{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# investigate easycom transcript\n",
    "# transcript jisoo gaved me is not align what I got\n",
    "# I want to know whether there are overlap or not"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/home/junhyeok/projects/turn-taking-dataset/notebooks\n",
      "27582\n",
      "6895\n"
     ]
    }
   ],
   "source": [
    "import json\n",
    "import os\n",
    "import numpy as np\n",
    "\n",
    "print(os.getcwd())\n",
    "ts_dir = '../mmaction2/data/EasyComDataset_fix/Main/Speech_Transcriptions/val/Session_1/'\n",
    "\n",
    "ts_list = os.listdir(ts_dir)\n",
    "ts_list.sort()\n",
    "\n",
    "sum_end_frame = 0\n",
    "\n",
    "for ts_path in ts_list:\n",
    "  with open(os.path.join(ts_dir, ts_path), 'r') as f:\n",
    "    ts = json.load(f)\n",
    "  \n",
    "  max_end_frame = 0\n",
    "  for utt in ts:\n",
    "    if utt['End_Frame'] > max_end_frame:\n",
    "      max_end_frame = utt['End_Frame']\n",
    "  \n",
    "  sum_end_frame += max_end_frame\n",
    "\n",
    "print(sum_end_frame)\n",
    "print(sum_end_frame//4)\n",
    "  \n",
    "  \n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "ttd",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.17"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}