35VG/Mp/oSQiAjIwN16tRRdShESnmb+ktv03cJwHyqO+ZTfb1NuQBvXz5paWmwtraWfn8qCwemqFpSVy+Y/szIyOit+AAXMjQ0ZD7VGPOp3phP9fa25POm/3FP/y1vY3/pbfkuKcR8qjfmU329TbkAb18+hb8/lba/St0bERERERERERGRkjgwRUREREREREREKsGBKaqWdHR0MHfuXOjo6Kg6lErBfKo35lO9MZ/q7W3Lh+hN8jZ9/t6mXADmU90xn+rrbcoFYD7KUhN8JjIREREREREREakAr5giIiIiIiIiIiKV4MAUERERERERERGpBAemiIiIiIiIiIhIJTRVHQD9d8TFxSExMRGNGjWCkZFRpZUpz34rQ1nrzcvLQ2RkJNTV1WFrawtNTdmP37179xAfHy+zTE9PDy1atKjUuItTlnzKEquqjk9sbCwePXoEBwcHGBoalrjt+fPnkZeXJ7fc0tIS9vb2AICoqCgkJCTIrNfX14erq2vlBV2ChIQEREVFoWnTpkq3ozJtUJZ2qkxlzScvLw937tyBpqYmGjZsKPf5UfXxiY+Px7179+Di4lJqO969exeJiYkyywwMDNC8eXO5bR88eICkpKTXfnzKks+5c+eQn58vt9zKygp2dnYAypYzEcm6ffs2srKy4OzsDC0trSor8zq8fPkSN2/ehI6ODho3bgw1NbVSy6SmpuL+/fuoV68eatWq9RqiVF5GRgYiIiJgbm6OBg0alKnspUuXkJ+fDzc3t6oJrhySkpIQExMDGxsb1K5dW+lyd+7cQV5entLH9HWJjo5GcnIyGjduDAMDg1K3z8vLw/3795Geng4bGxuYmZm9hiiVd/v2bTx58gQeHh5KbS+EwO3bt5GTkwNnZ2e5vpMqCSFw9epVaGlpwcXFRakyT58+RXR0NOrXrw9zc/MqjrBssrOzceXKFVhaWsLW1rZMZYODg6Guro7WrVtXUXRl9+TJE9y6dQsODg5KfxcIIXDnzh0IIeDo6Fj27wJBVMWysrLEe++9J3R1dUWTJk1EjRo1xC+//FLhMuXZb2UoT73z588XlpaWonHjxqJhw4aiTp06IiAgQGabyZMnC1NTU+Hh4SH9++CDD6owkwLlyUeZWFV1fDIzM4WPj4/Q1dUVjRs3Frq6uuL3338vsUzXrl1lcmndurUAIGbNmiVtM378eGFmZiaz3fDhw6s6HXHp0iUxcOBAUatWLQFAnDx5stQyyrRBedqpMpQnn++//15YWFiIJk2aiAYNGoi6deuKvXv3ymyjquMTHBwsBgwYIOUTGBhYapnRo0cLc3NzmVhHjhwps82LFy9Ev379hJ6ennR8/vjjj6pKQ1KefDp37iyTS6tWrQQA8dVXX0nbKJMzEcmKjo4WLi4uwtzcXDRo0EBYWFiU+p1ZnjKvy+nTp4WlpaWwsbER5ubmwtnZWdy7d6/Y7cPDw0Xfvn2FiYmJaN68uTAwMBDvvvuuSE1NfY1RF2/lypVCT09PODo6Cn19fdGzZ0/x7Nkzpcpu3LhRqKurCzMzsyqOUnnTpk0TOjo6wsnJSejo6Ihp06aVWubcuXPC0dFR1KlTR7i6ugpXV1cRERHxGqItWUZGhujevbvQ19eXjs+aNWtKLHP+/HlhZ2cn5aKrqyuGDx8ucnJyXk/QJdi8ebNwc3MTJiYmQkNDQ6kykZGRwsnJSdSqVUtYW1sLKysrERQUVMWRli4vL08sWrRI2NvbC2NjY+Hm5lZqmRs3boiePXsKU1NT4erqKvT19cXAgQNFRkbGa4i4ZCkpKeLzzz8XderUEXp6emLy5MllKr9q1Sqhrq4u6tatW0URls2dO3fEyJEjhZWVlQBQ6uem0JkzZ4SdnZ2oW7eucHV1Fa1atRJRUVFlqpsDU1TlvvzyS1GnTh0RFxcnhBBi7969AoA4e/ZshcqUZ7+VoTz1zp49Wzx+/Fh6PW/ePFGjRg1pH0IUDPb069evyuIuTnnyUSZWVR0fPz8/Ua9ePREfHy+EEGLXrl0CgLhw4YLS+9i4caMAIMLDw6Vl48ePFwMHDqz0eEuzZs0asXXrVnHv3j2lB3KUaYPKaKfyKGs+ubm5Ys6cOSIlJUVaNnfuXKGnpycSEhKkZao6PqtWrRLbt28XkZGRZRqYGjJkSInbzJgxQ1hbW4vExEQhhBDbtm0Tampq4vLly5USd3HKk09R69evF2pqaiIyMlJapkzORCSrffv2wtvbW/rDePr06cLMzEykpaVVapnXISMjQ9SqVUt89tlnQgghXr58Kbp06SLatm1bbJm9e/eKPXv2SK8fPXokGjVqJEaNGlXl8ZYmLCxMqKuri40bNwohhEhKShINGjQQvr6+pZa9e/euqFu3rpg4cWK1GZhau3at0NPTE6GhoUIIIa5evSp0dXXFunXrii1z//59UbNmTTFz5kyRl5cnhBDi5s2b4sSJE68l5pJMmDBBNGrUSOo7rFmzRmhoaIhbt24VW8bJyUkMGjRIyiU8PFzo6OiIlStXvpaYSzJ37lxx7tw5KQ9ltG7dWvTq1Uvk5uYKIQr67paWluL58+dVGWqpXrx4IaZPny4iIyPF5MmTlRqY2rVrlzhw4ID0Oj4+XjRo0EBMnDixKkNVSlhYmPjxxx9FUlKScHNzK9PA1O3bt0XdunXFuHHjqs3A1O7du8WqVavEixcvhIaGhlIDU3fu3BF6enri66+/Fvn5+UIIIa5du1bmPiQHpqjK1alTR3z55Zcyy5o3by5Gjx5doTLl2W9lqIx6k5OTBQCxe/duadnkyZNFjx49xJUrV8Tdu3elH8aqVp58lIlVVcfHwsJCfPPNNzLLnJ2dxfjx45Xeh7e3t+jQoYPMsvHjx4vevXu/9uNTKDY2VumBKWXaoDLaqSLKkk9RiYmJAoDYv3+/tEzVx+f+/ftlGpjq16+fuHz5soiKipKLNT8/X5iZmYnvv/9eZnmTJk3KfCauvMqST1FeXl6ic+fOMstKy5mIZN25c0cAEMeOHZOWJScnC01NTfHPP/9UWpnXZePGjUJDQ0MkJydLy44dOyYAiNu3byu9Hz8/P9GoUaOqCLFMPvvsM2Fvby+z7McffxRGRkbSQIAi2dnZolWrVmLt2rVi0aJF1WZgqkOHDuL999+XWfbee+8JLy+vYstMnDhRNGzYsMR8VSEnJ0cYGBiIX3/9VWa5tbW1mDlzZrHlatWqJRYvXiyzrG7duuKHH36okjjLQ9mBqbCwMAFA5gqp+Ph4oa6uLrZt21aVIZaJsgNTikybNk00bdq0kiOqmLIMTGVmZgoXFxexefNmMW/evGozMPUqZQemRo0aJRo3biwNSpUXJz+nKpWUlIT4+Hi0bNlSZnmbNm0QEhJS7jLl2W9lqKx6L126BADS/CuFjh07hpEjR6Jdu3awtrbGvn37Kh50CSqST0mxqur4xMfH49GjRxWqNzo6GidOnMDYsWPl1h0+fBgjR45E27ZtYWNjgwMHDlRK3JVJmTaojHZSpeI+P2/C8Sl04MABjBo1Cu7u7mjYsCEOHz4srYuNjUVKSorc8WndunW1Pz53797FmTNnFH5+SsqZiGQVftZf/R4wMzODra1tsd8D5SnzuoSEhKBBgwYyc/a0adNGWqesy5cvS3M/qlJISIjC39C0tDTcu3ev2HKzZs1Cw4YN8dFHH1V1iGVSXD4lHZvjx4+jV69eyM3NxdWrVxETEwMhRFWHWqrIyEg8e/ZMLp9WrVqVmM/8+fOxdOlSrF+/HkePHsXkyZNhaGiIjz/+uKpDrnSKvgusrKxQr149lX8XVJbq8l1QXtOnT0ezZs0wZMgQVYdSYcePH0fv3r2lebYePHhQru8CDkxRlXry5AkAyE0eaGZmJq0rT5ny7LcyVEa9KSkpmDx5Mnx8fPDOO+9Iy7t27Yq4uDiEhYUhISEBI0aMwKBBg3Dnzp3KS6CI8uZTWqxv8vFZvXo1jI2N8d5778ks7969Ox4+fCjlPHToUAwaNAhRUVGVE3wlqc6fn8qQnJwMX19fDB48GI6OjtLyN+X4AECvXr2kWOPj4/Hee+9h4MCBiI6OBvBmH5/Vq1fD1NQUAwYMkFleWs5EJOvJkyfQ0NCQezhEaf2nspZ5XZ48eSL3nVazZk1oaWkpHZu/vz9OnjyJ2bNnV0WIZaIon8LXxeVz8OBBbNu2DStWrKjy+MoiNzcXGRkZCvNJT09X+HAYoOAkV0pKCpo0aYLRo0ejVatWaNmyZZX2W5VR3t/QHj16oEmTJpg5cya++OILbNq0CRMnToSFhUWVxlsVnjx5Aj09PdSoUUNmeXX4LqgMy5Ytw4ULF+Dn56fqUMpl165dOHjwIJYtW6bqUCpFfHw8EhMT0bhxY4wZMwaurq5o06ZNiYP0inBgiqpU4ZNgsrKyZJZnZmZCW1u73GXKs9/KUNF609PT0atXL5iZmWHNmjUy6/r16yf9+Kmrq+P777+Hnp4edu/eXUnRyytvPqXF+qYen/z8fKxduxYffvih3I/5gAEDpKdSaGhoYMGCBdDS0sKePXsqKfrKUZ0/PxWVlpaGHj16wNLSEv7+/jLr3pTjAwA+Pj7Sk6U0NTWxcOFCqKmpYe/evQDe3OOTl5eHdevWYcSIEdDR0ZFZV1rORCRLS0sLeXl5ePnypczy0vpPZS3zumhpacl9p+Xm5iI3N1ep2AICAjBp0iT8+eefaN++fVWFqTRF+WRmZgKAwnyys7Px0UcfYezYsbh58yaCgoIQHR2N3NxcBAUFISkp6bXErYiGhgbU1dUV5qOurg4NDQ2F5bS0tLBv3z7s378fISEhePDgAUxNTVV+NVh5fkNzc3Ph7e0NCwsLxMbGIiQkBMHBwfj666/fyMEDLS0tZGdny121Uh2+Cypq69at+Oyzz+Dv7y9ddfkmefbsGUaPHo0JEyYgLCwMQUFBePDgAXJychAUFITk5GRVh1hmhd8FR48eRUhICGJjY6Grq4vRo0eXaT8cmKIqVbduXWhoaODhw4cyyx8+fAhra+tylynPfitDRerNyMhAjx49kJeXhyNHjpT6CHZ1dXWYm5vL1VWZKqsdi8aqquNTv359qKurl7veI0eOIDY2VuFtSEVpaGjAzMysSo9PeSjTBhVtJ1VIT09Ht27doKGhgUOHDqFmzZolbl9dj48impqaMDU1lWK1traGmpraG3V8gIKrAeLj4zFmzJhSty2aMxHJsrGxAVBwJvpV8fHxxX4PlKfM62JjY4P4+HiZP5QLX5cW2549e/D+++/jt99+w7hx46o6VKXY2Ngo/I4GoDCfly9fwsHBAYcPH4afnx/8/Pywf/9+PH/+HH5+frhy5cpriVsRNTU11K9fv8y/OQ0aNECHDh3QpEkTAICuri4++ugjXLx4EdnZ2VUac0kKPwdlySciIgKRkZEYP348NDU1AQCNGjVCt27dquUJrtLY2NggLy8Pjx49kpbl5+cjMTFR5d8FFbF9+3YMHz4cy5cvV/kAaHnl5OTAyckJe/bskb4LDh8+jLS0NPj5+eHatWuqDrHMGjRoAG9vbzRq1AgAoKenh+HDh+PcuXPIz89Xej8cmKIqVaNGDXh6esp8qWdmZuLo0aPo2rWrtCwqKgpXr15Vuoyy+60O+QD/G5TKycnB0aNHYWxsLLfvFy9eyLyOiorCvXv34OzsXPmJ/L/y5lNarKo6Pnp6emjXrp1Mvc+fP8exY8dk6r17967Ce+z9/f3h5uaGpk2byiwXQkhnQgtFRkYiJiamSo+Psl7NR5k2ULadVKXo8SkclAIKBg+L3qZS3Y9PZGQkQkNDARR0DIvGGh4ejtjYWCnWmjVrws3NTeb4ZGRk4MSJE9Xi+Lyaz6v8/f3h4eEBJycnmeXK5ExEstq2bQt9fX2Z74Hz588jKSlJ5nvg6tWr0i3LypZRha5duyI5ORnnz5+Xlu3evRt6enrw8PAAUPBdERQUJPPH9L59+zB48GD8/PPPmDhx4muPuzhdu3bFmTNnkJaWJi3bvXs3XF1dpVvI0tLSEBQUhOfPn8PAwABBQUEy/yZPngwjIyMEBQWhZ8+eqkoFQEE++/btkwYOhRDYu3evzPvm4cOHOHv2rPS68Bb6V8XFxcHQ0FDuqtnXydLSEs7OzjKfg5SUFJw9e1Ymn/DwcFy/fh0ApCt64+LiZPYVGxsrravuLl++jPv37wMAPD09oaOjI9MGp0+fRmpqqsq/C5RReCXh48ePpWU7duzAsGHD8Oeff75x836lpqYiKCgImZmZMDU1lfsuGDt2LGrVqoWgoCB4e3urOtxSxcbGynyXF/ddYGJiAnX1Mgw3VWjqdCIlBAUFCS0tLfH555+L3bt3i27duokGDRrIPLp4/PjxwtHRsUxllNmmOuTz8uVL4enpKczNzcXevXtFYGCg9K/wcfe5ubmiSZMmYvHixeLQoUPC399f2NraihYtWojMzMxqlY+ysarq+Jw6dUpoaWkJPz8/sXv3btGlSxdhZ2cnMjIypG1Gjx4t3nnnHZlyjx8/Ftra2sLf319un9nZ2aJJkyZiyZIl4tChQ+Lvv/8WDRo0EK1btxZZWVlVms+jR49EYGCg2LlzpwAgfv/9dxEYGChiYmKKzUeZNlBmm+qQT05OjmjXrp2oXbu22Ldvn8znJzExUQih2uOTmJgoAgMDxbZt2wQA8eeff4rAwEDx4MEDaZuPPvpINGvWTAhR8JhkJycn8fPPP4tDhw6JFStWCGtra+Hm5iays7OlMsePHxeamppi9uzZIiAgQHTu3Fk0atRIPHv2rFrl82o5TU1NsXbtWrl9KpszEclauHCh0NfXF3/99ZfYvHmzsLOzEwMGDJDZxtHRUeZpqsqUUZVBgwYJW1tbsWnTJrF8+XJhYGAg88SzjIwMAUCsXLlSCCHEiRMnhI6Ojhg+fLjMd/+5c+dUlYIkMzNTODk5iQ4dOoiAgADx9ddfCw0NDXHw4EFpm5MnTwoAIiQkROE+qtNT+aKiooSxsbEYMWKE2LNnjxg+fLgwNjYWUVFR0jaLFi2SeSJcUlKSqFu3rhg7dqw4dOiQWLp0qTA0NBQLFixQRQoy9u7dKzQ0NMS3334rdu3aJTw8PETTpk1lfnMGDhwoPDw8pNdDhgwRderUEf7+/uLQoUNi4sSJQkNDo1xPp61sERERIjAwUMyePVuKKTAwUKZPbWNjI6ZOnSq9/vbbb4WhoaH4+++/xcaNG4W1tbX44IMPVBC9vCtXrojAwEDh4+MjnJycpHwKn+r2+PFjAUB6muihQ4eElpaWGD16tMx3wYULF1SZhhBCiLy8PCkeJycn4ePjIwIDA8XVq1elbQ4ePFjiE0ir01P5UlNTpXw0NDTE7NmzRWBgoLhz5460zbx584S+vr70Oj4+XlhYWIhJkyaJw4cPi19//VXo6+uLX375pUx1qwlRDR6fQG+98+fPY9myZUhMTISzszP8/PxgZWUlrV+8eDFCQkKwYcMGpcsou42q83n27Bl69OihcD+ff/45+vXrBwBITEzE0qVLcfXqVRgbG8PT0xNjx46V7pWvLvmUJVZVHZ+zZ8/ijz/+wKNHj9C0aVP4+fnB0tJSWr9w4ULcvHkT69evl5bt2bMHP//8M/bt2wcDAwO5fSYkJGDp0qUICQmBiYkJPD09MWbMmCo/PgcPHsT8+fPlln/88cfSGSNF+ZTWBspuo+p80tLS0Lt3b4X78vPzQ58+fQCo7vjs27cPP/74o9zycePGYcSIEQCAH374AZGRkdK8cg8fPsSyZcsQGhoKExMTeHl5YfTo0dLtA4UCAwPx559/IikpCc2aNYOfn580j1Z1ygcomMhz6dKl2LdvH/T09OTKK5szEcnasGEDtmzZguzsbHTu3BnTpk2TuRpl2LBhcHV1xYwZM5Quoyo5OTn47bffcOzYMejo6GDQoEEYPny4tD4zMxNdu3aVvttXrlyJdevWye1HT08PR44ceZ2hK5ScnIyFCxciJCQEZmZmmDBhAjp16iStDwkJga+vL9auXavw6WGbN2/Gv//+W+VPYFZWeHg4Fi9ejHv37sHW1hYzZsxA48aNpfWbN2/G8uXLcerUKWnZw4cP8dNPP+HmzZuwsLDAoEGD0L9//9cfvALHjx/HihUrpKfc+vn5wdTUVFr/1VdfISUlBX/++SeAgtstV61ahRMnTiA1NRW2traYMGECmjdvrqIM/mfevHkKn2T7119/SVf5Dxo0CJ6enpgyZYq0fu3atdixYwdycnLQrVs3+Pr6Vos5poYNG4aYmBi55SdOnIC2trbU9/v666/RrVs3/PHHH9i0aZPc9sbGxir//BR+bxVla2sr9cuDg4Mxffp0bNy4UeGtlOvXr0dAQAB27txZ5fGWpvB7q6hevXpJD55Yv349/v33X5nv4QcPHuCnn35CeHg4LC0t8f7770t9dGVxYIqIiIiIiIiIiFSCc0wREREREREREZFKcGCKiIiIiIiIiIhUggNTRERERERERESkEhyYIiIiIiIiIiIileDAFBERERERERERqQQHpoiIiIiIiIiISCU4MEVERERERERERCqhqeoAiOjNderUKZiamsLFxeWtqOe/avfu3fD09ISpqamqQ1GZY8eOoU6dOnByclJ1KDIqM67U1FRcuHABqamp6N27N2rWrIl79+4hLCwMWVlZ0nY9e/aEkZFRheoKCgqCvr4+XF1dAQAHDx5EixYtYGFhUaH9EhG9aW7duoX4+Hh06dLlrajnv+r8+fMwMTFB48aNVR2KyoSFheHJkyfo2LGjqkORUZlx5ebm4ty5c0hISECrVq1gZ2eHlJQUXLx4EWlpadJ2LVu2RKNGjSpU1507d3D//n10794dAHDlyhVoa2ujadOmFdrvm4oDU0QVsGvXLmRnZwMA9PX10ahRo2r/g3XixAnUqlWrzF96isp9//33aNWqVZUPGL2ueiqqvG2rSvv378fs2bMRFhamcP3p06fRoEED2NjYKLW/s2fPIjY2Fs2bN5f7LISHhyM0NBT169eHh4eHzLoXL15gz549sLe3R6tWrYrdb1Hm5uaV0gn/8ssv0adPn2o3MFVZccXFxcHV1RVNmzZF7dq10aFDBxw4cABjx46Ft7c3dHR0kJWVhd27d+P69esVHphavHgxGjRoIA1MXb16FWvWrMHWrVsrtF8ievO8+v2tra0Na2truLq6QkNDQ8WRFe/GjRt49OgRvL29K1xu586d2LdvX5UPGL2ueiqqvG2rSomJiXj33Xdx+fJlhetv376NZ8+eoXXr1krtLzIyEleuXEGdOnXQoUMHmXVPnjzBkSNHoKOjgwEDBsiV3bdvH9TV1dGrV69i96vI4MGDoa5esZulNm7ciMuXL1e7ganKiis/Px9eXl5IS0uDs7MzzMzMIIRA69at0apVK5iZmQEADhw4gB9++KHCA1MHDhzA2rVrpYGprKwsDBw4ELdv34aurm6F9v0m4sAUUQWMHTsW9evXh6OjIzIyMnD69Gl07NgRO3fuhLa2tqrDU+i7775D+/btyzx4oqhcp06dYGtrW9khynld9VRUedtWlfz8/ODn5yf3B0JWVhY+/fRT/P333zA2NsbKlSvh4+NT6v5++eUX7NixAz179sSBAwdk1k2dOhVHjhzBwIED5Qamtm7dilGjRqFRo0a4c+eOwv2eP38enp6eMssbNWpU7Tvh1cHOnTtRp04dnDhxQlq2atUq+Pr6Yv78+QAKOt67d++ukvqnTJkCKysrXL16FS1atKiSOoioenr1+zsnJwfBwcEwNDTEvn37YGdnp+rwFNq+fTuOHTtW5sETReXeeecd6SRmVXpd9VRUedtWlb7//nu8++67Ck/SrVixAp999hlycnIwe/ZsfP3116UOuh4+fBi+vr4wMTFBfHw8atSoIbO/2bNnw8zMTG5g6v79++jXr5/0f2tra7n9fvrppxg4cKBcnQMHDqzwwNTb7saNGzh37hyePHkCExMTAMAPP/wAJycnHD16VNquqi5C8PDwgLW1NZYvX45PP/20SuqozjgwRVRBw4YNw4wZMwAUXEbt6uqKv/76C1OnTgUARERE4Pbt27CysoKrq6vMgFXhpafu7u44d+4cUlNT4ePjI92+Y2lpiZCQEGkEX1tbG0lJSQgODoapqSnatm0r8yOzf/9+NG7cWKajFxgYCAMDA7i6uuLs2bNISkrCrVu3sHnzZgBAv379cO3aNURHR0NNTQ3m5uZo1qwZzM3NpX0UV87Dw0Pu9q+srCycO3cO6enpaNasGRo2bCizvjA3KysrhISEQE1NDe7u7iWeGShaT3nbp7CchYWFVHeHDh2gpaVVrhxq164tdbDV1dXL1bZlaRMhBK5cuYL4+Hi4uLigQYMGMuvz8/Nx6dIlJCYmws7ODs7OzsW2KVBwi2R0dDTee+89ufzbtWsHFxcXuLu7w8vLC9OnT0dkZCRmzpxZ4j4BoF27djh+/Dji4uJQr149AMCDBw9w+vRptG3bVmEZf39/+Pr6YuXKlTh9+jS8vLzktnF1dZXatjR3795FZGQkevbsCQDIzMzE7t27Za7kunXrFhISEmQ6x0+fPi3xGJTWxuV5fwMFZ1sjIyNhbW0NFxcXuc5jSXGV9rk/ffo0Tp06hZcvX8q03507d1C7dm1s3rwZTk5OqF27tsLYlHlfpaen48yZMzAxMZGuknpVzZo10a9fP/z555/w9/cvsS2I6O3z6vf38+fP0bJlS8yYMQO7du0CACQlJeHixYvQ09NDixYtYGxsLJVNTEzEmTNnMGjQIFy6dAkPHjyAt7c3YmNjpT5UaGgoHj9+LPUXMjMzcfbsWeTn56Nt27aoWbOmtL9Lly4hNzdX5vcoIiIC0dHR6N69OyIjI3Hjxg0kJydLMbdt2xZ5eXm4ePEiAMDAwADvvPOOTP+guHKOjo4y9QMFv+eXL19GbGwsbGxs0LJlS5n1hf3Dtm3bIjQ0FE+ePEGrVq1Qq1atYtu4aD2v9jHL0j6F5dzc3BAaGork5GS0b99e+kO9rDm82sdt2rRpudq2rG1y9+5d3Lx5U7o6r6iS+uZFPXv2DOvWrZM72QYAEydOxPnz5zFy5EgkJCTg4sWL6Nu3r8JtizIzM4OlpSV27dqFoUOHSstXr14NLy8v3LhxQ67MqlWr0L59e2RlZWHNmjWYO3eu3DY6OjpK95XS09Nx4MABDBgwADo6OgAK+hQ1a9aUruRKTk7G8ePHMWjQIKlcdnZ2qcdAmb9/yvL+BoCEhASEhoaiZs2aaNmypVzfqqS4Svvc3759W7qq+/Dhw9I2QUFByMzMxObNm0u9Sr+091Vubi4CAwPx8uVLhe9LABg+fDh++uknDkwRUcU4OTnBwcEBoaGhyMnJwciRI3H8+HG0adMGMTExePnyJfbs2SNd+rlx40bs2rULmpqasLKyQr169eDj44Mvv/wSQggkJCTAxcUFISEhMDY2xpQpU7BgwQI0bdoUly5dQuvWrbF3716p/qlTp2LGjBkyf6AuXLgQ9vb2cHV1xaVLl5CcnIyIiAgEBAQAALp164YrV64gMDAQABAfH4/Q0FD4+/tj8ODBAFBsuaK32N24cQO9evWCvr4+6tevj7Nnz8LX1xc//vijFM+XX34JHR0dxMXFoUmTJggPD4eGhgaCg4NlOqKvKlpPedvnyy+/hLa2Nu7du4emTZvi5s2bMDIykm7BK2sOMTExaNq0KTw9PaGtrV2utlW2TR48eID+/fsjMTERrq6uiIiIwKhRozBnzhwAQGxsLN59911kZmbCwcEBISEhaNasGXbs2CF1Norat2+fwkGTGjVqYPHixfDy8oKXlxesra1x7Ngxpc+0WVlZoXv37li7di2+/PJLAAUdrd69eyMvL09u+/DwcFy4cAEbN27E06dP4e/vr3BgqiwePnyI/v37IzU1Fbq6ujhz5gyGDh2K4cOHY/369QCAefPmwdDQUBqYOnr0KNasWVPsMVCmjcvz/v7oo4+wf/9+eHh4ID4+Hnp6etizZ490O11pcZX2uT9//jwiIiKQnJwsvTcBIC0tDTdv3kRubi6EEAoHppTJOSQkBN26dYOlpSUsLCwQExMDHR0duYHTzp07S+9XIvrv0tfXR9euXbFv3z4AwJIlSzBv3jy4ubkhJycHYWFhWLt2Lfr27QsACA0NxdChQ7F27VokJSXB3t4eLVq0wMaNG7Fjxw7k5+fD3t4eCQkJePDgAX7//Xd89913sLe3R2xsLDIzM3Hx4kXppNCKFSvw7NkzmT9Q9+7di3///Rfdu3fH/fv3ER4ejpSUFOk7s169enj58qX0Oj09HYGBgRg/fjwWL14MAMWWO3XqlMwtdhkZGejTpw8iIiLQokULXL58Gc2aNcPu3buhp6cH4H/9wxo1asDCwgJpaWkIDw/HoUOHij3BU/RWvvK2z8aNG7Fz506oq6ujTp06ePLkCR48eID9+/dLdZclh1f7uAYGBuVqW2XbpLDvvXfvXrRt2xYpKSmwsrLCnj17oK6urlTfvKhTp04hLy8P7u7ucusmTJiA7777Dr/99htSUlKwYcMGXL16VeF+FPn444+xatUqaWDq1KlTUtsWHZjKy8vD2rVrsXDhQmRlZWHevHn46quvKnQVlI6ODkaNGoXatWujc+fOeP78OQYMGABLS0s8ePAAALBjxw78/PPPGDJkCICCfkGbNm1KPQbK/P1Tlvf3ihUrMGPGDGkgMz4+Hhs3bpQGeEqLq7TPfWRkJM6ePQsAMn2lqKgoPHv2DAEBAcVepa9MzqmpqfD29kZiYiJcXFwQFhYGBwcHuX117twZ48aNQ3h4eLWfHqbSCSIqNzMzM7Fo0SLp9fPnz4WJiYmYNWuW+Prrr0WbNm3E8+fPpfUTJkwQ3t7e0uuZM2cKAOLo0aMy+3VzcxP169cXKSkpQgghEhMThY6OjrCzsxOpqalCCCGio6OFhoaGOHv2rFTOzs5O/PXXXzL76t27t5g6dar02svLS8yZM6fEvLZs2SJMTU3FixcvSizn7e0tZs6cKb1u27atGDhwoMjLyxNCCBEYGCjU1NTEqVOnZHKrV6+eePz4sRBCiKysLNGgQQPx008/FRtP0XrK2z5ubm5CX19fREZGCiGEePbsmWjWrJmYNGlSmXMwNzcX8fHxMnGWt22VaZN27doJb29v6f2Ul5cn9u7dK63v0KGD+OSTT0R+fr4QQogXL14IV1dXMW/evGJj6dSpk5gyZUqJ8Xp4eMi9p0oycOBAMXDgQBEQECBsbW1Ffn6+yMvLE9bW1mL//v2iX79+YuDAgTJlZsyYIXr27CmEEOLMmTNCV1dXOo6v7tfV1VVs2rRJ5t/Vq1cVxpGVlSVq1Kghjh8/LoQQws/PT7Rq1UrUr19f2sbKykps3LhRCKHcMVCmjcv6/g4PDxcARExMjLTs3LlzIiEhQen9KfO5nzNnjvDy8pLZplmzZjLfXwkJCQKAuH79eplz/vDDD6Vtdu3aJQDI1C+EEMHBwQKAePDggcK2IKK308CBA0Xv3r1llnXv3l20bdtWHD9+XJiYmIg7d+5I6zZv3ixMTU1FWlqaEEKIgwcPCgByv68zZ84U6urq4sKFC0KIgt/FFi1aCC0tLem34eXLl8LR0VHMnz9fKjd69GgxZMgQmX0tWrRINGvWTHo9d+5c4eHhUWJeUVFRwsDAQKq/uHLz5s0Tbm5u0ms/Pz9hZ2cnkpOThRAFfZh69eqJb7/9Vi63V/sw77//vujVq1ex8RStp7ztU9g33bRpk7RszJgxwtnZWeobKZuDoj5uedtWmTaZNWuWsLS0FPfv35eW7d+/X+Tk5AghhFJ986K+/fZb4eLiUmK8c+bMkXtPlWTp0qXCzMxMJCUlCR0dHSneYcOGic8//1z88ssvwszMTKbMvn37hImJicjMzBTPnj0ThoaG4vDhw3L71dHRkesrHTp0qNhYOnbsKL766ishhBCHDh0Sjo6OombNmiIqKkoIUdDG48aNE0IodwyU/funrO9vS0tLsX79eul1XFycCA4OVnp/ynzujx49KooOj0ydOlXu+8vR0VEsXbq0TDl//vnnokmTJtL3WnR0tDAyMpKpv5Curq5Mrv8VvNGUqIJCQ0OxefNm+Pv7o2fPnhBCYPz48VizZg1cXFxw4MABbNu2DVu3bkXt2rURGBiI3NxcqXyTJk0Ujr4PGjRIun3NwsICDRs2xJAhQ6SrKGxsbGBlZaVwPp7ySEpKwvHjx7FlyxZkZ2fjyZMnuHv3rtLlY2JicP78ecycOVM6e9O+fXt4eXlhy5YtcrkVnpnT0dGBu7s7IiIiyhRvedtnwIABsLe3B1Bw1nby5MnSpbtlzcHKykqpWJVp25LaJDIyEufOncO3334rnYlUV1dHnz59AAD37t3DmTNnYGdnhx07dmDbtm3Yu3cv7OzscPLkyWLjSk5Olrs0v7L07t0bL168wKlTp3DkyBHk5eVJkzu+6uXLl1i/fj3GjRsHAPD09ESDBg2wYcMGuW0fPXqEgIAAmX/Xr19XWH9hGxbmf+rUKUyfPh2pqamIiopCREQEEhISZCbKLOkYlKWNy/L+rlGjBtTU1GQmn2/bti0sLS3Ltb/KpEzO0dHRCA4OxowZM6CmpgYA6N+/v8KzgIXvteTk5CqPnYiql/j4eGzevBn//PMPRo8ejSNHjuDzzz/H2rVr0bhxY1y7dk3qK+Xn5yM1NRXXrl2T2ceUKVPk9tu8eXO4ubkBKPhdbNOmDVq3bi1dRaGpqYnWrVtXWl8pMzMTQUFB2L59Oy5evAhLS0vpFjRlbd68GePHj5cmUrawsMCYMWPkbr9q1qwZ2rVrJ73u2LFjmb/7y9s+9erVw/vvvy+9/uKLL3Djxg3cvn27TDkU18dVRJm2La1N1q1bh08++UTmit1evXpJUzYo2zd/VVX2lWrVqoU+ffpgzZo1SE1Nxc6dO/Hxxx8r3Nbf3x8jRoxAjRo1oK+vjw8++EDhrfF5eXlyfaWS+oIdO3aU6St16dIFHh4e0rLC+XMLlXYMlG3jsr6/dXV1cevWLWkfdevWRZs2bcq9v8qkTM5bt27F+PHjYWhoCKDg75TCq9CKMjEx+U/2lXgrH1EFFd4Ko6+vjx49emDr1q0wNzdHbGwswsPDkZGRIbP9gAEDkJmZKd3PX9zgRtEfQR0dHYXLXn3Me3ktWbIEc+fORbNmzWBpaSn9gCclJSm9j5iYGACQm6Tczs5OWleo6LxU5cmjvO1T9Paihg0bIjk5GS9evChTDsoOSinbtiW1SeHl1Ir+2AcKBgeAgnmFXp0vS0NDo8SJpg0MDPD8+XOl8igrTU1NjBgxAqtXr0ZmZiZGjhypcDLQvXv3IjU1Fc+ePZM6tE2aNMGqVaswadIkmW3LMscUUNApOXr0KJ4/f46rV6/C29tb6mzl5eXBwcFB5jiWdAzK0sZleX/b2Nhg6dKlGD16NHR0dNCpUyeMHDkSnTp1Ktf+KpMyORe+NxV9rooqfK8VnWuFiN5+hScWtLS0UL9+fem24F9++QUpKSnYvn27zPaDBg2SmZ9FQ0ND4e3GyvYFnj17VuEcTp48icGDB8PS0hK2trbQ1dVFWlpamfpKQgjExsZWu75S0fYp7js9JiYGTk5OSuegbF9J2bYtqU0Kb+8qrq+Ul5endN/8VVXZVwKA0aNHY8KECTAzM0OLFi3QuHFjHDp0SGabR48eYd++fZgzZ47UDzI3N8fq1auRnJwsM3dpWeaYAgr6Sj/88IN0MnH69OmoX78+Tp48ifbt28udxCvtGCjbxmV9f69ZswaTJk3C8uXL4enpiYEDB2L48OHSiWRV9ZWUydnAwABxcXEKP1fBwcFy+3z+/Pl/sq/EgSmiCnp18vNX6evrY9CgQQrP8L2q8CqDyqCuro78/HyZZaV9KaempuLzzz/H4cOH0bVrVwAFZ4e2bNkCIYTSdRf+KD558kQ6g1b4uuhk36r09OlTudd6enrQ09MrUw7KHLfKatvCuYRSUlIUTgxZePZl7ty50jxcynBwcMD9+/eV3r6sPv74Y7i6uiI/Px8LFy5UuI2/vz9cXFykuUYAQEtLC7du3UJISEixk0Mqo2PHjpg/fz6OHDkCBwcH1KpVCx07dpTmiyjLY4XL28bKmDx5MiZOnIjr168jICAA3bp1Q0BAAHr37q1U+fJ87pWhTM6Fn5OnT59KVysWvi7q/v370NXVRf369SscGxG9WYo7sWBoaAhzc/My/SFdUeX9zpw+fTrGjRsnPckUAJydncv0e66mpgZTU1M8efJEZvmb0FcCCvp6ZclB2T5uZbSthoYGDAwMkJKSUux6Zfvmr3JwcMDKlSuV3r6sunfvjvz8fHz99df45ZdfFG6zbt061KlTB+Hh4QgPD5eWW1pa4p9//qnQRNnu7u5QV1fHkSNHcPXqVXh5eaF+/fpYunQpPD095U7ilaS8bawMLy8v3Lx5EzExMTh8+DD8/Pxw8+ZN/PTTT0qVr6q+krI5m5iYFPu5KrosLS0Njo6OFY7tTcNb+YiqSI8ePeDv7y93afDDhw+rrM66devK3CKWlpaGkJAQmW0MDAxkvoiTkpIghJD5Aix65lJRuaLs7e1haWmJnTt3SsuePn2K48ePo3379uXKpyrs378fL1++lF7v3LlTuvS3ojmUt21L4+LiAgsLC2nS7kKPHz8GAOlqrOXLl8uVjY+PL3a/3t7eOHfuXJk6fmXh6OiIyZMnY9q0aQofCR4XF4fDhw9j5cqV2Lx5s8y/bt26Vfjpbe7u7tDQ0MAPP/wgDUIVXrJe9NL00pS3jUuTkpKCrKwsqKuro1mzZpg7dy5atGih8AxacZT53JeHMjk3atQIFhYWMhOFRkdHK5z89ezZs/D09Cx2Mn4i+u/p0aMHDh06JF2hWSgxMVHuj8jKUvQ7Eyi4helVivo8iYmJMr/nN2/elBkkKK5cUe3bt5fpZwAFfYPq1Fe6ffu2zO19O3fuhLGxMZycnABULIfytq0yunXrhn///VemX5OSkiK9l8rTN+/cuTNSUlKq7LYwdXV1fPfdd+jVq5fMg3FetXr1anz66adyfaVp06Zh1apVFaq/cIqA+fPnSyfxWrVqhfT0dPj7+5eprwRUzd8/+fn5SExMBFBwpfm4ceMwYsQIXLhwQel9KPO5Ly9lcm7fvr1MXykvLw+7d++W29fZs2dRs2ZN6Rbc/xJeMUVURZYsWQJPT0+0bdsWw4cPh5qaGs6cOQM1NTVpTqPKNmLECHzyyScwMzODqakp1qxZI7dNq1atsH79ejg5OUFPTw99+vSBs7Mzhg0bhlGjRiE8PBzr1q0rtVy/fv1k1mtra2Px4sX4+OOP8eTJEzRo0AArVqxAo0aNMGrUqCrJtzyePXuGbt26YejQoQgODsbOnTtx5swZABXPobxtWxotLS2sWLECgwcPRkJCAjw8PHDjxg3Ex8dj69at0NLSwqpVqzBw4EA8fvwY3bp1Q3JyMvbu3YsPPvgAn3zyicL9Dhw4ENOmTSvzIE1ZLFq0qNh1a9asgbW1NZo3by63rn///vjss8+wZMkS1KhRA8D/5ih5la6urtx7sVBhZ+vUqVOYNWsWgIJjlJGRgYyMjDLlXN42Ls39+/cxbNgwDBo0CI0aNUJYWBiuX7+OZcuWKb0PZT735aFMztra2pg3bx4++eQTJCcnw9LSEsuWLVP4aPRt27bJPGGJiGjcuHHYvXs32rZtC19fX9SqVQuhoaE4evQobt68WaEnjhXn/fffx/z58zFhwgS0bt0aBw8eRFhYGOrUqSNt06pVK8yZMwe//vorLC0t0bZtW/Tv3x9ffvklMjIy8OLFC/z666/Q19eX2beickX98MMPcHNzw5AhQ+Dt7Y2DBw/ixo0bWLt2baXnWl6Ghobo3bs3pkyZgpSUFCxcuBALFiyAgYEBgIrlUN62VcZPP/2E9u3bo3Pnzhg8eDCSk5Oxbds2XLlyBerq6uXqm1tbW6Nbt27YtGkTvvnmmzLHpIxRo0YV288MDAxEREQE+vfvL7eusK904cIF6amBubm5Cq9A7N27d7G3h3Xs2BHffPMNfH19ARRcBeTh4YFDhw7hs88+K1MuVfH3T15eHjw9PdGtWze0aNECKSkp8Pf3L9PxUOZzX17K5Pztt9/C3d0dw4cPh6enJ7Zv346UlBTUrVtXZl9btmzB8OHDZW5l/q/gwBRRBfj4+BT7KE8bGxtcv34d69atQ2hoKAwNDfHBBx/I/LA0a9ZM4Y9E165d8c4778gs69Gjh1xdffr0kSbyBgp+2IyNjXH48GGkpaVh6dKlOH36tMxEyjNnzoSRkRGCgoLw4sULdOvWDSdPnsSyZctw+vRp1KtXDxcuXMCsWbNKLdepUyeZOQaGDRsGGxsbbN26FRcuXMCoUaMwZswYaGr+76tGUW5ubm4lnhktWk952wcApk6dikaNGiEwMBC6uro4f/68zBxB5c2hIm2rTJv069cPISEhWL9+Pc6dO4cWLVrIXL7cq1cv3LhxA//88w+CgoJQr149/PrrrzITQxalp6eH6dOn47fffit2kKZLly5ybViS0s6Wenp6SnNNPX/+HH5+fgq369u3Lw4fPozbt2/D1dUV7du3h6ampszZJqDgNsfiBqYAYOzYsbCwsICXlxeAgs7WjBkz8OjRI5lL05U5Bsq0cVnf361atcLx48exfv16nD59GlZWVrhy5QqaNGmi9P6U+dy7uLjIdXKKfmZ0dXUxZMgQ6dZRZXMeO3YsrKysEBAQgKysLPz999+4dOmSzG2nu3fvhoGBAQYMGKCwHYjo7dW+fXvk5OQoXKetrY1Dhw5h27ZtOHXqFGJiYuDq6opFixZJc9tZWVkpnCRYUR+qRYsWcn9otmnTRuYKHUdHR1y4cAHr169HSEgIhgwZgpEjRyIoKEjapmPHjvjnn39w4sQJBAcHo169evjtt9/g7OyMS5cuwdDQEFu2bMGJEydkvkcVlXvnnXeQnZ0tbdO4cWOEhobC398fQUFBcHZ2xm+//QZra+sSc2vYsGGJv3dF6ylv+wBAy5YtMW/ePOzatQvJycnYsGEDfHx8KpxDRdpWmTaxs7PD9evXsWrVKgQHB6Nhw4Y4cuSI9F5Spm+uyFdffYXBgwdj5syZ0NXVlVvv4uKicA604jg4OMi0Z1GOjo7S+vv372PKlCly8xMBBfn7+voiKioK7u7ucHBwQP/+/eX6SgDQoUOHYgem+vXrh9u3b2PQoEHSstGjR8PIyAidO3eWlilzDMr7909J728tLS1cu3YN69atw4ULF6Cvr4/NmzdL02Qosz9lPveWlpZy3zWKPjO9e/eWmctMmZxdXFwQHByMlStXIjQ0FCNGjMC0adOkk+MAkJCQgP379+PKlSsK2+Ftpyaq6h4OIqJqxt3dHX369MGXX36p6lCqjezsbEycOBELFy5UOH8VUWX47rvv0KlTJ3h6eqo6FCIiKoGfnx8uX76MY8eOqTqUauX7779Hhw4d0KFDB1WHQm+pwquoxo8fr+pQVIJXTBER/Yfp6Ohg9erVqg6D3nJff/21qkMgIiIqN57UpKr23nvvqToEleLAFBH9ZxR3Cx4RERERFX8LHhFRVeKtfEREREREREREpBKV/6gLIiIiIiIiIiIiJXBgioiIiIiIiIiIVIIDU0REREREREREpBIcmCIiIiIiIiIiIpXgwBQREREREREREakEB6aIiIiIiIiIiEglODBFREREREREREQqwYEpIiIiIiIiIiJSCQ5MERERERERERGRSvwf/1K/aLLBV8AAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 1200x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axes = plt.subplots(1, 2, figsize=(12, 5), sharex=False)\n",
    "for ax, frame, title in [\n",
    "    (axes[0], group_perm.head(10), \"Group CA\"),\n",
    "    (axes[1], inter_perm.head(10), \"Interpersonal CA\"),\n",
    "]:\n",
    "    ax.barh(frame[\"feature\"][::-1], frame[\"permutation_importance_mean\"][::-1], color=\"#C5050C\")\n",
    "    ax.set_title(title)\n",
    "    ax.set_xlabel(\"Permutation importance (\u2191 MAE when shuffled)\")\n",
    "fig.suptitle(\"Most important predictive features (raw fields)\", y=1.02)\n",
    "fig.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "3252fb44",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-07-25T15:40:25.492058Z",
     "iopub.status.busy": "2026-07-25T15:40:25.491936Z",
     "iopub.status.idle": "2026-07-25T15:40:25.500222Z",
     "shell.execute_reply": "2026-07-25T15:40:25.499842Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Encoded RF impurity importance \u2014 group CA (top 12)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>encoded_feature</th>\n",
       "      <th>importance</th>\n",
       "      <th>target</th>\n",
       "      <th>tier</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>num__LocationLatitude</td>\n",
       "      <td>0.206547</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>num__LocationLongitude</td>\n",
       "      <td>0.170181</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>num__Age</td>\n",
       "      <td>0.149166</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>cat__Q28_Never</td>\n",
       "      <td>0.080977</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>cat__Employment status_Full-Time</td>\n",
       "      <td>0.042388</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>cat__Q28_0-1 days a month</td>\n",
       "      <td>0.036864</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>cat__Q26_Never</td>\n",
       "      <td>0.029853</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>cat__Employment status_Other</td>\n",
       "      <td>0.025896</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>cat__Q28_4-8 days a month</td>\n",
       "      <td>0.022405</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>cat__Employment status_Part-Time</td>\n",
       "      <td>0.020618</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>cat__Q26_2-4 days a month</td>\n",
       "      <td>0.019124</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>cat__Q26_0-1 days a month</td>\n",
       "      <td>0.016333</td>\n",
       "      <td>gt_group_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     encoded_feature  importance       target     tier\n",
       "0              num__LocationLatitude    0.206547  gt_group_ca  transit\n",
       "1             num__LocationLongitude    0.170181  gt_group_ca  transit\n",
       "2                           num__Age    0.149166  gt_group_ca  transit\n",
       "3                     cat__Q28_Never    0.080977  gt_group_ca  transit\n",
       "4   cat__Employment status_Full-Time    0.042388  gt_group_ca  transit\n",
       "5          cat__Q28_0-1 days a month    0.036864  gt_group_ca  transit\n",
       "6                     cat__Q26_Never    0.029853  gt_group_ca  transit\n",
       "7       cat__Employment status_Other    0.025896  gt_group_ca  transit\n",
       "8          cat__Q28_4-8 days a month    0.022405  gt_group_ca  transit\n",
       "9   cat__Employment status_Part-Time    0.020618  gt_group_ca  transit\n",
       "10         cat__Q26_2-4 days a month    0.019124  gt_group_ca  transit\n",
       "11         cat__Q26_0-1 days a month    0.016333  gt_group_ca  transit"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Encoded RF impurity importance \u2014 interpersonal CA (top 12)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>encoded_feature</th>\n",
       "      <th>importance</th>\n",
       "      <th>target</th>\n",
       "      <th>tier</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>num__Age</td>\n",
       "      <td>0.188560</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>num__LocationLatitude</td>\n",
       "      <td>0.179784</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>num__LocationLongitude</td>\n",
       "      <td>0.175024</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>cat__Employment status_Full-Time</td>\n",
       "      <td>0.067052</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>cat__Q28_Never</td>\n",
       "      <td>0.041835</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>cat__Q26_Never</td>\n",
       "      <td>0.035163</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>cat__Q28_4-8 days a month</td>\n",
       "      <td>0.026819</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>cat__Q28_0-1 days a month</td>\n",
       "      <td>0.021704</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>cat__Employment status_Other</td>\n",
       "      <td>0.019836</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>cat__Q26_0-1 days a month</td>\n",
       "      <td>0.017471</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>cat__Student status_No</td>\n",
       "      <td>0.015968</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>cat__Q26_2-4 days a month</td>\n",
       "      <td>0.015005</td>\n",
       "      <td>gt_interpersonal_ca</td>\n",
       "      <td>transit</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     encoded_feature  importance               target     tier\n",
       "0                           num__Age    0.188560  gt_interpersonal_ca  transit\n",
       "1              num__LocationLatitude    0.179784  gt_interpersonal_ca  transit\n",
       "2             num__LocationLongitude    0.175024  gt_interpersonal_ca  transit\n",
       "3   cat__Employment status_Full-Time    0.067052  gt_interpersonal_ca  transit\n",
       "4                     cat__Q28_Never    0.041835  gt_interpersonal_ca  transit\n",
       "5                     cat__Q26_Never    0.035163  gt_interpersonal_ca  transit\n",
       "6          cat__Q28_4-8 days a month    0.026819  gt_interpersonal_ca  transit\n",
       "7          cat__Q28_0-1 days a month    0.021704  gt_interpersonal_ca  transit\n",
       "8       cat__Employment status_Other    0.019836  gt_interpersonal_ca  transit\n",
       "9          cat__Q26_0-1 days a month    0.017471  gt_interpersonal_ca  transit\n",
       "10            cat__Student status_No    0.015968  gt_interpersonal_ca  transit\n",
       "11         cat__Q26_2-4 days a month    0.015005  gt_interpersonal_ca  transit"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "print(\"Encoded RF impurity importance \u2014 group CA (top 12)\")\n",
    "display(importances[\"gt_group_ca__encoded_impurity\"].head(12))\n",
    "print(\"Encoded RF impurity importance \u2014 interpersonal CA (top 12)\")\n",
    "display(importances[\"gt_interpersonal_ca__encoded_impurity\"].head(12))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "054b4719",
   "metadata": {},
   "source": [
    "## 5. Predictor-space PCA\n",
    "\n",
    "PCA on the same encoded design matrix used by stage-one models \u2014 useful for seeing which dummy/numeric features dominate multivariate variance (not the same as predictive importance)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "297f6f8c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-07-25T15:40:25.501629Z",
     "iopub.status.busy": "2026-07-25T15:40:25.501524Z",
     "iopub.status.idle": "2026-07-25T15:40:25.522616Z",
     "shell.execute_reply": "2026-07-25T15:40:25.522320Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>component</th>\n",
       "      <th>variance_explained</th>\n",
       "      <th>cumulative_variance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>PC1</td>\n",
       "      <td>0.239463</td>\n",
       "      <td>0.239463</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>PC2</td>\n",
       "      <td>0.154496</td>\n",
       "      <td>0.393959</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>PC3</td>\n",
       "      <td>0.093844</td>\n",
       "      <td>0.487803</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>PC4</td>\n",
       "      <td>0.067878</td>\n",
       "      <td>0.555681</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>PC5</td>\n",
       "      <td>0.065064</td>\n",
       "      <td>0.620745</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>PC6</td>\n",
       "      <td>0.055321</td>\n",
       "      <td>0.676067</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  component  variance_explained  cumulative_variance\n",
       "0       PC1            0.239463             0.239463\n",
       "1       PC2            0.154496             0.393959\n",
       "2       PC3            0.093844             0.487803\n",
       "3       PC4            0.067878             0.555681\n",
       "4       PC5            0.065064             0.620745\n",
       "5       PC6            0.055321             0.676067"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>component</th>\n",
       "      <th>encoded_feature</th>\n",
       "      <th>loading</th>\n",
       "      <th>abs_loading</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>PC1</td>\n",
       "      <td>num__LocationLongitude</td>\n",
       "      <td>0.636019</td>\n",
       "      <td>0.636019</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>PC1</td>\n",
       "      <td>num__LocationLatitude</td>\n",
       "      <td>0.596743</td>\n",
       "      <td>0.596743</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>PC1</td>\n",
       "      <td>cat__Country of residence_United States</td>\n",
       "      <td>-0.315773</td>\n",
       "      <td>0.315773</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>PC1</td>\n",
       "      <td>cat__Country of residence_United Kingdom</td>\n",
       "      <td>0.286371</td>\n",
       "      <td>0.286371</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>PC1</td>\n",
       "      <td>num__Age</td>\n",
       "      <td>-0.095482</td>\n",
       "      <td>0.095482</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>PC2</td>\n",
       "      <td>num__Age</td>\n",
       "      <td>0.902284</td>\n",
       "      <td>0.902284</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>PC2</td>\n",
       "      <td>num__LocationLongitude</td>\n",
       "      <td>0.173629</td>\n",
       "      <td>0.173629</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>PC2</td>\n",
       "      <td>cat__Q28_Never</td>\n",
       "      <td>0.150152</td>\n",
       "      <td>0.150152</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>PC2</td>\n",
       "      <td>cat__Q27_1-2 rides in a typical day</td>\n",
       "      <td>0.129810</td>\n",
       "      <td>0.129810</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>PC2</td>\n",
       "      <td>cat__Q26_8 or more days a month</td>\n",
       "      <td>-0.126628</td>\n",
       "      <td>0.126628</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>PC3</td>\n",
       "      <td>num__LocationLatitude</td>\n",
       "      <td>0.724548</td>\n",
       "      <td>0.724548</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>PC3</td>\n",
       "      <td>num__LocationLongitude</td>\n",
       "      <td>-0.579306</td>\n",
       "      <td>0.579306</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>PC3</td>\n",
       "      <td>cat__Sex_Male</td>\n",
       "      <td>-0.188768</td>\n",
       "      <td>0.188768</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>PC3</td>\n",
       "      <td>cat__Sex_Female</td>\n",
       "      <td>0.188768</td>\n",
       "      <td>0.188768</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>PC3</td>\n",
       "      <td>num__Age</td>\n",
       "      <td>0.171392</td>\n",
       "      <td>0.171392</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>PC4</td>\n",
       "      <td>cat__Sex_Male</td>\n",
       "      <td>0.461929</td>\n",
       "      <td>0.461929</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>PC4</td>\n",
       "      <td>cat__Sex_Female</td>\n",
       "      <td>-0.461929</td>\n",
       "      <td>0.461929</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>PC4</td>\n",
       "      <td>cat__Employment status_Full-Time</td>\n",
       "      <td>0.411571</td>\n",
       "      <td>0.411571</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>PC4</td>\n",
       "      <td>cat__Employment status_Other</td>\n",
       "      <td>-0.307631</td>\n",
       "      <td>0.307631</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>PC4</td>\n",
       "      <td>cat__Q28_Never</td>\n",
       "      <td>-0.286636</td>\n",
       "      <td>0.286636</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>PC5</td>\n",
       "      <td>cat__Sex_Male</td>\n",
       "      <td>0.483570</td>\n",
       "      <td>0.483570</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>PC5</td>\n",
       "      <td>cat__Sex_Female</td>\n",
       "      <td>-0.483570</td>\n",
       "      <td>0.483570</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>PC5</td>\n",
       "      <td>cat__Employment status_Full-Time</td>\n",
       "      <td>-0.303239</td>\n",
       "      <td>0.303239</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>PC5</td>\n",
       "      <td>cat__Q26_Never</td>\n",
       "      <td>0.270645</td>\n",
       "      <td>0.270645</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>PC5</td>\n",
       "      <td>cat__Employment status_Other</td>\n",
       "      <td>0.270522</td>\n",
       "      <td>0.270522</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>PC6</td>\n",
       "      <td>cat__Q27_1-2 rides in a typical day</td>\n",
       "      <td>0.455849</td>\n",
       "      <td>0.455849</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>PC6</td>\n",
       "      <td>cat__Q27_3-4 rides in a typical day</td>\n",
       "      <td>-0.433993</td>\n",
       "      <td>0.433993</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42</th>\n",
       "      <td>PC6</td>\n",
       "      <td>cat__Employment status_Full-Time</td>\n",
       "      <td>0.324116</td>\n",
       "      <td>0.324116</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43</th>\n",
       "      <td>PC6</td>\n",
       "      <td>cat__Q26_8 or more days a month</td>\n",
       "      <td>-0.287976</td>\n",
       "      <td>0.287976</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>PC6</td>\n",
       "      <td>cat__Employment status_Other</td>\n",
       "      <td>-0.281181</td>\n",
       "      <td>0.281181</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   component                           encoded_feature   loading  abs_loading\n",
       "0        PC1                    num__LocationLongitude  0.636019     0.636019\n",
       "1        PC1                     num__LocationLatitude  0.596743     0.596743\n",
       "2        PC1   cat__Country of residence_United States -0.315773     0.315773\n",
       "3        PC1  cat__Country of residence_United Kingdom  0.286371     0.286371\n",
       "4        PC1                                  num__Age -0.095482     0.095482\n",
       "8        PC2                                  num__Age  0.902284     0.902284\n",
       "9        PC2                    num__LocationLongitude  0.173629     0.173629\n",
       "10       PC2                            cat__Q28_Never  0.150152     0.150152\n",
       "11       PC2       cat__Q27_1-2 rides in a typical day  0.129810     0.129810\n",
       "12       PC2           cat__Q26_8 or more days a month -0.126628     0.126628\n",
       "16       PC3                     num__LocationLatitude  0.724548     0.724548\n",
       "17       PC3                    num__LocationLongitude -0.579306     0.579306\n",
       "18       PC3                             cat__Sex_Male -0.188768     0.188768\n",
       "19       PC3                           cat__Sex_Female  0.188768     0.188768\n",
       "20       PC3                                  num__Age  0.171392     0.171392\n",
       "24       PC4                             cat__Sex_Male  0.461929     0.461929\n",
       "25       PC4                           cat__Sex_Female -0.461929     0.461929\n",
       "26       PC4          cat__Employment status_Full-Time  0.411571     0.411571\n",
       "27       PC4              cat__Employment status_Other -0.307631     0.307631\n",
       "28       PC4                            cat__Q28_Never -0.286636     0.286636\n",
       "32       PC5                             cat__Sex_Male  0.483570     0.483570\n",
       "33       PC5                           cat__Sex_Female -0.483570     0.483570\n",
       "34       PC5          cat__Employment status_Full-Time -0.303239     0.303239\n",
       "35       PC5                            cat__Q26_Never  0.270645     0.270645\n",
       "36       PC5              cat__Employment status_Other  0.270522     0.270522\n",
       "40       PC6       cat__Q27_1-2 rides in a typical day  0.455849     0.455849\n",
       "41       PC6       cat__Q27_3-4 rides in a typical day -0.433993     0.433993\n",
       "42       PC6          cat__Employment status_Full-Time  0.324116     0.324116\n",
       "43       PC6           cat__Q26_8 or more days a month -0.287976     0.287976\n",
       "44       PC6              cat__Employment status_Other -0.281181     0.281181"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pred_pca = predictor_pca(participants, tier=TIER)\n",
    "display(pred_pca[\"variance\"])\n",
    "display(pred_pca[\"top_loadings\"].groupby(\"component\").head(5))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c8183174",
   "metadata": {},
   "source": [
    "## 6. Persist ranked indicators\n",
    "\n",
    "Writes CSVs under `outputs/feature_importance/`, including `top_predictive_features.csv` (mean permutation importance across both CA targets)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "f1fb4e19",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-07-25T15:40:25.530407Z",
     "iopub.status.busy": "2026-07-25T15:40:25.530213Z",
     "iopub.status.idle": "2026-07-25T15:40:39.198420Z",
     "shell.execute_reply": "2026-07-25T15:40:39.197849Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "item_pca_loadings                        -> /workspace/outputs/feature_importance/ca_item_pca_loadings.csv\n",
      "item_pca_variance                        -> /workspace/outputs/feature_importance/ca_item_pca_variance.csv\n",
      "item_fa_loadings                         -> /workspace/outputs/feature_importance/ca_item_fa_loadings.csv\n",
      "item_communalities                       -> /workspace/outputs/feature_importance/ca_item_communalities.csv\n",
      "gt_group_ca__encoded_impurity            -> /workspace/outputs/feature_importance/importance_gt_group_ca__encoded_impurity.csv\n",
      "gt_group_ca__raw_permutation             -> /workspace/outputs/feature_importance/importance_gt_group_ca__raw_permutation.csv\n",
      "gt_interpersonal_ca__encoded_impurity    -> /workspace/outputs/feature_importance/importance_gt_interpersonal_ca__encoded_impurity.csv\n",
      "gt_interpersonal_ca__raw_permutation     -> /workspace/outputs/feature_importance/importance_gt_interpersonal_ca__raw_permutation.csv\n",
      "predictor_pca_variance                   -> /workspace/outputs/feature_importance/predictor_pca_variance.csv\n",
      "predictor_pca_top_loadings               -> /workspace/outputs/feature_importance/predictor_pca_top_loadings.csv\n",
      "top_predictive_features                  -> /workspace/outputs/feature_importance/top_predictive_features.csv\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>feature</th>\n",
       "      <th>permutation_importance_mean</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Q28</td>\n",
       "      <td>1.518146</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Age</td>\n",
       "      <td>1.325874</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>LocationLatitude</td>\n",
       "      <td>1.253573</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Employment status</td>\n",
       "      <td>1.074655</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>LocationLongitude</td>\n",
       "      <td>1.041415</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Q26</td>\n",
       "      <td>0.731446</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Sex</td>\n",
       "      <td>0.257538</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Student status</td>\n",
       "      <td>0.170066</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Q21</td>\n",
       "      <td>0.166650</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Country of residence</td>\n",
       "      <td>0.164674</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Q29</td>\n",
       "      <td>0.125486</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Q27</td>\n",
       "      <td>0.125306</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Q20</td>\n",
       "      <td>0.079849</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 feature  permutation_importance_mean\n",
       "0                    Q28                     1.518146\n",
       "1                    Age                     1.325874\n",
       "2       LocationLatitude                     1.253573\n",
       "3      Employment status                     1.074655\n",
       "4      LocationLongitude                     1.041415\n",
       "5                    Q26                     0.731446\n",
       "6                    Sex                     0.257538\n",
       "7         Student status                     0.170066\n",
       "8                    Q21                     0.166650\n",
       "9   Country of residence                     0.164674\n",
       "10                   Q29                     0.125486\n",
       "11                   Q27                     0.125306\n",
       "12                   Q20                     0.079849"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "paths = run_factor_and_importance_bundle(\n",
    "    PROLIFIC,\n",
    "    QUALTRICS,\n",
    "    OUT_DIR,\n",
    "    join_how=\"inner\",\n",
    "    tier=TIER,\n",
    ")\n",
    "for name, path in paths.items():\n",
    "    print(f\"{name:40s} -> {path}\")\n",
    "\n",
    "top = pd.read_csv(paths[\"top_predictive_features\"])\n",
    "top"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d19d2810",
   "metadata": {},
   "source": [
    "## 7. How to read these results against LLM / ML baselines\n",
    "\n",
    "1. Features with high **permutation importance** are the covariates a tabular learner needs most \u2014 if an LLM tier that omits them still matches band/score distance, the model is not using the sample\u2019s strongest signals the same way.\n",
    "2. **Item factor loadings** justify treating group vs interpersonal subscales as distinct evaluation targets (as the persona prompts already do).\n",
    "3. Re-run this notebook whenever the extract grows; rankings from the tiny excerpt are diagnostic, not definitive."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python (.venv psych755-jjb)",
   "language": "python",
   "name": "psych755-jjb"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.15"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}