junk_vs_real_cnn_scores.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas\n",
    "\n",
    "import seaborn\n",
    "from matplotlib import pyplot\n",
    "from IPython.display import Image, display\n",
    "\n",
    "from pathlib import Path\n",
    "\n",
    "import json\n",
    "\n",
    "from swasputils import DATA_LOCATION"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from matplotlib.font_manager import fontManager, FontProperties\n",
    "from matplotlib import rcParams\n",
    "font_path = '/Users/adam/Code/font/Serif/Palatino/Palatino Linotype.ttf'\n",
    "fontManager.addfont(font_path)\n",
    "seaborn.set_style({'font.family': FontProperties(fname=font_path).get_name()})\n",
    "\n",
    "# figure size in inches\n",
    "rcParams['figure.figsize'] = 8.3,5.53\n",
    "rcParams['figure.dpi'] = 140"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "FILTERED_REAL_SET = 91123\n",
    "FILTERING_WORKFLOW = 17313 \n",
    "DATA_LOCATION = Path(DATA_LOCATION)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Merge subject metadata with CNN scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3413060"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "zoo_subjects_cache = DATA_LOCATION / 'junk-filtering-cnn-subjects.pkl'\n",
    "if zoo_subjects_cache.exists():\n",
    "    zoo_subjects = pandas.read_pickle(zoo_subjects_cache)\n",
    "else:\n",
    "    zoo_lookup = pandas.read_csv(\n",
    "        DATA_LOCATION / 'lookup.dat',\n",
    "        delim_whitespace=True,\n",
    "        header=None,\n",
    "    )\n",
    "    zoo_lookup.columns = [\n",
    "        'subject_id',\n",
    "        'SWASP ID',\n",
    "        'Period',\n",
    "        'Period Number',\n",
    "    ]\n",
    "    zoo_subjects = pandas.read_csv(\n",
    "        DATA_LOCATION / 'superwasp-variable-stars-subjects.csv',\n",
    "    )\n",
    "    zoo_subjects = zoo_subjects[['subject_id', 'subject_set_id', 'workflow_id', 'retired_at', 'locations']]\n",
    "    zoo_subjects = zoo_subjects.merge(zoo_lookup)\n",
    "    zoo_subjects['file'] = zoo_subjects.apply(lambda r: f\"{r['SWASP ID']}_P{r['Period Number']}_fold.gif\", axis=1)\n",
    "    zoo_subjects['locations'] = zoo_subjects.apply(lambda r: list(json.loads(r['locations']).values())[0], axis=1)\n",
    "    zoo_subjects['retired_at'] = pandas.to_datetime(zoo_subjects['retired_at'])\n",
    "    zoo_subjects.rename({'locations': 'location'}, axis=1, inplace=True)\n",
    "    zoo_subjects.to_pickle(zoo_subjects_cache)\n",
    "len(zoo_subjects)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "filtered_real_subjects = zoo_subjects[zoo_subjects['subject_set_id'] == FILTERED_REAL_SET]\n",
    "retired_subjects = zoo_subjects[(zoo_subjects['workflow_id'] == FILTERING_WORKFLOW) & zoo_subjects['retired_at'].notna()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1569062"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cnn_scores = pandas.read_csv(DATA_LOCATION / 'svs-junk-predictions.csv', header=0, names=['file', 'prediction'])\n",
    "cnn_scores = cnn_scores.merge(zoo_subjects)\n",
    "cnn_scores = cnn_scores[['file', 'subject_id', 'prediction', 'location']]\n",
    "cnn_scores.drop_duplicates(inplace=True)\n",
    "len(cnn_scores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/k4/3wrsd3sj0yl87cjrql9bktm40000gn/T/ipykernel_97953/3914798590.py:2: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  cnn_scores_retired['zoo_classification_real'] = cnn_scores_retired['subject_id'].isin(filtered_real_subjects['subject_id'])\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "951984"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cnn_scores_retired = cnn_scores[cnn_scores['subject_id'].isin(retired_subjects['subject_id'])]\n",
    "cnn_scores_retired['zoo_classification_real'] = cnn_scores_retired['subject_id'].isin(filtered_real_subjects['subject_id'])\n",
    "len(cnn_scores_retired)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "617078"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cnn_scores_active = cnn_scores[~cnn_scores['subject_id'].isin(cnn_scores_retired['subject_id'])]\n",
    "len(cnn_scores_active)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plot distribution of scores vs human classifications"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[Text(0.5, 0, 'Classification'), Text(0, 0.5, 'Score')]"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1162x774.2 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "a = seaborn.boxplot(data=cnn_scores_retired, x='zoo_classification_real', y='prediction')\n",
    "a.set(xlabel='Classification', ylabel='Score')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Number of subjects')"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1162x774.2 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cnn_scores_retired[cnn_scores_retired['zoo_classification_real']]['prediction'].plot.hist()\n",
    "pyplot.xlabel('Score')\n",
    "pyplot.ylabel('Number of subjects')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Number of subjects')"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1162x774.2 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cnn_scores_retired[~cnn_scores_retired['zoo_classification_real']]['prediction'].plot.hist()\n",
    "pyplot.xlabel('Score')\n",
    "pyplot.ylabel('Number of subjects')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Count classifications"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "zoo_classifications_cache = DATA_LOCATION / 'junk-filtering-cnn-classifications.pkl'\n",
    "if zoo_classifications_cache.exists():\n",
    "    zoo_classifications = pandas.read_pickle(zoo_classifications_cache)\n",
    "else:\n",
    "    zoo_classifications = pandas.read_csv(\n",
    "        DATA_LOCATION / 'superwasp-variable-stars-classifications.csv',\n",
    "    )\n",
    "    zoo_classifications = zoo_classifications[zoo_classifications['workflow_id'] == FILTERING_WORKFLOW]\n",
    "    zoo_classifications.rename({'subject_ids': 'subject_id'}, axis=1, inplace=True)\n",
    "    zoo_classifications['created_at'] = pandas.to_datetime(zoo_classifications['created_at'])\n",
    "    zoo_classifications.to_pickle(zoo_classifications_cache)\n",
    "zoo_classifications.drop_duplicates(subset=['user_name', 'subject_id'], inplace=True)\n",
    "zoo_classifications = zoo_classifications.merge(retired_subjects[['subject_id', 'retired_at']]).drop_duplicates()\n",
    "zoo_classifications = zoo_classifications[zoo_classifications['created_at'] <= zoo_classifications['retired_at']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "classification_counts = pandas.pivot_table(\n",
    "    zoo_classifications,\n",
    "    values='classification_id',\n",
    "    aggfunc=len,\n",
    "    index='subject_id',\n",
    ")\n",
    "classification_counts.rename({'classification_id': 'classification_count'}, axis=1, inplace=True)\n",
    "classification_counts.reset_index(inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: ylabel='Frequency'>"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1162x774.2 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "classification_counts['classification_count'].plot.hist(logy=True, bins=max(classification_counts['classification_count']))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Plot sample of misclassified subjects"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>file</th>\n",
       "      <th>subject_id</th>\n",
       "      <th>prediction</th>\n",
       "      <th>location</th>\n",
       "      <th>zoo_classification_real</th>\n",
       "      <th>classification_count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1SWASPJ191851.17-181611.2_P1_fold.gif</td>\n",
       "      <td>23680962</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>https://panoptes-uploads.zooniverse.org/subjec...</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1SWASPJ193134.23-204007.3_P1_fold.gif</td>\n",
       "      <td>23727327</td>\n",
       "      <td>0.999987</td>\n",
       "      <td>https://panoptes-uploads.zooniverse.org/subjec...</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1SWASPJ192637.84-191505.6_P3_fold.gif</td>\n",
       "      <td>23701332</td>\n",
       "      <td>0.658785</td>\n",
       "      <td>https://panoptes-uploads.zooniverse.org/subjec...</td>\n",
       "      <td>True</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1SWASPJ193206.03+514729.8_P3_fold.gif</td>\n",
       "      <td>23732167</td>\n",
       "      <td>0.999991</td>\n",
       "      <td>https://panoptes-uploads.zooniverse.org/subjec...</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1SWASPJ192819.14-101839.7_P3_fold.gif</td>\n",
       "      <td>23707606</td>\n",
       "      <td>0.990684</td>\n",
       "      <td>https://panoptes-uploads.zooniverse.org/subjec...</td>\n",
       "      <td>True</td>\n",
       "      <td>2</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>163119</th>\n",
       "      <td>1SWASPJ203444.63+063744.9_P1_fold.gif</td>\n",
       "      <td>24048083</td>\n",
       "      <td>0.995545</td>\n",
       "      <td>https://panoptes-uploads.zooniverse.org/subjec...</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>163120</th>\n",
       "      <td>1SWASPJ203054.11+052207.7_P1_fold.gif</td>\n",
       "      <td>24028608</td>\n",
       "      <td>0.980465</td>\n",
       "      <td>https://panoptes-uploads.zooniverse.org/subjec...</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>163121</th>\n",
       "      <td>1SWASPJ204039.25-495734.5_P1_fold.gif</td>\n",
       "      <td>24060619</td>\n",
       "      <td>0.999885</td>\n",
       "      <td>https://panoptes-uploads.zooniverse.org/subjec...</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>163122</th>\n",
       "      <td>1SWASPJ205254.57+014844.0_P1_fold.gif</td>\n",
       "      <td>24812174</td>\n",
       "      <td>0.863901</td>\n",
       "      <td>https://panoptes-uploads.zooniverse.org/subjec...</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>163123</th>\n",
       "      <td>1SWASPJ205326.96+041833.2_P1_fold.gif</td>\n",
       "      <td>24814484</td>\n",
       "      <td>0.999159</td>\n",
       "      <td>https://panoptes-uploads.zooniverse.org/subjec...</td>\n",
       "      <td>True</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>163124 rows × 6 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                         file  subject_id  prediction  \\\n",
       "0       1SWASPJ191851.17-181611.2_P1_fold.gif    23680962    1.000000   \n",
       "1       1SWASPJ193134.23-204007.3_P1_fold.gif    23727327    0.999987   \n",
       "2       1SWASPJ192637.84-191505.6_P3_fold.gif    23701332    0.658785   \n",
       "3       1SWASPJ193206.03+514729.8_P3_fold.gif    23732167    0.999991   \n",
       "4       1SWASPJ192819.14-101839.7_P3_fold.gif    23707606    0.990684   \n",
       "...                                       ...         ...         ...   \n",
       "163119  1SWASPJ203444.63+063744.9_P1_fold.gif    24048083    0.995545   \n",
       "163120  1SWASPJ203054.11+052207.7_P1_fold.gif    24028608    0.980465   \n",
       "163121  1SWASPJ204039.25-495734.5_P1_fold.gif    24060619    0.999885   \n",
       "163122  1SWASPJ205254.57+014844.0_P1_fold.gif    24812174    0.863901   \n",
       "163123  1SWASPJ205326.96+041833.2_P1_fold.gif    24814484    0.999159   \n",
       "\n",
       "                                                 location  \\\n",
       "0       https://panoptes-uploads.zooniverse.org/subjec...   \n",
       "1       https://panoptes-uploads.zooniverse.org/subjec...   \n",
       "2       https://panoptes-uploads.zooniverse.org/subjec...   \n",
       "3       https://panoptes-uploads.zooniverse.org/subjec...   \n",
       "4       https://panoptes-uploads.zooniverse.org/subjec...   \n",
       "...                                                   ...   \n",
       "163119  https://panoptes-uploads.zooniverse.org/subjec...   \n",
       "163120  https://panoptes-uploads.zooniverse.org/subjec...   \n",
       "163121  https://panoptes-uploads.zooniverse.org/subjec...   \n",
       "163122  https://panoptes-uploads.zooniverse.org/subjec...   \n",
       "163123  https://panoptes-uploads.zooniverse.org/subjec...   \n",
       "\n",
       "        zoo_classification_real  classification_count  \n",
       "0                          True                     1  \n",
       "1                          True                     1  \n",
       "2                          True                     3  \n",
       "3                          True                     1  \n",
       "4                          True                     2  \n",
       "...                         ...                   ...  \n",
       "163119                     True                     2  \n",
       "163120                     True                     1  \n",
       "163121                     True                     1  \n",
       "163122                     True                     1  \n",
       "163123                     True                     3  \n",
       "\n",
       "[163124 rows x 6 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "misclassified_real_as_junk = cnn_scores_retired[cnn_scores_retired['zoo_classification_real'] & (cnn_scores_retired['prediction'] > 0.5)]\n",
    "misclassified_real_as_junk = misclassified_real_as_junk.merge(classification_counts)\n",
    "misclassified_real_as_junk"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification count:  1\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/5018dfbc-4ed0-4abc-8f36-598fe839978d.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/779e07b8-d533-452e-b8e8-1946fbc753ce.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/fb8a9735-3272-4693-91a1-34c37bdd8fde.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/c144752b-5969-4da6-9d64-0cdbba62bc0e.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/dac77292-a880-417b-bc47-e09580701d2e.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/69b7c605-ba82-4216-9813-d8b3bc550647.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/64add07f-d34f-4e67-96f9-e0db842849fb.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/ef7e491e-4898-49f2-8389-85d304eb3148.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/4b80efee-f2bb-476d-b01b-7acdeaff96be.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/a50baba6-548a-48f4-b16a-dd564859792d.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification count:  2\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/6e3a747c-e492-4503-8db2-1c631ac531cf.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/51d5b7f1-6126-4003-bfbe-d82a7e897dac.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/3a08387b-5d56-4139-8e73-9803e4fa3b2a.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/d70e1381-9058-48d2-a6d3-c1644ca7ce3e.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/ccd11e0d-7f62-400f-a340-0446cd614340.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/5a346749-ca72-4cfa-b871-cc46dbeec168.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/eecd1875-17f7-4ab6-a437-6998f64ab41c.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/25d78584-09c5-4b62-b74d-791849cece25.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/345323b2-5d6a-4c21-b892-22f41ec31a29.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/cb97fc46-89a4-4037-a3ae-323b8cedb32f.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification count:  3\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/ea2963e5-1b78-4991-a6b3-83c3a4711bfa.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/9513ddbc-50f7-4972-8e08-8f6943e9b6b7.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/76212abb-2e02-4c96-badd-0e5a5c3d4cfe.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/ca816d25-0254-4b49-b81b-e4ddad720a91.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/6c8a6392-b9ae-4fc7-b3fb-8dc841185fff.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/fddf5d03-a480-4848-a6c8-f955f2650801.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/73010b25-e697-4276-8ce0-5dc9fe2c447a.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/5ce3b16c-7e9f-4e04-b9b3-4c48841c2c2b.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/01fedab6-15d0-4586-84c0-c8e6c46a94de.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/49f790c0-d0a0-41a6-8419-45c1cc606e7d.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for classification_count in range(1, 4):\n",
    "    print('Classification count: ', classification_count)\n",
    "    for url in misclassified_real_as_junk[misclassified_real_as_junk['classification_count'] == classification_count].head(10)['location']:\n",
    "        display(Image(url=url))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Distribution of scores for active subjects"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: ylabel='Frequency'>"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABAMAAAKQCAYAAADqnLP5AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8hTgPZAAAACXBIWXMAABWIAAAViAHE10CgAABMYElEQVR4nO3de5RX5X0v/vfADCgJCMqACGQQFK1oYgSCmECHm1gvUZfxlhjOOdY2ahQj1aViTrSnUjj14P2YtOa0SapoMFZZBm2FCijWaUn0qJxC8Iapx/ECAbnJMMD394c/5jgZUBy+DMJ+vdZircyzP3s/n5nsNX73e/Z+dkWpVCoFAAAAKIx2e7oBAAAAoG0JAwAAAKBghAEAAABQMMIAAAAAKBhhAAAAABSMMAAAAAAKRhgAAAAABSMMAAAAgIIRBgAAAEDBCAMAAACgYIQBAAAAUDDCAAAAACiYyj3dALvmyCOPzKpVq9K/f/893QoAAABt6LXXXku3bt2ydOnST72vMGAvt2rVqmzYsGFPtwEAAEAb25VrQWHAXm7bHQHPPvvsHu4EAACAtjR8+PBW72vNAAAAACgYYQAAAAAUjDAAAAAACkYYAAAAAAUjDAAAAICCEQYAAABAwQgDAAAAoGCEAQAAAFAwwgAAAAAoGGEAAAAAFIwwAAAAAApGGAAAAAAFIwwAAACAghEGAAAAQMEIAwAAAKBghAEAAABQMMIAAAAAKBhhAAAAABSMMAAAAAAKRhgAAAAABSMMAAAAgIIRBgAAAEDBCAMAAACgYIQBAAAAUDDCAAAAACgYYQAAAAAUTOWeboBi6Hft7D3dwl5v+bRT9nQLAADAPsKdAQAAAFAwwgAAAAAoGGEAAAAAFMxeEwb87ne/y8SJE9OvX7906NAhffr0yR//8R+nvr6+Wd3s2bMzbty4jB49OkceeWQuu+yyrFmzpllNqVTK9OnTU1tbmxEjRmTQoEG59dZbW8y5Zs2aXHrppRk9enSGDBmS4cOH5/HHH29Rt3Tp0px55pkZO3ZsjjnmmJxxxhlZunRpi7qd6Q0AAAB2t71mAcH/8l/+S3r16pV77rknXbt2zSOPPJKpU6dm0aJFefHFF5MkDz30UK6//vo8/fTTqa6uzrvvvpsTTjghL7zwQp566qlUVFQkSa644orU19dn7ty5qaysTF1dXUaOHJmVK1fmpptuSpJs2rQpY8aMyQUXXJC77747STJ9+vScdtppmT17dsaPH58kee2111JbW5tZs2Zl2LBh2bx5c84555yMGDEiL774Ynr16vWpegMAAIDdba+4M2Dt2rXp3r17fvSjH2XcuHEZOnRopkyZkm9/+9t56aWXsmrVqjQ2NmbixImZNGlSqqurkyQ9evTIddddl4ULF+aJJ55IkixevDh33nlnpkyZksrKD7OQ448/Pueee26mT5+e1atXJ0nuueee1NfXZ+LEiU19TJo0Kb17986NN97YNHbNNddk9OjRGTZsWJKksrIyN998c1asWJE77rgjSXa6NwAAAGgLe0UY0Llz5/z4xz9uMd6uXbuMGjUq3bp1y/z58/PWW281XZRvM2bMmCTJvHnzkiT3339/unbtmoEDB7ao27hxY+rq6pIkM2bMyNChQ5v9xb6ioiKjRo3KokWLsn79+mzYsCGPPPJIizkHDBiQmpqapjl3tjcAAABoC3tFGJCkxW30s2bNypo1a/Lwww8nSZ577rkkSe/evZvV1dTUpKqqKq+++mpT3e/XJMnhhx+eJE11zz///A7rtmzZkuXLl+ell17K5s2bd1j30Tl3pjcAAABoC3vNmgHJh8/xDxo0KO3atcsrr7yS4447Lj/5yU8yceLEpoUEu3Tp0myfioqKHHjggVm7dm2SpL6+vkVNkhx00EFJPnwkYfXq1fnggw8+se7tt9/e7pzb6j4658709nGGDx++3fHFixfn6KOP/sT9AQAAYJu9KgzYunVr5s2blz59+mTVqlX55je/me9973tZs2ZN1q9fnySpqqpqsV+HDh2ydevWJMn69etzwAEHbLdm2xyfdKydrfvonDtTBwAAAG1hrwoD9ttvv/Tp0ydJ0q1bt9xzzz3p27dv7r333owaNSpJ0tDQkP3226/Zfg0NDencuXOSpH379mloaGhx7G1jnTt3Tvv27ZuNtbbuo3PuTG8f59lnn93u+I7uGAAAAIAd2WvWDNiePn36pLq6Ohs2bEiPHj2SJCtXrmxRt27dutTU1CT5cBX/7dVsu1W/pqYm3bt3T0VFxSfWfdyca9eubTbnzvQGAAAAbWGvCAMeffTR7Y6XSqVs3LgxI0aMSN++fZMkL7/8crOad955Jxs2bMjQoUOTJH379s0bb7yRxsbGZnWvv/56KioqMmTIkFRWVqZXr14tjrWtbuDAgenSpcsO59xW99E5d6Y3AAAAaAt7RRjw0ksv5bHHHmsxXldXly1btuQHP/hBxo0bl3bt2mXu3LnNahYuXJguXbrk5JNPTpKcdNJJaWxszIIFC1rUjRw5MoccckhTXV1dXdatW9ei7vzzz0/y4RsD+vfv32LO9957L8uWLWuq29neAAAAoC3sFWHAyJEjc8kll+Sll15qGnv33Xdz3XXX5eGHH86RRx6Zfv365eKLL85dd92VJUuWJPnwFvxp06Zl6tSpTYsGnnfeeTn22GMzefLkvP/++0mSZcuW5cEHH8z06dObjj958uR07NgxV199ddMCfw888EAaGhpy5ZVXNtVNmzYtzzzzTGbMmJHkw7sVbrjhhpx11lmpra1Nkp3uDQAAANrCXrGA4HHHHZfa2tqMHz8+PXv2TP/+/fOFL3whP/3pT5s9b3/77benZ8+eOf3001NdXZ2qqqpcfvnlmTBhQlNNVVVV5syZk6uvvjqDBw9Onz590qlTp8yaNSuDBw9uqhswYEAWLFiQq666KoMGDcrBBx+c3r17Z8GCBc0u3s8+++zMnDkzU6ZMyS233JJOnTpl5MiRue2225p9DzvTGwAAALSFilKpVNrTTdB6294msKO3DXxW9Lt29p5uYa+3fNope7oFAADgM2RXrgf3iscEAAAAgPIRBgAAAEDBCAMAAACgYIQBAAAAUDDCAAAAACgYYQAAAAAUjDAAAAAACkYYAAAAAAUjDAAAAICCEQYAAABAwQgDAAAAoGCEAQAAAFAwwgAAAAAoGGEAAAAAFIwwAAAAAApGGAAAAAAFIwwAAACAghEGAAAAQMEIAwAAAKBghAEAAABQMMIAAAAAKBhhAAAAABSMMAAAAAAKRhgAAAAABSMMAAAAgIIRBgAAAEDBCAMAAACgYIQBAAAAUDDCAAAAACgYYQAAAAAUjDAAAAAACkYYAAAAAAUjDAAAAICCEQYAAABAwQgDAAAAoGCEAQAAAFAwwgAAAAAoGGEAAAAAFIwwAAAAAApGGAAAAAAFIwwAAACAghEGAAAAQMEIAwAAAKBghAEAAABQMMIAAAAAKBhhAAAAABSMMAAAAAAKRhgAAAAABSMMAAAAgIIRBgAAAEDBCAMAAACgYIQBAAAAUDDCAAAAACgYYQAAAAAUjDAAAAAACkYYAAAAAAUjDAAAAICCEQYAAABAwQgDAAAAoGCEAQAAAFAwwgAAAAAoGGEAAAAAFIwwAAAAAApGGAAAAAAFIwwAAACAghEGAAAAQMEIAwAAAKBghAEAAABQMMIAAAAAKBhhAAAAABSMMAAAAAAKRhgAAAAABSMMAAAAgIIRBgAAAEDBCAMAAACgYIQBAAAAUDDCAAAAACgYYQAAAAAUzD4VBmzatCmNjY0plUp7uhUAAAD4zNprwoB/+7d/y2mnnZbevXunY8eOOeywwzJ58uSsW7euqWb8+PHp0KFD2rVrl4qKimb/VqxY0VRXKpUyffr01NbWZsSIERk0aFBuvfXWFnOuWbMml156aUaPHp0hQ4Zk+PDhefzxx1vULV26NGeeeWbGjh2bY445JmeccUaWLl3aom727NkZN25cRo8enSOPPDKXXXZZ1qxZU6afEAAAAOycyj3dwM468cQTc8kll2TSpEmprKzMj3/840ydOjWLFi3KnDlzknx4kX/ooYemQ4cOLfavrPx/3+oVV1yR+vr6zJ07N5WVlamrq8vIkSOzcuXK3HTTTUk+vMtgzJgxueCCC3L33XcnSaZPn57TTjsts2fPzvjx45Mkr732WmprazNr1qwMGzYsmzdvzjnnnJMRI0bkxRdfTK9evZIkDz30UK6//vo8/fTTqa6uzrvvvpsTTjghL7zwQp566qlUVFTs1p8fAABQHP2unb2nW9gnLJ92yp5uYbfZK+4MWLNmTU444YRMnTo1o0aNyogRI/LTn/40Y8eOzdy5c1NXV9dUW1dXl6VLl7b417Vr1yTJ4sWLc+edd2bKlClNAcHxxx+fc889N9OnT8/q1auTJPfcc0/q6+szceLEpmNPmjQpvXv3zo033tg0ds0112T06NEZNmxYkg9Dh5tvvjkrVqzIHXfckSRpbGzMxIkTM2nSpFRXVydJevTokeuuuy4LFy7ME088sbt+dAAAANDCXhEGNDY2ZtKkSS3GTz311CTJG2+80TTWrVu3jz3W/fffn65du2bgwIHNxseMGZONGzc2BQszZszI0KFDm/3FvqKiIqNGjcqiRYuyfv36bNiwIY888khTELDNgAEDUlNTk3nz5iVJ5s+fn7feeqtF3ZgxY5KkqQ4AAADawl4RBhx00EEZO3Zsi/ENGzYkSY444oimsR/+8If50pe+lM997nPp1atXLrnkkqxcubJp+3PPPZfevXu3ONbhhx+eJHn11VeTJM8///wO67Zs2ZLly5fnpZdeyubNm3dYt+1Yzz33XJK0qKupqUlVVVVTHQAAALSFvWbNgO159NFH87WvfS3HHnts09gLL7yQu+66K5WVlbn//vtz1113Zf78+fn1r3+dTp06pb6+Pl26dGlxrIMOOihJsnbt2qxevToffPDBJ9a9/fbbSbLDurVr1yZJ6uvrt1tXUVGRAw88sKnu4wwfPny744sXL87RRx/9ifsDAADANnttGPDYY4/lN7/5Tf71X/+1aezWW2/Nl7/85aavhw8fns9//vOZOnVqZsyYkYsuuijr16/PAQcc0OJ42xYd3Lp1a9avX58kqaqq2qW6rVu3JslO1wEAAEBb2CvDgN/+9rf53ve+l8ceeyyHHXZY0/hHg4BtLrrookydOjXPP/98kqR9+/ZpaGhoUbdtrHPnzmnfvn2zsdbWde7cuWnObWP77bffDus+zrPPPrvd8R3dMQAAAAA7slesGfBRq1atygUXXJB77723xYJ829OnT58kSadOnZJ8uIr/R9cQ2Gbbrfo1NTXp3r17KioqPrGuR48eSbLDupqamqY5d1S3bt26pjoAAABoC3tVGLBu3bpMmDAht99+e77yla8027ZixYrt7rPtAnz8+PFJkr59++aNN95IY2Njs7rXX389FRUVGTJkSCorK9OrV6+8/PLLLY73+uuvZ+DAgenSpUv69u2bJDusGzp0aNOc26t75513smHDhqY6AAAAaAt7TRiwcePG/Omf/mmmTp263ccBJk6cmC1btrQYnz17dk488cSmtxGcdNJJaWxszIIFC5rVLVy4MCNHjswhhxzSVFdXV5d169a1qDv//POTfPjGgP79+2fu3LnNat57770sW7asqW7cuHFp165di7qFCxemS5cuOfnkkz/NjwIAAAB2yV4RBjQ2Nubss8/OnDlz8o1vfCNHHnlkjjzyyBxxxBHp379/unfvniOOOCLXXnttNm7c2LRfXV1dZs2alfvvv79p7Lzzzsuxxx6byZMn5/3330+SLFu2LA8++GCmT5/eVDd58uR07NgxV199ddMCfw888EAaGhpy5ZVXNtVNmzYtzzzzTGbMmJEkKZVKueGGG3LWWWeltrY2SdKvX79cfPHFueuuu7JkyZIkH97lMG3atEydOnW7CxoCAADA7rJXLCB4ww035Je//GWSHT8OcNJJJ+W6667LoYcemqOOOiqHHHJIjjvuuPziF79Ix44dm+qqqqoyZ86cXH311Rk8eHD69OmTTp06ZdasWRk8eHBT3YABA7JgwYJcddVVGTRoUA4++OD07t07CxYsaHbxfvbZZ2fmzJmZMmVKbrnllnTq1CkjR47Mbbfd1qy/22+/PT179szpp5+e6urqVFVV5fLLL8+ECRPK+JMCAACAT1ZRKpVKe7qJT9LY2Jh27do1rcr/UaVSKVu2bEll5V6Ra5TdtrcJ7OhtA58V/a6dvadb2Ostn3bKnm4BAIC9hM/f5fFZ/wy+K9eDe8UVdFVV1Q63VVRUFDYIAAAAgNbYK9YMAAAAAMpHGAAAAAAFIwwAAACAghEGAAAAQMEIAwAAAKBghAEAAABQMMIAAAAAKBhhAAAAABSMMAAAAAAKRhgAAAAABSMMAAAAgIIRBgAAAEDBCAMAAACgYIQBAAAAUDDCAAAAACgYYQAAAAAUjDAAAAAACkYYAAAAAAUjDAAAAICCEQYAAABAwQgDAAAAoGCEAQAAAFAwwgAAAAAoGGEAAAAAFIwwAAAAAApGGAAAAAAFIwwAAACAghEGAAAAQMEIAwAAAKBghAEAAABQMMIAAAAAKBhhAAAAABSMMAAAAAAKRhgAAAAABSMMAAAAgIIRBgAAAEDBCAMAAACgYIQBAAAAUDDCAAAAACgYYQAAAAAUjDAAAAAACkYYAAAAAAUjDAAAAICCEQYAAABAwQgDAAAAoGCEAQAAAFAwwgAAAAAoGGEAAAAAFIwwAAAAAApGGAAAAAAFIwwAAACAghEGAAAAQMEIAwAAAKBghAEAAABQMMIAAAAAKBhhAAAAABSMMAAAAAAKRhgAAAAABSMMAAAAgIIRBgAAAEDBCAMAAACgYIQBAAAAUDDCAAAAACgYYQAAAAAUjDAAAAAACkYYAAAAAAUjDAAAAICCEQYAAABAwQgDAAAAoGCEAQAAAFAwwgAAAAAoGGEAAAAAFIwwAAAAAApGGAAAAAAFIwwAAACAghEGAAAAQMEIAwAAAKBghAEAAABQMHtNGPBv//ZvOe2009K7d+907Ngxhx12WCZPnpx169Y1q5s9e3bGjRuX0aNH58gjj8xll12WNWvWNKsplUqZPn16amtrM2LEiAwaNCi33npriznXrFmTSy+9NKNHj86QIUMyfPjwPP744y3qli5dmjPPPDNjx47NMccckzPOOCNLly5tUbczvQEAAMDuVrmnG9hZJ554Yi655JJMmjQplZWV+fGPf5ypU6dm0aJFmTNnTpLkoYceyvXXX5+nn3461dXVeffdd3PCCSfkhRdeyFNPPZWKiookyRVXXJH6+vrMnTs3lZWVqaury8iRI7Ny5crcdNNNSZJNmzZlzJgxueCCC3L33XcnSaZPn57TTjsts2fPzvjx45Mkr732WmprazNr1qwMGzYsmzdvzjnnnJMRI0bkxRdfTK9evT5VbwAAALC77RV3BqxZsyYnnHBCpk6dmlGjRmXEiBH56U9/mrFjx2bu3Lmpq6tLY2NjJk6cmEmTJqW6ujpJ0qNHj1x33XVZuHBhnnjiiSTJ4sWLc+edd2bKlCmprPwwCzn++ONz7rnnZvr06Vm9enWS5J577kl9fX0mTpzY1MekSZPSu3fv3HjjjU1j11xzTUaPHp1hw4YlSSorK3PzzTdnxYoVueOOO5Jkp3sDAACAtrBXhAGNjY2ZNGlSi/FTTz01SfLGG29k/vz5eeutt5ouyrcZM2ZMkmTevHlJkvvvvz9du3bNwIEDW9Rt3LgxdXV1SZIZM2Zk6NChzf5iX1FRkVGjRmXRokVZv359NmzYkEceeaTFnAMGDEhNTU3TnDvbGwAAALSFvSIMOOiggzJ27NgW4xs2bEiSHHHEEXnuueeSJL17925WU1NTk6qqqrz66qtJkueee65FTZIcfvjhSdJU9/zzz++wbsuWLVm+fHleeumlbN68eYd1H51zZ3oDAACAtrDXrBmwPY8++mi+9rWv5dhjj81PfvKTJEmXLl2a1VRUVOTAAw/M2rVrkyT19fUtapIPA4ckWbt2bVavXp0PPvjgE+vefvvt7c65re6jc+5Mbx9n+PDh2x1fvHhxjj766E/cHwAAALbZa8OAxx57LL/5zW/yr//6r0mS9evXJ0mqqqpa1Hbo0CFbt25tqjvggAO2W5MkW7du/cRj7WzdR+fcmToAAABoC3tlGPDb3/423/ve9/LYY4/lsMMOS5K0b98+SdLQ0JD99tuvWX1DQ0M6d+7cVNfQ0NDimNvGOnfu3OxYu1L30Tl3preP8+yzz253fEd3DAAAAMCO7BVrBnzUqlWrcsEFF+Tee+9ttiBfjx49kiQrV65ssc+6detSU1PTVLe9mm236tfU1KR79+6pqKj4xLqPm3Pt2rXN5tyZ3gAAAKAt7FVhwLp16zJhwoTcfvvt+cpXvtJsW9++fZMkL7/8crPxd955Jxs2bMjQoUOb6t544400NjY2q3v99ddTUVGRIUOGpLKyMr169WpxrG11AwcOTJcuXXY457a6j865M70BAABAW9hrwoCNGzfmT//0TzN16tR8+ctfbrF93LhxadeuXebOndtsfOHChenSpUtOPvnkJMlJJ52UxsbGLFiwoEXdyJEjc8ghhzTV1dXVZd26dS3qzj///CQfvjGgf//+LeZ87733smzZsqa6ne0NAAAA2sJeEQY0Njbm7LPPzpw5c/KNb3wjRx55ZI488sgcccQR6d+/f7p375527drl4osvzl133ZUlS5Yk+fBOgmnTpmXq1KlNiwaed955OfbYYzN58uS8//77SZJly5blwQcfzPTp05vmnDx5cjp27Jirr766aYG/Bx54IA0NDbnyyiub6qZNm5ZnnnkmM2bMSJKUSqXccMMNOeuss1JbW5sk6dev3071BgAAAG1hr1hA8IYbbsgvf/nLJMmKFSu2W9PQ0JDbb789PXv2zOmnn57q6upUVVXl8ssvz4QJE5rqqqqqMmfOnFx99dUZPHhw+vTpk06dOmXWrFkZPHhwU92AAQOyYMGCXHXVVRk0aFAOPvjg9O7dOwsWLGh28X722Wdn5syZmTJlSm655ZZ06tQpI0eOzG233dasv53pDQAAANpCRalUKu3pJj5JY2Nj2rVr17Qq/0eVSqVs2bIllZV7Ra5RdtveJrCjtw18VvS7dvaebmGvt3zaKXu6BQAA9hI+f5fHZ/0z+K5cD+4VV9BVVVU73FZRUVHYIAAAAABaY69YMwAAAAAoH2EAAAAAFIwwAAAAAApGGAAAAAAFIwwAAACAghEGAAAAQMEIAwAAAKBghAEAAABQMMIAAAAAKBhhAAAAABSMMAAAAAAKRhgAAAAABSMMAAAAgIIRBgAAAEDBCAMAAACgYIQBAAAAUDDCAAAAACgYYQAAAAAUjDAAAAAACkYYAAAAAAUjDAAAAICCKVsYUCqVynUoAAAAYDcqSxiwZcuWfPvb3y7HoQAAAIDdrCxhQPv27TNjxoxccMEFefXVV8txSAAAAGA3KdtjAscff3wGDx6c8847LxdccEEWL15crkMDAAAAZVRZrgNde+21+frXv54rr7wyzz33XO6+++787ne/y3e+852MGjWqXNMAAAAAu6hsdwZ8/etfb/rfxx13XO6+++7cc889qaury+DBg/PTn/40mzZtKtd0AAAAQCuVLQz493//92Zf19fX53/8j/+R22+/Pc8//3yuuuqqnHXWWfnzP//zvPfee+WaFgAAAPiUyhIGbN26Nd/5zneyZcuWzJ49O2effXb69euXv/iLv8igQYPy85//PG+99VYeffTRnHLKKfnud7+byZMnp7GxsRzTAwAAAJ9CWcKAdu3a5V/+5V/SrVu3fP3rX8+TTz6ZSy+9NEuWLMk///M/5+yzz05VVVWSZMiQIZk5c2Y2btyYK664ohzTAwAAAJ9C2R4TKJVKqampyV//9V/nzTffzK233pojjjhih/XHH3987rvvvnJNDwAAAOyksr1N4Nxzz83999+/0/Vvv/12jj322HJNDwAAAOykst0ZcOmll+aVV17Z7hsD1qxZk82bNzcbmzhxYhYsWFCu6QEAAICdVLYw4M/+7M9y7rnn5vbbb2+xbe3atfnBD35QrqkAAACAXVC2xwSWL1+eJUuW5KCDDmqxrXfv3nnttdfKNRUAAACwC8p2Z8DAgQO3GwRsU19fX66pAAAAgF1QtjDgsMMO2+G25cuX5//8n/9TrqkAAACAXVC2MOCP/uiP8vjjj7cYf/PNN3P66afni1/8YrmmAgAAAHZBWV8tOHz48Nx111350pe+lHXr1mXJkiV5+umns3nz5u0GBQAAAEDbK1sYkCQPP/xwJkyYkGnTpjWNHXzwwbnlllsybty4ck4FAAAAtFJZw4CDDz44TzzxRN57770sX748+++/f4466qi0a1e2pxEAAACAXbRbrtKrq6szdOjQHH300U1BwIQJE3bHVAAAAMCnVNY7A1auXJmnnnoq69atS6lUahrfsGGDNQMAAADgM6JsYcD8+fNz+umntwgCtqmoqCjXVAAAAMAuKFsY8P3vfz+DBg3Kd7/73fTq1Svt27dPkmzevDmbNm3KmjVryjUVAAAAsAvKFga88soree2119KpU6dyHRIAAADYDcq2gOBXvvIVQQAAAADsBcoWBnzxi1/M+vXrd7j9nnvuKddUAAAAwC4oWxgwceLETJ48ebtrA2zatCk/+tGPyjUVAAAAsAvKtmbAEUcckTVr1uSuu+4q1yEBAACA3aBsYcDIkSPTtWvXtGvX8maDhoaGzJo1q1xTAQAAALugbGHAWWedlQkTJuxw+6hRo8o1FQAAALALyrZmwJlnnvmx22fMmFGuqQAAAIBdULYw4HOf+1yuv/76HHLIIfna177WbNtDDz2Ujh07lmsqAAAAYBeULQy49dZbM3Xq1Oy3335ZvXp1s21f/vKXc+utt5ZrKgAAAGAXlC0M+PGPf5xHH300r732Wqqrq5tt69+/f/793/+9XFMBAAAAu6BsYcDnP//5nHLKKUmSioqKFtvffvvtck0FAAAA7IKyhQG9evXa4bYlS5bktddeK9dUAAAAwC4oWxgwZMiQPPnkk0mSUqnUNL5o0aKcfvrpLRYVBAAAAPaMynIdaNKkSfniF7+YQw45JEuXLs3JJ5+cV155Ja+++mo6duyYX/ziF+WaCgAAANgFZV0z4J/+6Z/StWvXbNy4Mf/4j/+YV155JcOHD8+TTz6ZL37xi+WaCgAAANgFZbszIEkOP/zw/PKXv0ySvPfee+nSpUs6duxYzikAAACAXVS2OwN+X3V1dbMg4NJLL91dUwEAAACfQlnuDFi1alWeffbZ7LfffqmsbHnIhoaGzJ49uxxTAQAAALuoLGFAqVTK6aefnq1bt5bjcAAAAMBuVJYw4MADD8yXvvSlfPvb307Xrl2bxkulUpYvX54nn3wyZ511VjmmAgAAAHZR2RYQHDp0aK644oodbjvooIPKNRUAAACwC8q2gOAPf/jDHW475ZRTcu+995ZrKgAAAGAX7La3Cfy+RYsWtdVUAAAAwMcoy2MCW7duzZo1a/L5z3++xdsE1q5dm4ceeigvv/xyOaYCAAAAdlFZwoA33ngjhx122MfWfOtb3yrHVAAAAMAuKksYcOihh+awww7LMccck86dOzeNV1RUZP/998+gQYNy4YUXlmMqAAAAYBeV7W0CX/rSlzJz5sxyHQ4AAADYTcq2gOCtt976qff527/923JNDwAAAOyksoUBixcvzurVq3e6vrGxMf/zf/7Pck0PAAAA7KSyPSbw6KOP5uSTTy7X4QAAAIDdpGxhwKZNm/LFL34xtbW1zcZLpVIeeeSRjBo1qtl4Y2NjHn300XJNDwAAAOyksoUBc+bMyZw5c7b7isH//J//c1avXt0iKDj11FNbPd+WLVvSvn37ZmObNm1KRUVFKisrU1FR0epjAwAAwL6sbGsGHHLIIdsNApLk2GOPzeOPP95i/Gc/+9mnnufNN9/M97///fTr16/FtvHjx6dDhw5p165dKioqmv1bsWJFU12pVMr06dNTW1ubESNGZNCgQdtdAHHNmjW59NJLM3r06AwZMiTDhw/f7vexdOnSnHnmmRk7dmyOOeaYnHHGGVm6dGmLutmzZ2fcuHEZPXp0jjzyyFx22WVZs2bNp/4ZAAAAwK4o250BPXv2/NjtjzzySP77f//vzcYOPPDAnT7+gw8+mL//+7/P/vvvn9mzZ6djx44takqlUg499NB06NChxbbKyv/3rV5xxRWpr6/P3LlzU1lZmbq6uowcOTIrV67MTTfdlOTDuwzGjBmTCy64IHfffXeSZPr06TnttNMye/bsjB8/Pkny2muvpba2NrNmzcqwYcOyefPmnHPOORkxYkRefPHF9OrVK0ny0EMP5frrr8/TTz+d6urqvPvuuznhhBPywgsv5KmnnnInAwAAAG2mbHcGnHTSSZkxY8Z2t/3N3/xN3n///V06/u9+97vceeed+fnPf57u3bunU6dO262rq6vL0qVLW/zr2rVrkg/fenDnnXdmypQpTQHB8ccfn3PPPTfTp09veiPCPffck/r6+kycOLHp2JMmTUrv3r1z4403No1dc801GT16dIYNG5bkw9Dh5ptvzooVK3LHHXck+XB9hIkTJ2bSpEmprq5OkvTo0SPXXXddFi5cmCeeeGKXfjYAAADwaZQtDPjjP/7jTJ06Nd/+9rdz3333Ze7cufm7v/u7nHrqqbnkkkty/vnn79Lxv/Od76Smpqbp6/3222+7dd26dfvY49x///3p2rVrBg4c2Gx8zJgx2bhxY+rq6pIkM2bMyNChQ5v9xb6ioiKjRo3KokWLsn79+mzYsCGPPPJIUxCwzYABA1JTU5N58+YlSebPn5+33nqrRd2YMWOSpKkOAAAA2kLZHhNo3759fv7zn+eUU07Jfffdl4qKipRKpVRUVORb3/pWpk2bVq6pkmS7jwIkyQ9/+MP8r//1v/LKK6+kS5cuOeOMM3LTTTfloIMOSpI899xz6d27d4v9Dj/88CTJq6++miR5/vnnc+GFF263bsuWLVm+fHnWrVuXzZs37/B4//t//++mOZO0qKupqUlVVVXTnAAAANAWyhYGJMlRRx2V3/zmN3niiSeyZMmSdOvWLcOHD8+gQYPKOU2S7PAZ+xdeeCF33XVXKisrc//99+euu+7K/Pnz8+tf/zqdOnVKfX19unTp0mK/bWHB2rVrs3r16nzwwQefWPf2228nyQ7r1q5dmySpr6/fbl1FRUUOPPDAprqPM3z48O2OL168OEcfffQn7g8AAADblDUMSD585d9Xv/rVXXptYGvdeuut+fKXv9z09fDhw/P5z38+U6dOzYwZM3LRRRdl/fr1OeCAA1rsu+1Og61bt2b9+vVJkqqqql2q27p1a5LsdB0AAAC0hbKtGbBx48Z861vfygEHHJBx48Y12/aTn/wkb775Zrmm2qGPBgHbXHTRRUk+vO0/+fBxhoaGhhZ128Y6d+6c9u3bNxtrbV3nzp2b5tyZuo/z7LPPbvefuwIAAAD4tMoWBtx00035h3/4h3zta1/Lpk2bmm37oz/6o9xyyy3lmupT6dOnT5I0vX2gR48eWblyZYu6bbfq19TUpHv37qmoqPjEuh49eiTJDuu2LXj4cXXr1q1rtjAiAAAA7G5lCwMeeOCBPPPMM3nyySebnqvfpmfPnnnjjTfKNVWSZPPmzc2+XrFixXbrtl2Ajx8/PknSt2/fvPHGG2lsbGxW9/rrr6eioiJDhgxJZWVlevXqlZdffrnF8V5//fUMHDgwXbp0Sd++fZNkh3VDhw5tmnN7de+88042bNjQVAcAAABtoWxhQHV1dY477rgk21/cr5xhwNatW1uEARMnTsyWLVta1M6ePTsnnnhixo4dmyQ56aST0tjYmAULFjSrW7hwYUaOHJlDDjmkqa6uri7r1q1rUbftNYmHH354+vfvn7lz5zaree+997Js2bKmunHjxqVdu3Yt6hYuXJguXbrk5JNP/rQ/AgAAAGi1soYBO/Iv//IvTSvv76rNmzdn1apVef/991MqlZrGv/rVr+baa6/Nxo0bm8bq6uoya9as3H///U1j5513Xo499thMnjw577//fpJk2bJlefDBBzN9+vSmusmTJ6djx465+uqrmxb4e+CBB9LQ0JArr7yyqW7atGl55plnMmPGjCRJqVTKDTfckLPOOiu1tbVJkn79+uXiiy/OXXfdlSVLliT58PGAadOmZerUqdtd0BAAAAB2l7K9TWDs2LH52c9+lgkTJjRdPG/ZsiX/8A//kMsuu2yX//r95ptv5sILL8xvf/vbrFu3LuvWrcsxxxyT7t2757HHHst5552Xs88+O4ceemiOOuqoHHLIITnuuOPyi1/8Ih07dmw6TlVVVebMmZOrr746gwcPTp8+fdKpU6fMmjUrgwcPbqobMGBAFixYkKuuuiqDBg3KwQcfnN69e2fBggXNLt7PPvvszJw5M1OmTMktt9ySTp06ZeTIkbntttua9X/77benZ8+eOf3001NdXZ2qqqpcfvnlmTBhwi79XAAAAODTqih99M/ru2DTpk0ZOnRoVq5cmffff7/p2fyNGzemW7du+fWvf22hvN1g+PDhST5828BnWb9rZ+/pFvZ6y6edsqdbAABgL+Hzd3l81j+D78r1YNkeE+jQoUPmzp2bM844oykI6NKlS84///z867/+qyAAAAAAPiPK9pjAhg0bUl1dnbvuuqtchwQAAAB2g7LdGdCvX7906NAhN954Y7kOCQAAAOwGZQsDampq0rVr1x2GAY8++mi5pgIAAAB2QdnCgHvvvTc9evRIfX39drffeuut5ZoKAAAA2AVlWzNg5syZGTt2bEaMGJFTTz013bp1S0VFRZIP3zTwwgsvlGsqAAAAYBeULQxYvHhxHnzwwSTJHXfc0WL7tmAAAAAA2LPKFgaMGTMmPXr0yFVXXdViW2NjY77xjW+UayoAAABgF5QtDBgxYkR69uyZmpqa7W6fMGFCuaYCAAAAdkGrw4CxY8fmiCOOyCGHHJLrr78+f/AHf5A/+IM/2GH9pEmTWjsVAAAAUEatDgPmzZuX//bf/luOOuqoJB++OvAnP/lJVq1alW7dumXUqFG57LLLytYoAAAAUB6tDgO+8IUv5IQTTmj6+rTTTsuJJ56Y6urqrFy5MlVVVWVpEAAAACivdq3dsVu3bi3GOnbsmMMPP1wQAAAAAJ9hrQ4DKiu3f1PBAQccsN3x8ePHt3YqAAAAoIxa/ZjAe++9l//7f/9v9t9//2bBwObNm7N27dqUSqWmsffffz8vv/zyrnUKAAAAlEWrw4A33ngjX/jCF7a7rWvXrq09LAAAALCbtToMqKqqynHHHZf99tvvE2tXr16dF198sbVTAQAAAGXU6jDgnHPOyd///d/vdP2QIUNaOxUAAABQRq1eQPCkk076VPXf/OY3WzsVAAAAUEatDgO+9a1vfar6SZMmtXYqAAAAoIxaHQYAAAAAeydhAAAAABSMMAAAAAAKRhgAAAAABSMMAAAAgIIRBgAAAEDBCAMAAACgYIQBAAAAUDDCAAAAACgYYQAAAAAUjDAAAAAACkYYAAAAAAUjDAAAAICCEQYAAABAwQgDAAAAoGCEAQAAAFAwwgAAAAAoGGEAAAAAFIwwAAAAAApGGAAAAAAFIwwAAACAghEGAAAAQMEIAwAAAKBghAEAAABQMMIAAAAAKBhhAAAAABSMMAAAAAAKRhgAAAAABSMMAAAAgIIRBgAAAEDBCAMAAACgYIQBAAAAUDDCAAAAACgYYQAAAAAUjDAAAAAACkYYAAAAAAUjDAAAAICCEQYAAABAwQgDAAAAoGCEAQAAAFAwwgAAAAAoGGEAAAAAFIwwAAAAAApGGAAAAAAFIwwAAACAghEGAAAAQMEIAwAAAKBghAEAAABQMMIAAAAAKBhhAAAAABSMMAAAAAAKRhgAAAAABSMMAAAAgIIRBgAAAEDBCAMAAACgYIQBAAAAUDDCAAAAACgYYQAAAAAUzF4bBmzZsmVPtwAAAAB7pb0uDHjzzTfz/e9/P/369dvu9tmzZ2fcuHEZPXp0jjzyyFx22WVZs2ZNs5pSqZTp06entrY2I0aMyKBBg3Lrrbe2ONaaNWty6aWXZvTo0RkyZEiGDx+exx9/vEXd0qVLc+aZZ2bs2LE55phjcsYZZ2Tp0qWt6g0AAAB2t8o93cDOevDBB/P3f//32X///TN79ux07NixRc1DDz2U66+/Pk8//XSqq6vz7rvv5oQTTsgLL7yQp556KhUVFUmSK664IvX19Zk7d24qKytTV1eXkSNHZuXKlbnpppuSJJs2bcqYMWNywQUX5O67706STJ8+Paeddlpmz56d8ePHJ0lee+211NbWZtasWRk2bFg2b96cc845JyNGjMiLL76YXr16fareAAAAYHfba+4M+N3vfpc777wzP//5z9O9e/d06tSp2fbGxsZMnDgxkyZNSnV1dZKkR48eue6667Jw4cI88cQTSZLFixfnzjvvzJQpU1JZ+WEWcvzxx+fcc8/N9OnTs3r16iTJPffck/r6+kycOLFpjkmTJqV379658cYbm8auueaajB49OsOGDUuSVFZW5uabb86KFStyxx13fKreAAAAoC3sNWHAd77zndTU1DR9vd9++zXbPn/+/Lz11ltNF+XbjBkzJkkyb968JMn999+frl27ZuDAgS3qNm7cmLq6uiTJjBkzMnTo0GZ/sa+oqMioUaOyaNGirF+/Phs2bMgjjzzSYs4BAwakpqamac6d7Q0AAADawl4TBvy+Dh06NPv6ueeeS5L07t272XhNTU2qqqry6quvNtX9fk2SHH744UnSVPf888/vsG7Lli1Zvnx5XnrppWzevHmHdR+dc2d6AwAAgLaw16wZ8Pt+/xn7+vr6JEmXLl1a1B144IFZu3ZtU93v1yTJQQcdlCRZu3ZtVq9enQ8++OAT695+++3tzrmt7qNz7kxvH2f48OHbHV+8eHGOPvroT9wfAAAAttlr7wz4fevXr0+SVFVVtdjWoUOHbN26taluRzVJsnXr1k881s7WfXTOnakDAACAtrDX3hnw+9q3b58kaWhoaLGeQENDQzp37txU19DQ0GL/bWOdO3dudqxdqfvonDvT28d59tlntzu+ozsGAAAAYEf2mTsDevTokSRZuXJli23r1q1rWnywR48e263Zdqt+TU1NunfvnoqKik+s+7g5165d22zOnekNAAAA2sJeGwZs3ry52dd9+/ZNkrz88svNxt95551s2LAhQ4cObap744030tjY2Kzu9ddfT0VFRYYMGZLKysr06tWrxbG21Q0cODBdunTZ4Zzb6j465870BgAAAG1hrwwDtm7d2iIMGDduXNq1a5e5c+c2G1+4cGG6dOmSk08+OUly0kknpbGxMQsWLGhRN3LkyBxyyCFNdXV1dVm3bl2LuvPPPz/Jh28M6N+/f4s533vvvSxbtqypbmd7AwAAgLaw14UBmzdvzqpVq/L++++nVCo1jffr1y8XX3xx7rrrrixZsiTJh7fgT5s2LVOnTs0BBxyQJDnvvPNy7LHHZvLkyXn//feTJMuWLcuDDz6Y6dOnNx1v8uTJ6dixY66++uqmBf4eeOCBNDQ05Morr2yqmzZtWp555pnMmDEjSVIqlXLDDTfkrLPOSm1t7afqDQAAANrCXrOA4JtvvpkLL7wwv/3tb7Nu3bqsW7cuxxxzTLp3757HHnssnTp1yu23356ePXvm9NNPT3V1daqqqnL55ZdnwoQJTcepqqrKnDlzcvXVV2fw4MHp06dPOnXqlFmzZmXw4MFNdQMGDMiCBQty1VVXZdCgQTn44IPTu3fvLFiwoNnF+9lnn52ZM2dmypQpueWWW9KpU6eMHDkyt912W7P+d6Y3AAAAaAsVpY/+eZ29zra3CezobQOfFf2unb2nW9jrLZ92yp5uAQCAvYTP3+XxWf8MvivXg3vdYwIAAADArhEGAAAAQMEIAwAAAKBghAEAAABQMMIAAAAAKBhhAAAAABSMMAAAAAAKRhgAAAAABSMMAAAAgIIRBgAAAEDBCAMAAACgYIQBAAAAUDDCAAAAACgYYQAAAAAUjDAAAAAACkYYAAAAAAUjDAAAAICCEQYAAABAwQgDAAAAoGCEAQAAAFAwwgAAAAAoGGEAAAAAFIwwAAAAAApGGAAAAAAFIwwAAACAghEGAAAAQMEIAwAAAKBghAEAAABQMMIAAAAAKBhhAAAAABSMMAAAAAAKRhgAAAAABSMMAAAAgIIRBgAAAEDBCAMAAACgYIQBAAAAUDDCAAAAACgYYQAAAAAUjDAAAAAACkYYAAAAAAUjDAAAAICCEQYAAABAwQgDAAAAoGCEAQAAAFAwwgAAAAAoGGEAAAAAFIwwAAAAAApGGAAAAAAFIwwAAACAghEGAAAAQMEIAwAAAKBghAEAAABQMMIAAAAAKBhhAAAAABSMMAAAAAAKRhgAAAAABSMMAAAAgIIRBgAAAEDBCAMAAACgYIQBAAAAUDDCAAAAACgYYQAAAAAUjDAAAAAACkYYAAAAAAUjDAAAAICCEQYAAABAwQgDAAAAoGCEAQAAAFAwwgAAAAAoGGEAAAAAFIwwAAAAAApGGAAAAAAFIwwAAACAghEGAAAAQMEIAwAAAKBghAEAAABQMMIAAAAAKBhhAAAAABTMPhcGbN68OZs2bcrWrVv3dCsAAADwmbTPhQFTpkxJx44d0759+1RUVDT794tf/KKpbvbs2Rk3blxGjx6dI488MpdddlnWrFnT7FilUinTp09PbW1tRowYkUGDBuXWW29tMeeaNWty6aWXZvTo0RkyZEiGDx+exx9/vEXd0qVLc+aZZ2bs2LE55phjcsYZZ2Tp0qXl/yEAAADAx6jc0w2UW6lUSs+ePdO1a9cW27p06ZIkeeihh3L99dfn6aefTnV1dd59992ccMIJeeGFF/LUU0+loqIiSXLFFVekvr4+c+fOTWVlZerq6jJy5MisXLkyN910U5Jk06ZNGTNmTC644ILcfffdSZLp06fntNNOy+zZszN+/PgkyWuvvZba2trMmjUrw4YNy+bNm3POOedkxIgRefHFF9OrV682+OkAAADAPnhnQJLccccdWbp0aYt/J554YhobGzNx4sRMmjQp1dXVSZIePXrkuuuuy8KFC/PEE08kSRYvXpw777wzU6ZMSWXlh5nJ8ccfn3PPPTfTp0/P6tWrkyT33HNP6uvrM3HixKb5J02alN69e+fGG29sGrvmmmsyevToDBs2LElSWVmZm2++OStWrMgdd9zRBj8VAAAA+NA+GQZ069Zth9vmz5+ft956q+mifJsxY8YkSebNm5ckuf/++9O1a9cMHDiwRd3GjRtTV1eXJJkxY0aGDh3adDdBklRUVGTUqFFZtGhR1q9fnw0bNuSRRx5pMeeAAQNSU1PTNCcAAAC0hX0yDJg3b15qa2vTrVu3dO3aNV//+tfz4osvJkmee+65JEnv3r2b7VNTU5Oqqqq8+uqrTXW/X5Mkhx9+eJI01T3//PM7rNuyZUuWL1+el156KZs3b95h3bZjAQAAQFvY59YMSJKnn3461113XQ4++OAsWLAgN954Y44//vgsWrQo9fX1Sf7f+gHbVFRU5MADD8zatWuTJPX19S1qkuSggw5KkqxduzarV6/OBx988Il1b7/99nbn3Fa3bc6PM3z48O2OL168OEcfffQn7g8AAADb7HNhwLnnnptrr702++23X5LkuOOOy6BBgzJ+/PhMnTo1+++/f5Kkqqqqxb4dOnRoeiXh+vXrc8ABB2y3Jkm2bt2a9evXf+yxdrbOaxABAABoS/tcGPAHf/AHLcZOPPHE9O3bN88//3xGjBiRJGloaGgKDLZpaGhI586dkyTt27dPQ0NDi2NtG+vcuXPat2/fbKy1ddvm/DjPPvvsdsd3dMcAAAAA7Mg+uWbA9vTp0yedOnVKjx49kiQrV65sUbNu3brU1NQk+fANA9ur2XZLf01NTbp3756KiopPrPu4OdeuXds0JwAAALSFfSoMWLFixQ63rVy5MuPHj0/fvn2TJC+//HKz7e+88042bNiQoUOHJkn69u2bN954I42Njc3qXn/99VRUVGTIkCGprKxMr169WhxrW93AgQPTpUuXHc65rW7bnAAAANAW9qkw4Oabb87y5ctbjL/88stZu3ZtJk2alHHjxqVdu3aZO3dus5qFCxemS5cuOfnkk5MkJ510UhobG7NgwYIWdSNHjswhhxzSVFdXV5d169a1qDv//POTfPjGgP79+7eY87333suyZcua6gAAAKAt7FNhwIQJE3LRRRflnXfeaRp76623ctVVV+WXv/xlDjzwwPTr1y8XX3xx7rrrrixZsiTJh48HTJs2LVOnTm1aNPC8887Lsccem8mTJ+f9999PkixbtiwPPvhgpk+f3nT8yZMnp2PHjrn66qubFgJ84IEH0tDQkCuvvLKpbtq0aXnmmWcyY8aMJEmpVMoNN9yQs846K7W1tbv15wIAAAAftU8tIDho0KAMGDAgX/rSl3LooYempqYm/fr1y9/8zd+kZ8+eTXW33357evbsmdNPPz3V1dWpqqrK5ZdfngkTJjTVVFVVZc6cObn66qszePDgpjUHZs2alcGDBzfVDRgwIAsWLMhVV12VQYMG5eCDD07v3r2zYMGCZm8jOPvsszNz5sxMmTIlt9xySzp16pSRI0fmtttua5OfDQAAAGxTUSqVSnu6CVpv29sEdvS2gc+KftfO3tMt7PWWTztlT7cAAMBewufv8visfwbflevBfeoxAQAAAOCTCQMAAACgYIQBAAAAUDDCAAAAACgYYQAAAAAUjDAAAAAACkYYAAAAAAUjDAAAAICCEQYAAABAwQgDAAAAoGCEAQAAAFAwwgAAAAAoGGEAAAAAFIwwAAAAAApGGAAAAAAFIwwAAACAghEGAAAAQMEIAwAAAKBghAEAAABQMMIAAAAAKBhhAAAAABSMMAAAAAAKRhgAAAAABSMMAAAAgIIRBgAAAEDBCAMAAACgYIQBAAAAUDDCAAAAACgYYQAAAAAUjDAAAAAACkYYAAAAAAUjDAAAAICCEQYAAABAwQgDAAAAoGCEAQAAAFAwwgAAAAAoGGEAAAAAFIwwAAAAAApGGAAAAAAFIwwAAACAghEGAAAAQMEIAwAAAKBghAEAAABQMMIAAAAAKBhhAAAAABSMMAAAAAAKRhgAAAAABSMMAAAAgIIRBgAAAEDBCAMAAACgYIQBAAAAUDDCAAAAACgYYQAAAAAUjDAAAAAACkYYAAAAAAUjDAAAAICCEQYAAABAwQgDAAAAoGCEAQAAAFAwwgAAAAAoGGEAAAAAFIwwAAAAAAqmck83AAAA8FH9rp29p1uAfZ47AwAAAKBghAEAAABQMMIAAAAAKBhhAAAAABSMBQRhL2EhnfJYPu2UPd0CAADsce4MAAAAgIIRBgAAAEDBeEwAAADKxGN9wN7CnQEAAABQMO4MAArFX2zKw0KMAAB7N3cGAAAAQMG4MwCAT80dFuXhDovycD4CwKcnDACAPcRFLACwp3hMAAAAAApGGAAAAAAFIwwAAACAghEGtLGlS5fmzDPPzNixY3PMMcfkjDPOyNKlS/d0WwAAABSIMKANvfbaa6mtrc21116buXPn5vnnn0+7du0yYsSI1NfX7+n2AAAAKAhhQBu65pprMnr06AwbNixJUllZmZtvvjkrVqzIHXfcsYe7AwAAoCiEAW1kw4YNeeSRR5qCgG0GDBiQmpqazJs3bw91BgAAQNEIA9rISy+9lM2bN6d3794tth1++OF59dVX90BXAAAAFFHlnm6gKLatCdClS5cW2w466KCsXbv2Y/cfPnz4dsd/9atfZb/99tvh9s+K+t+u2tMtAAAAfCrDF9y0p1v4WIsXL06nTp1ata8woI2sX78+SVJVVdViW4cOHbJ169ZWHbeysrLV/+e3lcWLF6dDkqOPPnpPt0LBLV68OIlzkc8G5yOfJc5HPkucj3xWLF68OIsX/9/P9LnYqVOndOvWrVX7CgPaSPv27ZMkDQ0NLbY1NDSkc+fOH7v/s88+u1v6agvb7lrYm78H9g3ORT5LnI98ljgf+SxxPvJZsa+fi9YMaCM9evRIkqxcubLFtrVr16ampqatWwIAAKCghAFtpG/fvkmSl19+ucW2119/PUOHDm3rlgAAACgoYUAbOfzww9O/f//MnTu32fh7772XZcuW5fzzz99DnQEAAFA0woA2NG3atDzzzDOZMWNGkqRUKuWGG27IWWedldra2j3bHAAAAIUhDGhDZ599dmbOnJm/+qu/ypAhQ/KHf/iHOfDAA3Pvvffu6dYAAAAokIpSqVTa000AAAAAbcedAQAAAFAwwgAAAAAoGGEAAAAAFIwwAAAAAApGGAAAAAAFIwwAgM+o119/PfPnz9/TbUCWLVuWG264IU888cSebgWg7LZs2bKnW9gjhAHskqVLl+bMM8/M2LFjc8wxx+SMM87I0qVLP3afUqmU6dOnp7a2NiNGjMigQYNy6623tlHHAG2jNb8fP+rJJ5/MYYcdlp/85Ce7r0kKYVfOxVWrVuXCCy/MpEmTcvrpp+fEE0/czd2yr2vN+bh27dpcd911GTNmTEaMGJGBAwdm4sSJWbt2bRt1zb7qzTffzPe///3069dvp+r3ueuYErTSq6++WurZs2eprq6uVCqVSo2NjaUzzzyz1L1799Jbb721w/0uv/zy0je+8Y1SY2NjqVQqlZ599tlSVVVV6frrr2+Tvtl3LVmypHTGGWeUxowZUzr66KNLp59+emnJkiU7rN+4cWPpL/7iL0qDBg0qff7zny916dKlNG7cuNL8+fPbsGv2Ra39/bjNqlWrSoMHDy4lKf2n//SfdnO37Mt25VxcvHhxqV+/fqXbbrutLVqlAFpzPjY0NJS+9KUvla688srS5s2bS6VSqfQf//EfpZqamtI3vvGNNuudfcvMmTNLp512Wumcc84pfe5znysdeOCBO7XfvnYdIwyg1b7xjW+Uzj///GZjr7zySilJ6dprr93uPi+99FIpSek3v/lNs/ELLrigtN9++5VWrVq1u9plH9eaDxjTp08vffGLXyzNmDGj9Ktf/ar005/+tHTwwQeXKisrS/PmzWvD7tnXtOb340ddcMEFpYcfflgYwC5r7bn4xhtvlHr27FmaOnXq7m6RAmnN+XjfffeVkpR+97vfNRv/y7/8y1KnTp12W6/s2370ox+Vli9fXiqVSqWamppSnz59PnGfffE6xmMCtMqGDRvyyCOPZNiwYc3GBwwYkJqamsybN2+7+91///3p2rVrBg4c2Gx8zJgx2bhxY+rq6nZbz+zbrrnmmowePbrpnKysrMzNN9+cFStW5I477tjuPm+++WbuueeenH/++Rk8eHAmTJiQX/ziF9m8eXOmTZvWlu2zD2nt78dtHn744WzdujVnnHHGbuySItiVc/Giiy5K9+7dc8011+zuNimI1p6Pb775ZpLk/fffbzbeqVOnDBo0aPc0yz7vO9/5Tmpqapq+3m+//T5xn33xOkYYQKu89NJL2bx5c3r37t1i2+GHH55XX311u/s999xzO9wnyQ73g4/T2g8Yxx9/fIYOHdps7Ktf/Wq6du2aN954Y7f1y76ttb8fk2TFihX5/ve/n9tuu203dkhRtPZcrKury5w5c/Ld7343FRUVu7tNCqK15+NXv/rVJMmf/Mmf5N13302SNDY25pe//GV++MMf7r6GKZQOHTp8Ys2+eB0jDKBV6uvrkyRdunRpse2ggw7a4YIu9fX1O9wniYVgaJXWfsA455xzWnzQ3bp1azZu3Jgjjjhit/TKvq+1vx+T5Iorrsg111yT6urq3dYfxdHac/HBBx9Mkhx22GH5r//1v+bkk0/OoEGDMmHChCxfvny39cu+rbXn41e/+tX81V/9VebPn5+BAwfmz//8z/Mnf/InueeeezJ48ODd2jPFsTPB5754HVO5pxtg77R+/fokSVVVVYttHTp0yNatW3e43wEHHLDdfZLscD/4OLty8fX75s6dm40bN+byyy8vW38US2t/P/7jP/5j3nzzzUyYMGG39kdxtPZcXLRoUZLkZz/7WW655ZZUV1fn17/+dU4++eTMnz8/S5Ysyec+97nd1zj7pNaej0ly8cUX57e//W2OOuqoPPzww5kzZ06ee+65/OhHP8oJJ5yw23qGj9oXr2PcGUCrtG/fPknS0NDQYltDQ0M6d+68w/12tE+SHe4HH2dXPmB81JYtW3LDDTdk4sSJGTNmTFl7pDha8/tx48aNmThxYu68887d3h/F0dr/Vr/77rv53Oc+l7/+679uuktl8ODB+cEPfpD/+I//yKOPPrr7mmaf1drzceXKlTnxxBMzceLEXHLJJXniiScyb968rFq1KuPGjXO3Cm1mX7yOEQbQKj169Ejy4S/o37d27dpmC3L8/n472ifJDveDj9PaDxi/7/rrr0///v337vfFsse15vfjP//zP6dUKmXixImpra1NbW1tvva1ryX58G6V2tra/MVf/MXubZx9Tmv/W11RUZEDDzwwnTp1ajZeW1ubJC6+aJXWno833HBDDjvssKbnspMPz8V77703GzZsyMMPP7x7Gobfsy9ex3hMgFbp27dvkuTll19use3111/PyJEjd7hfXV1dGhsbm/0V9/XXX09FRUWGDBmyexpmn9baDxgfdc8996S+vj4/+9nP0q6dnJTWa83vx1NOOSWnnHJKs7Hly5fn0EMPzUUXXZQbb7xxt/TKvq21/60eMGBA5s+fn1Kp1Ow52lKplCTp1avXbuiWfV1rz8cFCxbk5JNPbjH+la98JUmy//77l7FLimrz5s2fWLMvXsf4xEurHH744enfv3/mzp3bbPy9997LsmXLcv755ydJPvjgg/zud79r2n7SSSelsbExCxYsaLbfwoULM3LkyBxyyCG7v3n2OZ/0AeP33xjw+2bOnJlf//rX+bu/+7umuwygtVr7+/H3ffDBB0l27gMKbE9rz8VTTz0169evz69+9atm+z377LP53Oc+t90LM/gkrT0fu3TpkhdffLHF8V566aV07Ngxp5122u5tnH3e1q1bt/vf2kJcx5SglWbOnFlKUrrvvvtKpVKptHXr1tIll1xSOvvss5tqjjrqqFKnTp1Ky5cvL5VKpdKmTZtKxx57bGno0KGl1atXl0qlUuk3v/lNqWfPnqVf/epXbf9NsM/o379/6atf/WqzsXfffbfUrl270rx580qlUqm0YcOG0sqVK5vV/PKXvyxdeeWVpa1bt7ZVqxRAa34//r4nn3yylKR04YUXtknP7Jtacy5u3LixNGjQoNIf/uEfltauXVsqlUqlN998s9S/f//S3/7t37b9N8E+ozXn48yZM0sVFRXNzr233nqrNGTIkNKPfvSjtv0G2Oc0NjaWPv/5z5e6d+/e4rNgEa5jKkql//+eL2iFBx98MFOmTEllZWU6deqUkSNH5gc/+EHTqprjx4/P0qVLs2jRoqZbuVesWJGrr746Tz/9dPr06ZNOnTrlhhtuaPGOePg0HnzwwZxzzjm577778s1vfjOlUinf/e53s2LFisycOTNJMmjQoCxfvjz//u//npqamsybNy+nnHJKvvCFLzQ71pYtW7Jx48aceuqp3mFMq7Xm92OS3Hvvvfnrv/7rLFmyJCtXrkxFRUW+/OUv57TTTvO4AK3SmnPxnXfeyZ/92Z/l2WefzRe+8IW0b98+kyZNclcAu6w15+M//dM/5S//8i/z/vvvp7q6Ovvvv3++973vZfTo0XvyW2Ev9uabb+bCCy/Mb3/72/zmN79J8uHnxO7du+exxx5Lp06dCnEdIwwA9hmf5gNG+/bt069fv6xbt26Hx/vmN7+Z++67r63aBwCANiMMAApr06ZNqays3O6CgaVSKVu3brWGAAAA+yRhAAAAABSMtwkAAABAwQgDAAAAoGCEAQAAAFAwwgAAAAAoGGEAAAAAFIwwAAAAAApGGAAAAAAFIwwAAACAghEGAAAAQMEIAwAAAKBghAEAAABQMMIAAAAAKBhhAAAAABSMMAAAAAAKRhgAAAAABSMMAAAAgIL5/wAIxtqdeNBHcAAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 1162x774.2 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cnn_scores_active['prediction'].plot.hist()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Sample light curves of active subjects"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "REAL_THRESHOLD = 0.1\n",
    "JUNK_THRESHOLD = 0.9"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Real"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "305337"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/6e49e274-87e4-4c68-9813-56fa75149faf.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/7a9b4099-23b5-4ae2-96e5-c23a8551b5b9.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/4a4f71d3-846e-477a-a510-694d30a50b68.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/be79a551-c208-453f-ae29-dacdffb87cf6.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/ac39075e-6d19-4816-8925-ba444a0e9331.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/6f8b98ab-ab50-4218-bdc9-7ac495f4b46f.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/50a61883-3d88-438c-8ed1-3589e3fcce39.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/5ed7725d-24bf-4c4a-b360-e783a5c51c58.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/d1a54ad7-ada0-4b33-b0e7-e34e13eaefcf.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/fdb55e78-f09d-4bd5-8940-e017af8c102b.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "active_real = cnn_scores_active[cnn_scores_active['prediction'] < REAL_THRESHOLD]\n",
    "display(len(active_real))\n",
    "for url in active_real.head(10)['location']:\n",
    "    display(Image(url=url))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Junk"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "252613"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/b272f5c7-e8ea-41b6-9bb7-d3ff4f33d837.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/b5a3a282-16fb-4d18-b135-fd774f9d11ef.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/c98597bd-d34c-4c6e-89f4-e5ab0463a8cd.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/7cef315a-2e09-4cf8-a88e-d4dbe504354e.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/f53d026c-6df8-41b8-8161-3dbd93c97508.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/66fdb063-5fe1-48ef-a3c7-870307d4b9c0.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/7a69b05a-a0b5-4a9c-8da4-9d734bc69243.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/4298c8f3-56f0-4fa5-a47c-784898dadeeb.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/f02afb3a-b264-4e0a-ae8b-8ca140cc1152.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/f6eaaed6-54a9-4e88-9183-6f6a05159d77.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "active_junk = cnn_scores_active[cnn_scores_active['prediction'] >= JUNK_THRESHOLD]\n",
    "display(len(active_junk))\n",
    "for url in active_junk.head(10)['location']:\n",
    "    display(Image(url=url))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### In the middle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "59128"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/fbd01b85-e007-4c6e-92e8-ef49ac488625.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/2f7b19b2-d7ee-4c5e-8da4-fcfe64bfb4e4.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/ecd7b73b-ad3d-40c3-8b4e-34b862380884.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/b1d1a791-1e87-4586-816f-cfd5fa4962e7.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/f6177c9c-1720-41f7-8cea-6afc734e11cb.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/ae0d922a-a1c5-4a84-b3a0-51685083bd3d.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/334097fe-e2ea-4bcf-b6cf-d5e05cc0903b.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/a4d40eab-c30d-45ba-9cfa-40772c754102.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/a89b522d-717b-489b-8bdf-bdb53f651df9.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"https://panoptes-uploads.zooniverse.org/subject_location/719a7606-0c23-4973-9716-3be7931d09aa.gif\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "active_middle = cnn_scores_active[(cnn_scores_active['prediction'] >= REAL_THRESHOLD) & (cnn_scores_active['prediction'] < JUNK_THRESHOLD)]\n",
    "display(len(active_middle))\n",
    "for url in active_middle.head(10)['location']:\n",
    "    display(Image(url=url))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "superwasp_tools",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}