{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import torch\n",
    "import numpy as np\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.DataFrame()\n",
    "\n",
    "for file in os.listdir(\"./dataset/train\"):\n",
    "    df = pd.concat([df, pd.read_parquet(f\"./dataset/train/{file}\")])\n",
    "    \n",
    "grouped_df = df.groupby('subject_id', group_keys=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing subject_id:  1\n",
      "Finished processing subject_id:  1\n",
      "Processing subject_id:  2\n",
      "Finished processing subject_id:  2\n",
      "Processing subject_id:  3\n",
      "Finished processing subject_id:  3\n",
      "Processing subject_id:  4\n",
      "Finished processing subject_id:  4\n",
      "Processing subject_id:  5\n",
      "Finished processing subject_id:  5\n",
      "Processing subject_id:  6\n",
      "Finished processing subject_id:  6\n",
      "Processing subject_id:  7\n",
      "Finished processing subject_id:  7\n",
      "Processing subject_id:  8\n",
      "Finished processing subject_id:  8\n"
     ]
    }
   ],
   "source": [
    "for subject_id, group in grouped_df:\n",
    "    print(\"Processing subject_id: \", subject_id)\n",
    "    arr = np.array([np.array([y for y in x]) for x in group[\"EEG\"].values])\n",
    "    mean_each_channel = arr.mean(axis=2)\n",
    "    std_each_channel = arr.std(axis=2)\n",
    "    # Reshape mean and std to match arr dimensions for broadcasting\n",
    "    mean_each_channel = mean_each_channel[:, :, np.newaxis]\n",
    "    std_each_channel = std_each_channel[:, :, np.newaxis]\n",
    "    normalized_arr = (arr - mean_each_channel) / std_each_channel\n",
    "    \n",
    "    normalized_df = pd.DataFrame()\n",
    "    normalized_df[\"EEG\"] = normalized_arr.tolist()\n",
    "    normalized_df[\"subject_id\"] = subject_id\n",
    "    normalized_df[\"coco_id\"] = group[\"coco_id\"].values\n",
    "    normalized_df[\"curr_time\"] = group[\"curr_time\"].values\n",
    "    normalized_df[\"session\"] = group[\"session\"].values\n",
    "    normalized_df[\"trial\"] = group[\"trial\"].values\n",
    "    normalized_df[\"block\"] = group[\"block\"].values\n",
    "    normalized_df[\"73k_id\"] = group[\"73k_id\"].values\n",
    "    \n",
    "    normalized_df.to_parquet(f\"./dataset/normalized_subject/subject_{subject_id}.parquet\")\n",
    "    \n",
    "    print(\"Finished processing subject_id: \", subject_id)\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": ".venv",
   "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": "3.12.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}