Datasets
Models
Downloads: 1
[fc9ccf]:
/
2a-preprocess-pickle-study_to_data.ipynb
696 lines (695 with data), 24.5 kB
{
"cells": [
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"import fastai\n",
"import pickle\n",
"from fastai.utils.mem import *"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [],
"source": [
"stage = 'stage_2'\n",
"data_dir = Path('data')\n",
"fn_to_study_ix = pickle.load(open(f'{data_dir}/{stage}_fn_to_study_ix.pickle', 'rb'))\n",
"study_ix_to_fn = pickle.load(open(f'{data_dir}/{stage}_study_ix_to_fn.pickle', 'rb'))\n",
"fn_to_labels = pickle.load( open(f'{data_dir}/{stage}_train_fn_to_labels.pickle', 'rb'))"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>fn</th>\n",
" <th>study</th>\n",
" <th>any</th>\n",
" <th>epidural</th>\n",
" <th>intraparenchymal</th>\n",
" <th>intraventricular</th>\n",
" <th>subarachnoid</th>\n",
" <th>subdural</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>ID_76d55d9d0</td>\n",
" <td>ID_0000298a7d</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>ID_96d282ea9</td>\n",
" <td>ID_0000298a7d</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>ID_7d8a7c29d</td>\n",
" <td>ID_0000298a7d</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>ID_4d4401491</td>\n",
" <td>ID_0000298a7d</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>ID_8f5ded0b7</td>\n",
" <td>ID_0000298a7d</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" <td>False</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" fn study any epidural intraparenchymal \\\n",
"0 ID_76d55d9d0 ID_0000298a7d False False False \n",
"1 ID_96d282ea9 ID_0000298a7d False False False \n",
"2 ID_7d8a7c29d ID_0000298a7d False False False \n",
"3 ID_4d4401491 ID_0000298a7d False False False \n",
"4 ID_8f5ded0b7 ID_0000298a7d False False False \n",
"\n",
" intraventricular subarachnoid subdural \n",
"0 False False False \n",
"1 False False False \n",
"2 False False False \n",
"3 False False False \n",
"4 False False False "
]
},
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"classes = ['any', 'epidural', 'intraparenchymal', 'intraventricular', 'subarachnoid', 'subdural']\n",
"labels = []\n",
"for fn, lbls in fn_to_labels.items():\n",
" row = {'fn':fn,'study':fn_to_study_ix[fn][0]}\n",
" for c in classes:\n",
" row[c] = c in fn_to_labels[fn]\n",
" labels.append(row)\n",
"\n",
"labels_df = pd.DataFrame(labels)\n",
"labels_df.head() "
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>any</th>\n",
" <th>epidural</th>\n",
" <th>intraparenchymal</th>\n",
" <th>intraventricular</th>\n",
" <th>subarachnoid</th>\n",
" <th>subdural</th>\n",
" <th>any_size</th>\n",
" <th>epidural_size</th>\n",
" <th>intraparenchymal_size</th>\n",
" <th>intraventricular_size</th>\n",
" <th>subarachnoid_size</th>\n",
" <th>subdural_size</th>\n",
" <th>strat_class</th>\n",
" </tr>\n",
" <tr>\n",
" <th>study</th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>ID_0c828b6688</th>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>ID_fe788cb8e6</th>\n",
" <td>6.0</td>\n",
" <td>0.0</td>\n",
" <td>2.0</td>\n",
" <td>0.0</td>\n",
" <td>4.0</td>\n",
" <td>0.0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>101010.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>ID_676bad1cd8</th>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>ID_d0df6455dc</th>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>ID_db74a340fd</th>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>ID_29ecf20f76</th>\n",
" <td>7.0</td>\n",
" <td>0.0</td>\n",
" <td>5.0</td>\n",
" <td>5.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>101100.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>ID_2d80aed85a</th>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>ID_d9cf56413e</th>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>ID_a2d95718f5</th>\n",
" <td>15.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>15.0</td>\n",
" <td>2</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>2</td>\n",
" <td>200002.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>ID_8fd4e085bf</th>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0.0</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" any epidural intraparenchymal intraventricular \\\n",
"study \n",
"ID_0c828b6688 0.0 0.0 0.0 0.0 \n",
"ID_fe788cb8e6 6.0 0.0 2.0 0.0 \n",
"ID_676bad1cd8 0.0 0.0 0.0 0.0 \n",
"ID_d0df6455dc 0.0 0.0 0.0 0.0 \n",
"ID_db74a340fd 0.0 0.0 0.0 0.0 \n",
"ID_29ecf20f76 7.0 0.0 5.0 5.0 \n",
"ID_2d80aed85a 0.0 0.0 0.0 0.0 \n",
"ID_d9cf56413e 0.0 0.0 0.0 0.0 \n",
"ID_a2d95718f5 15.0 0.0 0.0 0.0 \n",
"ID_8fd4e085bf 0.0 0.0 0.0 0.0 \n",
"\n",
" subarachnoid subdural any_size epidural_size \\\n",
"study \n",
"ID_0c828b6688 0.0 0.0 0 0 \n",
"ID_fe788cb8e6 4.0 0.0 1 0 \n",
"ID_676bad1cd8 0.0 0.0 0 0 \n",
"ID_d0df6455dc 0.0 0.0 0 0 \n",
"ID_db74a340fd 0.0 0.0 0 0 \n",
"ID_29ecf20f76 0.0 0.0 1 0 \n",
"ID_2d80aed85a 0.0 0.0 0 0 \n",
"ID_d9cf56413e 0.0 0.0 0 0 \n",
"ID_a2d95718f5 0.0 15.0 2 0 \n",
"ID_8fd4e085bf 0.0 0.0 0 0 \n",
"\n",
" intraparenchymal_size intraventricular_size \\\n",
"study \n",
"ID_0c828b6688 0 0 \n",
"ID_fe788cb8e6 1 0 \n",
"ID_676bad1cd8 0 0 \n",
"ID_d0df6455dc 0 0 \n",
"ID_db74a340fd 0 0 \n",
"ID_29ecf20f76 1 1 \n",
"ID_2d80aed85a 0 0 \n",
"ID_d9cf56413e 0 0 \n",
"ID_a2d95718f5 0 0 \n",
"ID_8fd4e085bf 0 0 \n",
"\n",
" subarachnoid_size subdural_size strat_class \n",
"study \n",
"ID_0c828b6688 0 0 0.0 \n",
"ID_fe788cb8e6 1 0 101010.0 \n",
"ID_676bad1cd8 0 0 0.0 \n",
"ID_d0df6455dc 0 0 0.0 \n",
"ID_db74a340fd 0 0 0.0 \n",
"ID_29ecf20f76 0 0 101100.0 \n",
"ID_2d80aed85a 0 0 0.0 \n",
"ID_d9cf56413e 0 0 0.0 \n",
"ID_a2d95718f5 0 2 200002.0 \n",
"ID_8fd4e085bf 0 0 0.0 "
]
},
"execution_count": 4,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"study_labels = labels_df.groupby('study').agg('sum').reindex()\n",
"for c in classes :\n",
" study_labels[c+'_size'] = 0 # 0 :not present, 1:small, 2:big\n",
" c_idx = study_labels.query(f'{c}>0').sort_values(c).index\n",
" c_small, c_big = c_idx[:len(c_idx)//2], c_idx[len(c_idx)//2:]\n",
" study_labels.loc[c_small,c+'_size'] = 1\n",
" study_labels.loc[c_big,c+'_size'] = 2\n",
"size_classes = [c+'_size' for c in classes]\n",
"study_labels['strat_class'] = study_labels[size_classes].astype(str).sum(axis=1)\n",
"study_labels.sample(10)"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/home/antor/anaconda3/lib/python3.7/site-packages/sklearn/model_selection/_split.py:667: UserWarning: The least populated class in y has only 1 members, which is less than n_splits=19.\n",
" % (min_groups, self.n_splits)), UserWarning)\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>any</th>\n",
" <th>epidural</th>\n",
" <th>intraparenchymal</th>\n",
" <th>intraventricular</th>\n",
" <th>subarachnoid</th>\n",
" <th>subdural</th>\n",
" <th>any_size</th>\n",
" <th>epidural_size</th>\n",
" <th>intraparenchymal_size</th>\n",
" <th>intraventricular_size</th>\n",
" <th>subarachnoid_size</th>\n",
" <th>subdural_size</th>\n",
" <th>strat_class</th>\n",
" <th>fold</th>\n",
" </tr>\n",
" <tr>\n",
" <th>study</th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>ID_0000298a7d</th>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0.0</td>\n",
" <td>6</td>\n",
" </tr>\n",
" <tr>\n",
" <th>ID_0004c4e54c</th>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0.0</td>\n",
" <td>15</td>\n",
" </tr>\n",
" <tr>\n",
" <th>ID_000a935543</th>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0.0</td>\n",
" <td>2</td>\n",
" </tr>\n",
" <tr>\n",
" <th>ID_000f6fd7db</th>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0.0</td>\n",
" <td>9</td>\n",
" </tr>\n",
" <tr>\n",
" <th>ID_0010b2528e</th>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0.0</td>\n",
" <td>6</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" any epidural intraparenchymal intraventricular \\\n",
"study \n",
"ID_0000298a7d 0.0 0.0 0.0 0.0 \n",
"ID_0004c4e54c 0.0 0.0 0.0 0.0 \n",
"ID_000a935543 0.0 0.0 0.0 0.0 \n",
"ID_000f6fd7db 0.0 0.0 0.0 0.0 \n",
"ID_0010b2528e 0.0 0.0 0.0 0.0 \n",
"\n",
" subarachnoid subdural any_size epidural_size \\\n",
"study \n",
"ID_0000298a7d 0.0 0.0 0 0 \n",
"ID_0004c4e54c 0.0 0.0 0 0 \n",
"ID_000a935543 0.0 0.0 0 0 \n",
"ID_000f6fd7db 0.0 0.0 0 0 \n",
"ID_0010b2528e 0.0 0.0 0 0 \n",
"\n",
" intraparenchymal_size intraventricular_size \\\n",
"study \n",
"ID_0000298a7d 0 0 \n",
"ID_0004c4e54c 0 0 \n",
"ID_000a935543 0 0 \n",
"ID_000f6fd7db 0 0 \n",
"ID_0010b2528e 0 0 \n",
"\n",
" subarachnoid_size subdural_size strat_class fold \n",
"study \n",
"ID_0000298a7d 0 0 0.0 6 \n",
"ID_0004c4e54c 0 0 0.0 15 \n",
"ID_000a935543 0 0 0.0 2 \n",
"ID_000f6fd7db 0 0 0.0 9 \n",
"ID_0010b2528e 0 0 0.0 6 "
]
},
"execution_count": 5,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"from sklearn.model_selection import StratifiedKFold\n",
"\n",
"#this will throw a warning that can be ignored\n",
"skf = StratifiedKFold(n_splits=19, shuffle=True, random_state=1972)\n",
"study_labels['fold'] = -1\n",
"for fold, (oof_idx,f_idx) in enumerate(skf.split(study_labels, study_labels.strat_class)):\n",
" study_labels.loc[study_labels.iloc[f_idx].index, 'fold' ] = fold\n",
"study_labels.head()"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [],
"source": [
"study_to_data = {}\n",
"for study in study_ix_to_fn.keys():\n",
" if study in study_labels.index:\n",
" study_to_data[study] = {'fold': study_labels.loc[study].fold}\n",
" else:\n",
" study_to_data[study] = {'fold': -1} #study not in label set\n",
"pickle.dump(study_to_data, open(f\"data/{stage}_study_to_data.pickle\", \"wb\" ))"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"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": "3.7.3"
}
},
"nbformat": 4,
"nbformat_minor": 2
}