handling_annotation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Tagtog에서 annotation 파일 다운 받기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import json\n",
    "import glob\n",
    "import re\n",
    "import os\n",
    "import pandas as pd\n",
    "import requests\n",
    "import zipfile\n",
    "import shutil\n",
    "import html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "#로그인 위치\n",
    "url = \"https://tagtog.net/-login\"\n",
    "\n",
    "#다운로드 위치\n",
    "file_url = 'https://tagtog.net/leegi/relation_extraction/-downloads/dataset-as-anndoc' \n",
    "o_file = 'abc.zip'  \n",
    "if os.path.exists(o_file):\n",
    "    os.remove(o_file)\n",
    "\n",
    "#로그인 정보\n",
    "#log_info.json file로부터 아이디(계정) 비밀번호 받아오기\n",
    "with open(\"log_info.json\",\"r\") as f:\n",
    "    log_info = json.load(f)\n",
    "login_info = {\n",
    "    'loginid' : log_info[\"login_id\"],\n",
    "    'password' : log_info[\"login_pw\"],\n",
    "}\n",
    "#로그인\n",
    "with requests.Session() as s:\n",
    "    login_req = s.post(url, data=login_info)\n",
    "\n",
    "    # #download annotation zip file from tagtog.net\n",
    "    r = s.get(file_url)\n",
    "\n",
    "    with open(o_file,\"wb\") as output:\n",
    "        output.write(r.content)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "#압축파일 풀기\n",
    "folder_path = \"./tagtog_relation_extraction/\"\n",
    "\n",
    "zip_ = zipfile.ZipFile(\"abc.zip\")\n",
    "if os.path.exists(folder_path):\n",
    "    shutil.rmtree(folder_path)\n",
    "zip_.extractall(folder_path)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# annotation 결과 전처리"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "#폴더 경로\n",
    "folder_name = \"./tagtog_relation_extraction/relation_extraction/\"\n",
    "\n",
    "#context list\n",
    "context_name_list = os.listdir(folder_name + \"ann.json/master/pool\")\n",
    "\n",
    "#relation 폴더 경로\n",
    "relation_folder_paths = glob.glob(folder_name + \"ann.json/master/pool/*\")\n",
    "\n",
    "#context 폴더 경로\n",
    "# contexts_folders_paths = glob.glob(folder_name + \"plain.html/pool/*\")\n",
    "contexts_folders_paths = [folder_name + \"plain.html/pool/\" + c for c in context_name_list]\n",
    "\n",
    "\n",
    "#anntation_lenged 정보\n",
    "annotation_legend = folder_name + \"annotations-legend.json\"\n",
    "with open(annotation_legend,\"r\") as f:\n",
    "    annotation_legend = json.load(f)\n",
    "#relation dictionary 파일로부터 subject와 object의 entity 정보를 추출해주는 함수\n",
    "def get_needed_relation_data(tmp_relation):\n",
    "    subject_token = re.findall(\"\\(+(.+)+\\)\",annotation_legend[tmp_relation[\"relations\"][0]['classId']])[0].split(\"|\")[0]\n",
    "    if subject_token == tmp_relation['entities'][0]['classId']:\n",
    "        sub_entity, obj_entity = tmp_relation['entities']\n",
    "    else:\n",
    "        obj_entity, sub_entity  = tmp_relation['entities']\n",
    "    \n",
    "    # get preprocessed entities\n",
    "    def _get_entity(entity):\n",
    "        outputs = entity['offsets'][0]\n",
    "        outputs['type'] = annotation_legend[entity['classId']].split(\"-\")[1].lower()\n",
    "        return outputs\n",
    "    \n",
    "    output_subject = _get_entity(sub_entity)\n",
    "    output_object = _get_entity(obj_entity)\n",
    "    return output_subject, output_object\n",
    "\n",
    "\n",
    "#relation dictionary 파일로부터 title(relation)을 추출해주는 함수\n",
    "def get_label(relation_json):\n",
    "    label_tag = relation_json['relations'][0]['classId'] #r_6\n",
    "    try:\n",
    "        _,sub_type, label = annotation_legend[re.findall(\"\\(+(.+)+\\|\",annotation_legend[label_tag])[0]].split(\"-\")\n",
    "        return f\"{sub_type}:{label}\"\n",
    "    except:\n",
    "        _,sub_type, = annotation_legend[re.findall(\"\\(+(.+)+\\|\",annotation_legend[label_tag])[0]].split(\"-\")\n",
    "        return f\"{sub_type}:no_relation\"\n",
    "\n",
    "\n",
    "#html 파일로부터 text만 추출해주는 함수\n",
    "def get_context_from_html(html_file):\n",
    "    html_file = re.sub(r\"\\n\",\"\", html_file)\n",
    "    html_file = html.unescape(html_file)\n",
    "    return re.findall(\"(<pre.+>)(.+)(</pre>)\",html_file)[0][1]\n",
    "\n",
    "\n",
    "#entity 정보가 포함된 sentence를 생성해주는 함수\n",
    "def get_sentence_with_entites(subject_entity, object_entity, sentence):\n",
    "    if subject_entity['start'] < object_entity['start']:\n",
    "        entity1,entity2 = subject_entity, object_entity\n",
    "    else:\n",
    "        entity1,entity2 = object_entity, subject_entity\n",
    "    \n",
    "    #entity 시작 위치 및 길이 \n",
    "    ett1_stt, ett1_len = entity1['start'], len(entity1['text'])\n",
    "    ett2_stt, ett2_len = entity2['start'], len(entity2['text'])\n",
    "    \n",
    "    #문장 분리\n",
    "    bf, ett1, mid, ett2, af = sentence[:ett1_stt], \\\n",
    "                            sentence[ett1_stt:ett1_stt+ett1_len], \\\n",
    "                            sentence[ett1_stt+ett1_len:ett2_stt], \\\n",
    "                            sentence[ett2_stt:ett2_stt+ett2_len], \\\n",
    "                            sentence[ett2_stt+ett2_len:]\n",
    "\n",
    "    \n",
    "    if subject_entity['start'] < object_entity['start']:\n",
    "        ett1,ett2 = f\"<sbj:{ett1}>\", f\"<obj:{ett2}>\"\n",
    "    else:\n",
    "        ett1,ett2 = f\"<obj:{ett1}>\", f\"<sbj:{ett2}>\"\n",
    "    \n",
    "    return \"\".join([bf, ett1, mid, ett2, af])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "#dataframe column\n",
    "# id : context title(e.g : 카카오게임, 11번가 등)\n",
    "# sentence w/o entity\n",
    "# sentence w entity\n",
    "# subject_entity\n",
    "# object_entity\n",
    "# class\n",
    "\n",
    "id_list = []\n",
    "sentence_list = []\n",
    "sentence_with_entities_list = []\n",
    "subject_entity_list = []\n",
    "object_entity_list = []\n",
    "relation_list = []\n",
    "\n",
    "\n",
    "# tagtog 데이터를 CSV 형태로 변경\n",
    "for context_name, relation_folder, contexts_folder in zip(context_name_list, relation_folder_paths, contexts_folders_paths):\n",
    "    # relation files와 context files 리스트 출력\n",
    "    file_ids = [file_name.split(\".txt.\")[0] for file_name in os.listdir(relation_folder)]\n",
    "    file_nums = [ids.split(\"-\")[1] for ids in file_ids]\n",
    "    relation_files = [relation_folder + \"/\"+ file_id + \".txt.ann.json\" for file_id in file_ids]\n",
    "    context_files = [contexts_folder + \"/\"+ file_id + \".txt.plain.html\" for file_id in file_ids]\n",
    "    \n",
    "    #json으로 된 relation data와 html로 된 context 데이터 읽기\n",
    "    for relation_file, context_file, file_num in zip(relation_files,context_files, file_nums):\n",
    "        \n",
    "        #subject, object 정보 추출\n",
    "        with open(relation_file, \"r\") as f:\n",
    "            relation_json = json.load(f)\n",
    "        try:\n",
    "            tmp_subject, tmp_object = get_needed_relation_data(relation_json) #subject, object\n",
    "            tmp_label = get_label(relation_json)\n",
    "        except:\n",
    "            continue\n",
    "        \n",
    "        #sentence, sentence with entities 정보 추출\n",
    "        with open(context_file, \"r\") as f:\n",
    "            context_json = f.read()\n",
    "        tmp_sentence = get_context_from_html(context_json)\n",
    "        tmp_sentence_w_entities = get_sentence_with_entites(tmp_subject,tmp_object,tmp_sentence)\n",
    "        \n",
    "        #각 list에 데이터 저장\n",
    "        id_list.append(f\"{context_name}_{file_num}\")\n",
    "        sentence_list.append(tmp_sentence)\n",
    "        sentence_with_entities_list.append(tmp_sentence_w_entities)\n",
    "        subject_entity_list.append(tmp_subject)\n",
    "        object_entity_list.append(tmp_object)\n",
    "        relation_list.append(tmp_label.lower())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "# class 통합\n",
    "class_dict = {\n",
    "    \"org:location\" : \"location\",\n",
    "    \"per:place_of_birth\" : \"location\",\n",
    "    \"org:related\" : \"related\",\n",
    "    'per:coworker' : \"related\",\n",
    "    'per:alternate' : \"alternate\",\n",
    "    'org:alternate' : \"alternate\",\n",
    "    'org:product' : \"product\",\n",
    "    'per:product' : 'product',\n",
    "    'org:no_relation' : \"no_relation\",\n",
    "    'per:no_relation' : \"no_relation\",\n",
    "    'poh:no_relation' : \"no_relation\"\n",
    "}\n",
    "relation_list = list(map(lambda x : class_dict[x] if x in class_dict.keys() else x, relation_list))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 전처리 결과를 Google Sheet으로 옮기기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "# google sheet와 연결\n",
    "# 참고 자료 : https://pbj0812.tistory.com/184\n",
    "\n",
    "# !pip install gspread\n",
    "# !pip install oauth2client\n",
    "\n",
    "import gspread\n",
    "from oauth2client.service_account import ServiceAccountCredentials\n",
    "\n",
    "scope = ['https://spreadsheets.google.com/feeds',\n",
    "         'https://www.googleapis.com/auth/drive']\n",
    "\n",
    "#json key file 위치\n",
    "json_file_name = './stable-house-327308-c4a6a7447abd.json'\n",
    "\n",
    "# json key file을 이용하여 접속\n",
    "credentials = ServiceAccountCredentials.from_json_keyfile_name(json_file_name, scope)\n",
    "gc = gspread.authorize(credentials)\n",
    "\n",
    "#구글 스프레드 시트 주소\n",
    "spreadsheet_url = \"https://docs.google.com/spreadsheets/d/1Zfhy3xh-2or5dUsSRm9d3yIN4Jhaleh32Rxqejy8hbI/edit#gid=0\"\n",
    "\n",
    "# 스프레드시트 문서 가져오기\n",
    "doc = gc.open_by_url(spreadsheet_url)\n",
    "## gc.create(spreadsheet_name) # 스프레드시트 생성\n",
    "\n",
    "\n",
    "#시트 선택하기\n",
    "sheet_name = \"annotation\"\n",
    "\n",
    "#해당 시트가 있는 경우 불러오기\n",
    "cy = doc.worksheet(\"창용\")\n",
    "sh = doc.worksheet(\"상하\")\n",
    "ys = doc.worksheet(\"유석\")\n",
    "mc = doc.worksheet(\"명철\")\n",
    "gs = doc.worksheet(\"기성\")\n",
    "yb = doc.worksheet(\"예빈\")\n",
    "bj = doc.worksheet(\"범진\")\n",
    "worksheet = doc.worksheet(sheet_name)\n",
    "## 403 error가 뜰 경우, google sheets API를 활성시켜줘야 함\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "# # 특정 셀 데이터 가져오기\n",
    "# print(\"특정 셀 데이터 가져오기:\",worksheet.acell(\"B1\").value)\n",
    "\n",
    "# # 행 데이터 가져오기\n",
    "# print(\"행 데이터 가져오기:\",worksheet.row_values(1))\n",
    "\n",
    "# # 열 데이터 가져오기\n",
    "# print(\"행 데이터 가져오기:\",worksheet.col_values(1))\n",
    "\n",
    "# # 특정 영역 선택하여 가져오기\n",
    "# range_list = worksheet.range(\"A1:D1\")\n",
    "# print(\"영역 데이터 가져오기:\",[cell.value for cell in range_list])\n",
    "# # 특정 셀에 데이터 쓰기\n",
    "# worksheet.update_acell(\"A1\",\"ids\")\n",
    "\n",
    "# # 행으로 데이터 추가하기\n",
    "# worksheet.append_row([1,2,3,4])\n",
    "\n",
    "# # 특정 index 행에 데이터 추가\n",
    "# worksheet.insert_row([1,2,3,4], 10)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 시트의 모든 데이터 가져오기\n",
    "values = worksheet.get_all_values()\n",
    "header, rows = values[0], values[1:]\n",
    "data = pd.DataFrame(rows, columns=header)\n",
    "\n",
    "column_list = [\"id\",\"sentence\",\"sentence_with_entity\",\"subject_entity\",\"object_entity\",\"class\"]\n",
    "data = data[column_list]\n",
    "data.head()\n",
    "\n",
    "sen_list =  list(data.sentence_with_entity.values)\n",
    "len(sen_list)\n",
    "#각 sheet의 cell을 업데이트하는 방식\n",
    "for el in [\n",
    "    worksheet,\n",
    "    cy,sh,ys,mc,gs,yb,bj\n",
    "    ]:\n",
    "    el.resize(len(id_list)+1,10)\n",
    "    list_range = f\"a2:f{len(id_list)+1}\"\n",
    "    cell_list = el.range(list_range)\n",
    "    \n",
    "    for i in range(len(cell_list)//len(column_list)):\n",
    "        cell_list[(6*i)].value = id_list[i]\n",
    "        cell_list[(6*i)+1].value = sentence_list[i]\n",
    "        cell_list[(6*i)+2].value = sentence_with_entities_list[i]\n",
    "        cell_list[(6*i)+3].value = str(subject_entity_list[i])\n",
    "        cell_list[(6*i)+4].value = str(object_entity_list[i])\n",
    "        cell_list[(6*i)+5].value = relation_list[i]\n",
    "\n",
    "    el.update_cells(cell_list)\n",
    "\n",
    "\n",
    "# # pandas dataframe을 sheet에 입력하는 방식\n",
    "# # from gspread_dataframe import set_with_dataframe\n",
    "# data = pd.DataFrame(\n",
    "#     {'id':id_list,\n",
    "#     'sentence':sentence_list,\n",
    "#     'sentence_with_entity':sentence_with_entities_list,\n",
    "#     'subject_entity':subject_entity_list,\n",
    "#     'object_entity':object_entity_list,\n",
    "#     'class':relation_list,\n",
    "#     'work1':[0]*len(id_list),\n",
    "#     'Data Error':[0]*len(id_list)\n",
    "#     }\n",
    "# )\n",
    "# set_with_dataframe(cy,data)\n",
    "# set_with_dataframe(sh,data)\n",
    "# set_with_dataframe(ys,data)\n",
    "# set_with_dataframe(mc,data)\n",
    "# set_with_dataframe(gs,data)\n",
    "# set_with_dataframe(yb,data)\n",
    "# set_with_dataframe(bj,data)\n",
    "# set_with_dataframe(worksheet,data)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 306,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from pytz import timezone\n",
    "from datetime import datetime\n",
    "today = datetime.now(timezone('Asia/Seoul'))\n",
    "time_now = \"%s/%s/%s %s:%s:%s\" %(today.year, today.month,today.day,today.hour, today.minute, today.second)\n",
    "\n",
    "# !pip install seaborn\n",
    "# !pip install matplotlib\n",
    "values = worksheet.get_all_values()\n",
    "header, rows = values[0], values[1:]\n",
    "data = pd.DataFrame(rows, columns=header)\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "total_class_num = len(data)\n",
    "tmp = data['class'].value_counts(ascending=True).to_frame()\n",
    "idx = tmp.index\n",
    "values = list(tmp[\"class\"].values)\n",
    "plt,ax = plt.subplots(1,1, figsize=(20,10))\n",
    "ax.barh(y=idx,width=values)\n",
    "for i,s in enumerate(values):\n",
    "    ax.annotate(text = f\"{s}\\n({round(s/total_class_num*100,2)}%)\",\n",
    "                xy = (s+int(values[-1]*0.037),i), va=\"center\", ha=\"center\",fontsize=12,fontweight=\"semibold\")\n",
    "\n",
    "# ax.set_title(f\"num(ratio) of class   [ {total_class_num} ea ]\", fontsize=20,fontweight=\"bold\")\n",
    "ax.set_title(f\"num(ratio) of class   ||   {total_class_num} ea (all)   ||   {time_now}\", fontsize=20,fontweight=\"bold\")\n",
    "ax.spines['top'].set_visible(False)\n",
    "ax.spines['right'].set_visible(False)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "**************************************************\n",
      "**************************************************\n",
      "********** 현재 시간: 2021/11/17 0:4:41 **********\n",
      "**************************************************\n",
      "************************************************** \n",
      "\n",
      "<Worksheet '창용' id:885499266> : 814 [100.0%]\n",
      "<Worksheet '상하' id:505361167> : 814 [100.0%]\n",
      "<Worksheet '유석' id:1649326117> : 814 [100.0%]\n",
      "<Worksheet '명철' id:990973882> : 814 [100.0%]\n",
      "<Worksheet '기성' id:677835246> : 813 [99.88%]\n",
      "<Worksheet '예빈' id:261734676> : 814 [100.0%]\n",
      "<Worksheet '범진' id:933366861> : 814 [100.0%]\n"
     ]
    }
   ],
   "source": [
    "from pytz import timezone\n",
    "from datetime import datetime\n",
    "today = datetime.now(timezone('Asia/Seoul'))\n",
    "time_now = \"%s/%s/%s %s:%s:%s\" %(today.year, today.month,today.day,today.hour, today.minute, today.second)\n",
    "print(\"*\"*50)\n",
    "print(\"*\"*50)\n",
    "print(f\"********** 현재 시간: {time_now} **********\")\n",
    "print(\"*\"*50)\n",
    "print(\"*\"*50,\"\\n\")\n",
    "for el in [cy,sh,ys,mc,gs,yb,bj]:\n",
    "    values = el.get_all_values()\n",
    "    headers,rows = values[0], values[1:]\n",
    "    data = pd.DataFrame(rows,columns=headers)\n",
    "    annot = sum(data.work1.map(lambda x : len(x)>0))\n",
    "    print(f\"{el} : {annot} [{round(annot/len(data)*100,2)}%]\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Annotation 결과 분석"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "import gspread\n",
    "from oauth2client.service_account import ServiceAccountCredentials\n",
    "\n",
    "scope = ['https://spreadsheets.google.com/feeds',\n",
    "         'https://www.googleapis.com/auth/drive']\n",
    "\n",
    "#json key file 위치\n",
    "json_file_name = './stable-house-327308-c4a6a7447abd.json'\n",
    "\n",
    "# json key file을 이용하여 접속\n",
    "credentials = ServiceAccountCredentials.from_json_keyfile_name(json_file_name, scope)\n",
    "gc = gspread.authorize(credentials)\n",
    "\n",
    "#구글 스프레드 시트 주소\n",
    "spreadsheet_url = \"https://docs.google.com/spreadsheets/d/1Zfhy3xh-2or5dUsSRm9d3yIN4Jhaleh32Rxqejy8hbI/edit#gid=0\"\n",
    "\n",
    "# 스프레드시트 문서 가져오기\n",
    "doc = gc.open_by_url(spreadsheet_url)\n",
    "## gc.create(spreadsheet_name) # 스프레드시트 생성\n",
    "\n",
    "\n",
    "#시트 선택하기\n",
    "sheet_name = \"annotation\"\n",
    "\n",
    "#해당 시트가 있는 경우 불러오기\n",
    "cy = doc.worksheet(\"창용\").get_all_values()\n",
    "sh = doc.worksheet(\"상하\").get_all_values()\n",
    "ys = doc.worksheet(\"유석\").get_all_values()\n",
    "mc = doc.worksheet(\"명철\").get_all_values()\n",
    "gs = doc.worksheet(\"기성\").get_all_values()\n",
    "yb = doc.worksheet(\"예빈\").get_all_values()\n",
    "bj = doc.worksheet(\"범진\").get_all_values()\n",
    "worksheet = doc.worksheet(sheet_name).get_all_values()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "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>id</th>\n",
       "      <th>sentence</th>\n",
       "      <th>sentence_with_entity</th>\n",
       "      <th>subject_entity</th>\n",
       "      <th>object_entity</th>\n",
       "      <th>class</th>\n",
       "      <th>work1</th>\n",
       "      <th>Data Error</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>최소기능제품_67</td>\n",
       "      <td>MVP는 사용자가 제품의 특징과 미래를 정의할 수 있도록 사용자의 말을 듣는 공개 ...</td>\n",
       "      <td>&lt;sbj:MVP&gt;는 사용자가 제품의 특징과 미래를 정의할 수 있도록 사용자의 말을 ...</td>\n",
       "      <td>{'start': 0, 'text': 'MVP', 'type': 'poh'}</td>\n",
       "      <td>{'start': 60, 'text': '조기 릴리스', 'type': 'poh'}</td>\n",
       "      <td>no_relation</td>\n",
       "      <td></td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>최소기능제품_22</td>\n",
       "      <td>MVP는 핵심적인 기능만 있는 제품이기 때문에 외관상 조악하고 부실하게 보일 수 있...</td>\n",
       "      <td>&lt;obj:MVP&gt;는 핵심적인 기능만 있는 제품이기 때문에 외관상 조악하고 부실하게 ...</td>\n",
       "      <td>{'start': 52, 'text': '창업자', 'type': 'per'}</td>\n",
       "      <td>{'start': 0, 'text': 'MVP', 'type': 'poh'}</td>\n",
       "      <td>product</td>\n",
       "      <td></td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>최소기능제품_37</td>\n",
       "      <td>예를 들어, 2015년에 시드니 대학의 전문가들이 농장과 잡초 관리를 자동화하기 위...</td>\n",
       "      <td>예를 들어, 2015년에 &lt;sbj:시드니 대학&gt;의 전문가들이 농장과 잡초 관리를 자...</td>\n",
       "      <td>{'start': 14, 'text': '시드니 대학', 'type': 'org'}</td>\n",
       "      <td>{'start': 48, 'text': '리파 로봇', 'type': 'poh'}</td>\n",
       "      <td>product</td>\n",
       "      <td></td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>최소기능제품_122</td>\n",
       "      <td>예약하기 전 본인의 위치 주위에 우버 드라이버들을 보여주고, 승차요금을 예상하고, ...</td>\n",
       "      <td>예약하기 전 본인의 위치 주위에 &lt;sbj:우버 드라이버&gt;들을 보여주고, 승차요금을 ...</td>\n",
       "      <td>{'start': 18, 'text': '우버 드라이버', 'type': 'per'}</td>\n",
       "      <td>{'start': 108, 'text': '모바일 앱', 'type': 'poh'}</td>\n",
       "      <td>no_relation</td>\n",
       "      <td></td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>최소기능제품_100</td>\n",
       "      <td>Airbnb는 2007년 샌프란시스코에서 시작되었다.</td>\n",
       "      <td>&lt;sbj:Airbnb&gt;는 2007년 &lt;obj:샌프란시스코&gt;에서 시작되었다.</td>\n",
       "      <td>{'start': 0, 'text': 'Airbnb', 'type': 'org'}</td>\n",
       "      <td>{'start': 14, 'text': '샌프란시스코', 'type': 'loc'}</td>\n",
       "      <td>location</td>\n",
       "      <td></td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           id                                           sentence  \\\n",
       "0   최소기능제품_67  MVP는 사용자가 제품의 특징과 미래를 정의할 수 있도록 사용자의 말을 듣는 공개 ...   \n",
       "1   최소기능제품_22  MVP는 핵심적인 기능만 있는 제품이기 때문에 외관상 조악하고 부실하게 보일 수 있...   \n",
       "2   최소기능제품_37  예를 들어, 2015년에 시드니 대학의 전문가들이 농장과 잡초 관리를 자동화하기 위...   \n",
       "3  최소기능제품_122  예약하기 전 본인의 위치 주위에 우버 드라이버들을 보여주고, 승차요금을 예상하고, ...   \n",
       "4  최소기능제품_100                      Airbnb는 2007년 샌프란시스코에서 시작되었다.   \n",
       "\n",
       "                                sentence_with_entity  \\\n",
       "0  <sbj:MVP>는 사용자가 제품의 특징과 미래를 정의할 수 있도록 사용자의 말을 ...   \n",
       "1  <obj:MVP>는 핵심적인 기능만 있는 제품이기 때문에 외관상 조악하고 부실하게 ...   \n",
       "2  예를 들어, 2015년에 <sbj:시드니 대학>의 전문가들이 농장과 잡초 관리를 자...   \n",
       "3  예약하기 전 본인의 위치 주위에 <sbj:우버 드라이버>들을 보여주고, 승차요금을 ...   \n",
       "4          <sbj:Airbnb>는 2007년 <obj:샌프란시스코>에서 시작되었다.   \n",
       "\n",
       "                                    subject_entity  \\\n",
       "0       {'start': 0, 'text': 'MVP', 'type': 'poh'}   \n",
       "1      {'start': 52, 'text': '창업자', 'type': 'per'}   \n",
       "2   {'start': 14, 'text': '시드니 대학', 'type': 'org'}   \n",
       "3  {'start': 18, 'text': '우버 드라이버', 'type': 'per'}   \n",
       "4    {'start': 0, 'text': 'Airbnb', 'type': 'org'}   \n",
       "\n",
       "                                    object_entity        class work1  \\\n",
       "0  {'start': 60, 'text': '조기 릴리스', 'type': 'poh'}  no_relation         \n",
       "1      {'start': 0, 'text': 'MVP', 'type': 'poh'}      product         \n",
       "2   {'start': 48, 'text': '리파 로봇', 'type': 'poh'}      product         \n",
       "3  {'start': 108, 'text': '모바일 앱', 'type': 'poh'}  no_relation         \n",
       "4  {'start': 14, 'text': '샌프란시스코', 'type': 'loc'}     location         \n",
       "\n",
       "  Data Error  \n",
       "0          0  \n",
       "1          1  \n",
       "2          1  \n",
       "3          0  \n",
       "4          0  "
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.DataFrame(worksheet[1:], columns = worksheet[0])\n",
    "data['subject_entity'] = data.subject_entity.map(eval)\n",
    "data['object_entity'] = data.object_entity.map(eval)\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Error Data 개수: 82\n"
     ]
    }
   ],
   "source": [
    "# Data Error 데이터 정리\n",
    "# 작업자의 annotation 결과 정리\n",
    "from collections import defaultdict, Counter\n",
    "annotation = defaultdict(list)\n",
    "data_error = defaultdict(int)\n",
    "for el in [cy,sh,ys,mc,gs,yb,bj]:\n",
    "    tmp = pd.DataFrame(el[1:], columns = el[0])\n",
    "    for i in range(len(data[\"id\"])):\n",
    "        annotation[i].append(tmp.loc[i,\"work1\"])\n",
    "        data_error[i] += eval(tmp.loc[i, \"Data Error\"])\n",
    "\n",
    "for i in range(len(annotation)):\n",
    "    annotation[i] = Counter(annotation[i])\n",
    "\n",
    "data_error = {k:v for k,v in data_error.items() if v !=0}\n",
    "print(\"Error Data 개수:\",len(data_error))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([{'no_relation': 6, 'poh:type': 1},\n",
       "       {'no_relation': 5, 'product': 2}, {'product': 6, 'no_relation': 1},\n",
       "       {'no_relation': 6, 'poh:type': 1}, {'location': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 6, 'poh:type': 1},\n",
       "       {'no_relation': 6, 'poh:type': 1},\n",
       "       {'product': 6, 'no_relation': 1}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'poh:type': 5, 'no_relation': 2},\n",
       "       {'no_relation': 6, 'poh:type': 1}, {'related': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 6, 'poh:type': 1},\n",
       "       {'product': 7}, {'org:members': 6, 'location': 1},\n",
       "       {'no_relation': 7}, {'no_relation': 4, 'product': 3},\n",
       "       {'poh:type': 7}, {'poh:type': 6, 'no_relation': 1},\n",
       "       {'poh:start_date': 5, 'no_relation': 2}, {'poh:type': 7},\n",
       "       {'poh:type': 6, 'no_relation': 1}, {'no_relation': 7},\n",
       "       {'product': 7}, {'product': 6, 'related': 1},\n",
       "       {'related': 5, 'no_relation': 2},\n",
       "       {'no_relation': 6, 'poh:type': 1},\n",
       "       {'poh:type': 6, 'no_relation': 1}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'per:member_of': 6, 'product': 1},\n",
       "       {'no_relation': 6, 'product': 1}, {'product': 7},\n",
       "       {'poh:type': 6, 'no_relation': 1}, {'no_relation': 7},\n",
       "       {'no_relation': 6, 'poh:type': 1}, {'poh:type': 7},\n",
       "       {'no_relation': 5, 'product': 2},\n",
       "       {'no_relation': 6, 'org:members': 1}, {'poh:start_date': 7},\n",
       "       {'alternate': 7}, {'poh:type': 4, 'no_relation': 3},\n",
       "       {'alternate': 7}, {'poh:type': 7},\n",
       "       {'poh:type': 6, 'no_relation': 1}, {'product': 6, 'org:field': 1},\n",
       "       {'related': 4, 'poh:type': 1, 'org:field': 1, 'no_relation': 1},\n",
       "       {'poh:type': 6, 'org:field': 1}, {'no_relation': 5, 'related': 2},\n",
       "       {'poh:type': 5, 'no_relation': 2}, {'poh:type': 5, 'alternate': 2},\n",
       "       {'poh:type': 6, 'no_relation': 1}, {'no_relation': 7},\n",
       "       {'poh:type': 7}, {'poh:type': 4, 'alternate': 2, 'no_relation': 1},\n",
       "       {'poh:type': 7}, {'poh:type': 5, 'no_relation': 2},\n",
       "       {'no_relation': 4, 'related': 3}, {'location': 7},\n",
       "       {'org:field': 7}, {'org:event': 7},\n",
       "       {'per:title': 6, 'no_relation': 1}, {'per:title': 7},\n",
       "       {'per:title': 7}, {'no_relation': 6, 'per:member_of': 1},\n",
       "       {'no_relation': 7}, {'no_relation': 4, 'related': 3},\n",
       "       {'product': 6, 'no_relation': 1},\n",
       "       {'no_relation': 4, 'per:title': 3},\n",
       "       {'poh:start_date': 4, 'org:founded': 3}, {'related': 7},\n",
       "       {'related': 4, 'no_relation': 3}, {'related': 7}, {'related': 7},\n",
       "       {'poh:type': 7}, {'no_relation': 7},\n",
       "       {'product': 3, 'no_relation': 3, 'org:field': 1},\n",
       "       {'poh:type': 4, 'no_relation': 3}, {'related': 7},\n",
       "       {'poh:type': 6, 'no_relation': 1}, {'org:event': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 7}, {'org:event': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'product': 5, 'org:field': 2}, {'no_relation': 6, 'product': 1},\n",
       "       {'org:field': 5, 'product': 2},\n",
       "       {'related': 4, 'org:members': 1, 'poh:type': 1, 'no_relation': 1},\n",
       "       {'org:founded': 6, 'poh:start_date': 1}, {'related': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 7}, {'product': 7},\n",
       "       {'per:member_of': 4, 'no_relation': 3},\n",
       "       {'related': 6, 'no_relation': 1}, {'no_relation': 7},\n",
       "       {'product': 5, 'no_relation': 2},\n",
       "       {'no_relation': 4, 'per:title': 2, 'poh:type': 1},\n",
       "       {'per:member_of': 6, 'no_relation': 1},\n",
       "       {'no_relation': 6, 'poh:type': 1}, {'no_relation': 7},\n",
       "       {'no_relation': 6, 'related': 1}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'poh:type': 4, 'no_relation': 3},\n",
       "       {'poh:type': 5, 'no_relation': 2}, {'no_relation': 7},\n",
       "       {'no_relation': 5, 'poh:type': 2},\n",
       "       {'no_relation': 6, 'poh:type': 1}, {'no_relation': 7},\n",
       "       {'poh:type': 7}, {'no_relation': 6, 'poh:type': 1},\n",
       "       {'no_relation': 7}, {'no_relation': 6, 'poh:type': 1},\n",
       "       {'no_relation': 7}, {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'poh:type': 4, 'no_relation': 3},\n",
       "       {'org:field': 7}, {'location': 7}, {'no_relation': 7},\n",
       "       {'org:field': 6, 'no_relation': 1},\n",
       "       {'org:founded': 5, 'poh:start_date': 2},\n",
       "       {'poh:type': 4, 'no_relation': 3},\n",
       "       {'no_relation': 4, 'poh:type': 2, 'org:field': 1},\n",
       "       {'alternate': 7}, {'no_relation': 7},\n",
       "       {'poh:type': 6, 'no_relation': 1}, {'no_relation': 7},\n",
       "       {'no_relation': 4, 'poh:type': 3},\n",
       "       {'poh:type': 6, 'no_relation': 1},\n",
       "       {'poh:type': 4, 'no_relation': 3},\n",
       "       {'poh:type': 6, 'no_relation': 1}, {'related': 7},\n",
       "       {'product': 6, 'related': 1}, {'product': 7},\n",
       "       {'no_relation': 4, 'related': 3}, {'related': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'product': 6, 'no_relation': 1}, {'product': 6, 'no_relation': 1},\n",
       "       {'no_relation': 5, 'product': 1, 'poh:type': 1}, {'product': 7},\n",
       "       {'product': 6, 'no_relation': 1}, {'no_relation': 5, 'product': 2},\n",
       "       {'product': 5, 'no_relation': 2}, {'no_relation': 6, 'product': 1},\n",
       "       {'no_relation': 7},\n",
       "       {'org:event': 3, 'org:field': 2, 'no_relation': 1, 'product': 1},\n",
       "       {'no_relation': 5, 'related': 2},\n",
       "       {'no_relation': 6, 'org:event': 1},\n",
       "       {'related': 4, 'no_relation': 3}, {'product': 6, 'related': 1},\n",
       "       {'poh:type': 7}, {'related': 7}, {'product': 6, 'no_relation': 1},\n",
       "       {'no_relation': 4, 'poh:type': 3}, {'poh:start_date': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 6, 'poh:type': 1},\n",
       "       {'no_relation': 7}, {'org:field': 4, 'no_relation': 3},\n",
       "       {'poh:type': 7}, {'poh:type': 7}, {'no_relation': 7},\n",
       "       {'poh:type': 3, 'no_relation': 2, 'alternate': 2},\n",
       "       {'alternate': 3, 'no_relation': 2, 'poh:type': 2}, {'poh:type': 7},\n",
       "       {'no_relation': 5, 'poh:type': 2},\n",
       "       {'no_relation': 3, 'alternate': 3, 'poh:type': 1},\n",
       "       {'no_relation': 7}, {'poh:type': 7},\n",
       "       {'related': 5, 'no_relation': 2},\n",
       "       {'poh:type': 5, 'no_relation': 2},\n",
       "       {'poh:type': 6, 'no_relation': 1}, {'product': 5, 'org:field': 2},\n",
       "       {'poh:start_date': 7}, {'no_relation': 6, 'poh:start_date': 1},\n",
       "       {'poh:start_date': 7}, {'related': 7}, {'poh:start_date': 7},\n",
       "       {'poh:start_date': 6, 'no_relation': 1}, {'poh:start_date': 7},\n",
       "       {'poh:type': 7}, {'no_relation': 6, 'poh:type': 1},\n",
       "       {'poh:type': 5, 'no_relation': 1, 'alternate': 1},\n",
       "       {'org:event': 7}, {'org:field': 7},\n",
       "       {'related': 4, 'no_relation': 3}, {'no_relation': 6, 'product': 1},\n",
       "       {'product': 5, 'no_relation': 2}, {'no_relation': 7},\n",
       "       {'poh:start_date': 7}, {'product': 4, 'no_relation': 3},\n",
       "       {'no_relation': 7}, {'poh:start_date': 6, 'org:founded': 1},\n",
       "       {'no_relation': 6, 'poh:type': 1}, {'no_relation': 7},\n",
       "       {'product': 7}, {'poh:type': 7}, {'product': 6, 'no_relation': 1},\n",
       "       {'no_relation': 7}, {'related': 4, 'no_relation': 3},\n",
       "       {'related': 2, 'poh:type': 2, 'org:members': 1, 'no_relation': 1, 'org:field': 1},\n",
       "       {'no_relation': 7}, {'location': 7},\n",
       "       {'no_relation': 6, 'poh:type': 1}, {'org:field': 7},\n",
       "       {'poh:end_date': 7}, {'org:founded': 7}, {'org:event': 7},\n",
       "       {'related': 5, 'no_relation': 2},\n",
       "       {'location': 4, 'no_relation': 3}, {'no_relation': 7},\n",
       "       {'related': 6, 'no_relation': 1},\n",
       "       {'no_relation': 5, 'poh:type': 2},\n",
       "       {'no_relation': 4, 'poh:type': 3}, {'no_relation': 7},\n",
       "       {'no_relation': 6, 'product': 1},\n",
       "       {'no_relation': 6, 'poh:type': 1}, {'poh:type': 7},\n",
       "       {'no_relation': 6, 'poh:start_date': 1}, {'poh:type': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 7}, {'poh:type': 7},\n",
       "       {'poh:type': 7}, {'poh:type': 4, 'no_relation': 3},\n",
       "       {'no_relation': 7}, {'no_relation': 6, 'poh:type': 1},\n",
       "       {'poh:type': 4, 'no_relation': 3}, {'poh:type': 7},\n",
       "       {'org:field': 5, 'poh:type': 1, 'product': 1},\n",
       "       {'no_relation': 6, 'poh:type': 1}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'poh:type': 7}, {'no_relation': 7},\n",
       "       {'no_relation': 6, 'poh:type': 1},\n",
       "       {'no_relation': 6, 'poh:type': 1}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 4, 'poh:type': 3},\n",
       "       {'no_relation': 6, 'poh:type': 1}, {'no_relation': 7},\n",
       "       {'no_relation': 5, 'poh:type': 2}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 6, 'poh:type': 1},\n",
       "       {'alternate': 6, 'related': 1}, {'poh:type': 6, 'no_relation': 1},\n",
       "       {'no_relation': 6, 'poh:type': 1}, {'poh:type': 7},\n",
       "       {'no_relation': 7},\n",
       "       {'org:field': 5, 'per:member_of': 1, 'no_relation': 1},\n",
       "       {'poh:type': 6, 'no_relation': 1}, {'poh:type': 7},\n",
       "       {'per:title': 7}, {'poh:type': 7}, {'org:field': 6, 'poh:type': 1},\n",
       "       {'poh:start_date': 7}, {'org:field': 7}, {'no_relation': 7},\n",
       "       {'product': 7}, {'related': 6, 'no_relation': 1}, {'org:field': 7},\n",
       "       {'org:event': 7}, {'product': 7}, {'product': 7},\n",
       "       {'poh:start_date': 7}, {'no_relation': 4, 'related': 3},\n",
       "       {'org:founded': 6, 'poh:start_date': 1},\n",
       "       {'related': 6, 'product': 1}, {'no_relation': 7},\n",
       "       {'poh:start_date': 7}, {'poh:type': 7},\n",
       "       {'related': 4, 'no_relation': 3}, {'no_relation': 6, 'related': 1},\n",
       "       {'poh:start_date': 7}, {'no_relation': 6, 'org:event': 1},\n",
       "       {'per:title': 7}, {'poh:start_date': 7},\n",
       "       {'product': 5, 'poh:type': 1, 'no_relation': 1},\n",
       "       {'product': 5, 'poh:type': 2}, {'poh:start_date': 7},\n",
       "       {'poh:start_date': 7}, {'product': 4, 'no_relation': 3},\n",
       "       {'poh:start_date': 6, 'per:member_of': 1},\n",
       "       {'no_relation': 6, 'related': 1}, {'poh:start_date': 7},\n",
       "       {'no_relation': 5, 'poh:start_date': 2}, {'org:members': 7},\n",
       "       {'related': 5, 'no_relation': 2}, {'product': 7},\n",
       "       {'no_relation': 6, 'org:event': 1}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'org:founded': 6, 'poh:start_date': 1},\n",
       "       {'poh:start_date': 7}, {'poh:start_date': 7},\n",
       "       {'poh:start_date': 7}, {'no_relation': 6, 'related': 1},\n",
       "       {'per:member_of': 5, 'no_relation': 2},\n",
       "       {'related': 5, 'product': 2}, {'no_relation': 6, 'related': 1},\n",
       "       {'related': 7}, {'no_relation': 6, 'poh:type': 1},\n",
       "       {'product': 5, 'org:field': 2}, {'poh:type': 7},\n",
       "       {'org:event': 5, 'product': 2}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'poh:start_date': 7}, {'poh:type': 7},\n",
       "       {'poh:start_date': 7}, {'no_relation': 5, 'poh:type': 2},\n",
       "       {'no_relation': 7}, {'poh:start_date': 7}, {'poh:end_date': 7},\n",
       "       {'related': 5, 'no_relation': 2},\n",
       "       {'org:founded': 6, 'poh:start_date': 1}, {'poh:start_date': 7},\n",
       "       {'no_relation': 4, 'related': 3}, {'per:title': 7},\n",
       "       {'poh:start_date': 7}, {'poh:start_date': 7}, {'org:event': 7},\n",
       "       {'poh:type': 5, 'no_relation': 2}, {'poh:start_date': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'org:field': 5, 'product': 2}, {'poh:start_date': 7},\n",
       "       {'poh:start_date': 6, 'no_relation': 1}, {'no_relation': 7},\n",
       "       {'poh:start_date': 7}, {'per:member_of': 7}, {'no_relation': 7},\n",
       "       {'poh:end_date': 7}, {'org:event': 7},\n",
       "       {'no_relation': 6, 'poh:type': 1}, {'no_relation': 7},\n",
       "       {'per:title': 7}, {'related': 6, 'no_relation': 1},\n",
       "       {'poh:start_date': 7}, {'no_relation': 7}, {'poh:start_date': 7},\n",
       "       {'related': 7}, {'per:title': 7}, {'no_relation': 7},\n",
       "       {'no_relation': 6, 'product': 1}, {'poh:start_date': 7},\n",
       "       {'location': 5, 'no_relation': 2},\n",
       "       {'org:field': 4, 'no_relation': 3}, {'poh:start_date': 7},\n",
       "       {'org:field': 6, 'no_relation': 1}, {'poh:end_date': 7},\n",
       "       {'poh:type': 5, 'org:field': 1, 'no_relation': 1},\n",
       "       {'poh:type': 6, 'org:field': 1}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'product': 7},\n",
       "       {'org:event': 4, 'no_relation': 3}, {'poh:end_date': 7},\n",
       "       {'no_relation': 7}, {'per:member_of': 7}, {'poh:start_date': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 7}, {'poh:end_date': 7},\n",
       "       {'product': 7}, {'no_relation': 5, 'poh:type': 2},\n",
       "       {'related': 5, 'product': 2}, {'no_relation': 7},\n",
       "       {'no_relation': 6, 'related': 1}, {'per:title': 7},\n",
       "       {'product': 5, 'no_relation': 2}, {'poh:type': 7},\n",
       "       {'no_relation': 4, 'per:title': 3},\n",
       "       {'poh:type': 5, 'alternate': 2}, {'no_relation': 7},\n",
       "       {'poh:type': 4, 'alternate': 3}, {'no_relation': 7},\n",
       "       {'poh:type': 7}, {'org:members': 7}, {'no_relation': 7},\n",
       "       {'poh:type': 7}, {'poh:type': 7},\n",
       "       {'no_relation': 4, 'org:field': 2, 'poh:type': 1},\n",
       "       {'no_relation': 6, 'related': 1},\n",
       "       {'org:members': 5, 'no_relation': 2}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 7}, {'poh:type': 7},\n",
       "       {'poh:type': 7}, {'poh:type': 7}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 7}, {'org:members': 7},\n",
       "       {'product': 7}, {'poh:start_date': 7}, {'org:members': 7},\n",
       "       {'product': 5, 'org:field': 2},\n",
       "       {'poh:type': 4, 'product': 2, 'org:field': 1}, {'alternate': 7},\n",
       "       {'product': 7}, {'per:title': 6, 'no_relation': 1},\n",
       "       {'org:founded': 6, 'poh:start_date': 1}, {'org:members': 7},\n",
       "       {'product': 7}, {'poh:start_date': 7},\n",
       "       {'location': 6, 'no_relation': 1}, {'no_relation': 7},\n",
       "       {'related': 7}, {'location': 7}, {'alternate': 5, 'related': 2},\n",
       "       {'org:field': 4, 'product': 2, 'no_relation': 1},\n",
       "       {'related': 4, 'no_relation': 3}, {'poh:end_date': 7},\n",
       "       {'related': 7}, {'alternate': 7}, {'poh:type': 4, 'alternate': 3},\n",
       "       {'no_relation': 5, 'related': 2}, {'poh:start_date': 7},\n",
       "       {'product': 7}, {'alternate': 5, 'related': 2}, {'related': 7},\n",
       "       {'related': 7}, {'product': 7}, {'alternate': 5, 'poh:type': 2},\n",
       "       {'org:members': 5, 'poh:start_date': 2},\n",
       "       {'org:founded': 6, 'poh:start_date': 1}, {'related': 7},\n",
       "       {'related': 6, 'no_relation': 1}, {'per:title': 7}, {'related': 7},\n",
       "       {'poh:start_date': 7}, {'product': 7}, {'poh:start_date': 7},\n",
       "       {'related': 7}, {'product': 6, 'no_relation': 1}, {'per:title': 7},\n",
       "       {'product': 7}, {'related': 6, 'no_relation': 1},\n",
       "       {'poh:start_date': 7}, {'no_relation': 6, 'related': 1},\n",
       "       {'no_relation': 7}, {'product': 5, 'no_relation': 2},\n",
       "       {'related': 6, 'no_relation': 1}, {'org:members': 7},\n",
       "       {'related': 7}, {'no_relation': 4, 'related': 3},\n",
       "       {'no_relation': 5, 'product': 1, 'org:event': 1}, {'product': 7},\n",
       "       {'org:founded': 6, 'poh:start_date': 1},\n",
       "       {'org:field': 3, 'no_relation': 2, 'product': 2},\n",
       "       {'product': 6, 'no_relation': 1},\n",
       "       {'org:founded': 6, 'poh:start_date': 1},\n",
       "       {'product': 6, 'no_relation': 1}, {'alternate': 4, 'poh:type': 3},\n",
       "       {'product': 7}, {'poh:start_date': 7}, {'product': 7},\n",
       "       {'product': 3, 'alternate': 3, 'no_relation': 1},\n",
       "       {'no_relation': 5, 'related': 2}, {'related': 7}, {'product': 7},\n",
       "       {'product': 7}, {'product': 7}, {'alternate': 7}, {'related': 7},\n",
       "       {'alternate': 7}, {'poh:start_date': 7}, {'poh:end_date': 7},\n",
       "       {'related': 6, 'product': 1}, {'product': 7}, {'related': 7},\n",
       "       {'alternate': 5, 'poh:type': 2}, {'product': 6, 'related': 1},\n",
       "       {'org:founded': 6, 'poh:start_date': 1}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 4, 'poh:type': 3},\n",
       "       {'no_relation': 7}, {'poh:type': 5, 'no_relation': 2},\n",
       "       {'related': 5, 'poh:type': 1, 'no_relation': 1},\n",
       "       {'poh:type': 5, 'no_relation': 2}, {'no_relation': 7},\n",
       "       {'no_relation': 6, 'poh:type': 1},\n",
       "       {'poh:type': 5, 'no_relation': 2},\n",
       "       {'no_relation': 5, 'related': 2},\n",
       "       {'no_relation': 5, 'poh:type': 2},\n",
       "       {'product': 6, 'no_relation': 1}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'poh:type': 4, 'no_relation': 3},\n",
       "       {'no_relation': 7}, {'no_relation': 7}, {'poh:type': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 5, 'poh:type': 2},\n",
       "       {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'poh:type': 4, 'no_relation': 3}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 5, 'poh:type': 2},\n",
       "       {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'poh:type': 5, 'no_relation': 2}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'poh:type': 6, 'no_relation': 1},\n",
       "       {'no_relation': 7}, {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'no_relation': 6, 'poh:type': 1}, {'no_relation': 7},\n",
       "       {'poh:type': 7}, {'no_relation': 7},\n",
       "       {'poh:type': 5, 'no_relation': 2}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'related': 4, 'no_relation': 3},\n",
       "       {'per:title': 7}, {'no_relation': 7}, {'org:founded': 7},\n",
       "       {'per:member_of': 6, 'product': 1}, {'related': 7},\n",
       "       {'related': 6, 'no_relation': 1},\n",
       "       {'per:member_of': 6, 'product': 1}, {'related': 7},\n",
       "       {'alternate': 4, 'no_relation': 3}, {'org:event': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 4, 'location': 3},\n",
       "       {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'poh:start_date': 6, 'no_relation': 1}, {'no_relation': 7},\n",
       "       {'no_relation': 5, 'location': 2}, {'related': 7},\n",
       "       {'no_relation': 7}, {'location': 7}, {'no_relation': 7},\n",
       "       {'poh:start_date': 7}, {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'poh:type': 7}, {'no_relation': 5, 'per:member_of': 2},\n",
       "       {'poh:type': 7}, {'no_relation': 7}, {'org:founded': 7},\n",
       "       {'product': 6, 'no_relation': 1},\n",
       "       {'per:member_of': 6, 'product': 1}, {'alternate': 7},\n",
       "       {'poh:type': 7}, {'org:field': 7}, {'no_relation': 7},\n",
       "       {'org:members': 6, 'product': 1}, {'related': 7}, {'per:title': 7},\n",
       "       {'per:member_of': 6, 'product': 1}, {'no_relation': 7},\n",
       "       {'no_relation': 6, 'org:event': 1},\n",
       "       {'no_relation': 6, 'related': 1}, {'org:field': 7}, {'related': 7},\n",
       "       {'location': 6, 'no_relation': 1}, {'location': 7},\n",
       "       {'per:title': 7}, {'related': 7}, {'product': 7},\n",
       "       {'product': 6, 'org:field': 1}, {'poh:end_date': 7},\n",
       "       {'org:members': 6, 'per:member_of': 1}, {'location': 7},\n",
       "       {'org:members': 6, 'per:member_of': 1}, {'org:founded': 7},\n",
       "       {'no_relation': 4, 'poh:type': 3}, {'poh:type': 7},\n",
       "       {'poh:type': 7}, {'no_relation': 6, 'product': 1},\n",
       "       {'no_relation': 7}, {'poh:type': 7}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'no_relation': 6, 'poh:type': 1}, {'poh:type': 7},\n",
       "       {'no_relation': 7}, {'poh:type': 7}, {'no_relation': 7},\n",
       "       {'alternate': 7}, {'alternate': 4, 'poh:type': 3},\n",
       "       {'no_relation': 7}, {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'no_relation': 6, 'per:member_of': 1}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'alternate': 4, 'no_relation': 3}, {'location': 7},\n",
       "       {'product': 7}, {'poh:start_date': 7},\n",
       "       {'no_relation': 5, 'related': 2}, {'no_relation': 7},\n",
       "       {'no_relation': 7},\n",
       "       {'org:field': 4, 'no_relation': 2, 'related': 1},\n",
       "       {'no_relation': 7}, {'no_relation': 6, 'org:event': 1},\n",
       "       {'no_relation': 7}, {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'org:event': 6, 'no_relation': 1}, {'poh:start_date': 7},\n",
       "       {'no_relation': 7}, {'product': 6, 'no_relation': 1},\n",
       "       {'poh:type': 6, 'org:field': 1}, {'no_relation': 7},\n",
       "       {'related': 5, 'no_relation': 2}, {'alternate': 4, 'poh:type': 3},\n",
       "       {'alternate': 4, 'poh:type': 3},\n",
       "       {'org:event': 3, 'no_relation': 2, 'org:field': 1, 'product': 1},\n",
       "       {'alternate': 4, 'poh:type': 3}, {'alternate': 4, 'poh:type': 3},\n",
       "       {'org:event': 3, 'product': 3, 'org:field': 1}, {'no_relation': 7},\n",
       "       {'poh:type': 7}, {'poh:type': 7}, {'product': 5, 'no_relation': 2},\n",
       "       {'no_relation': 7}, {'alternate': 7}, {'related': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'product': 4, 'per:member_of': 3},\n",
       "       {'related': 5, 'no_relation': 2},\n",
       "       {'no_relation': 4, 'poh:type': 3},\n",
       "       {'product': 6, 'no_relation': 1}, {'no_relation': 4, 'related': 3},\n",
       "       {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'no_relation': 4, 'org:field': 3},\n",
       "       {'no_relation': 4, 'poh:type': 3}, {'poh:start_date': 7},\n",
       "       {'location': 6, 'org:founded': 1},\n",
       "       {'poh:type': 4, 'no_relation': 3}, {'no_relation': 7},\n",
       "       {'product': 5, 'org:field': 1, 'no_relation': 1},\n",
       "       {'poh:start_date': 7}, {'product': 7},\n",
       "       {'no_relation': 6, 'related': 1}, {'product': 6, 'no_relation': 1},\n",
       "       {'per:title': 7}, {'location': 7}, {'related': 4, 'product': 3},\n",
       "       {'related': 7}, {'no_relation': 7},\n",
       "       {'no_relation': 5, 'product': 2}, {'product': 6, 'no_relation': 1},\n",
       "       {'per:title': 7}, {'no_relation': 7}, {'org:members': 7},\n",
       "       {'location': 7}, {'per:title': 7},\n",
       "       {'no_relation': 6, 'related': 1}, {'product': 7},\n",
       "       {'per:title': 6, 'related': 1}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'per:member_of': 5, 'location': 2},\n",
       "       {'product': 6, 'no_relation': 1},\n",
       "       {'location': 4, 'no_relation': 3},\n",
       "       {'no_relation': 4, 'related': 3}, {'alternate': 7},\n",
       "       {'per:member_of': 7}, {'per:member_of': 7}, {'no_relation': 7},\n",
       "       {'org:members': 7}, {'no_relation': 7},\n",
       "       {'no_relation': 5, 'org:members': 2},\n",
       "       {'no_relation': 6, 'related': 1}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'product': 7}, {'no_relation': 7},\n",
       "       {'per:title': 7}, {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'no_relation': 4, 'product': 3}, {'product': 7},\n",
       "       {'poh:start_date': 7}, {'product': 7}, {'related': 7},\n",
       "       {'org:field': 7}, {'product': 6, 'org:field': 1}, {'product': 7},\n",
       "       {'per:member_of': 6, 'per:title': 1}, {'org:event': 7},\n",
       "       {'related': 7}, {'location': 7}, {'alternate': 7},\n",
       "       {'alternate': 7}, {'no_relation': 4, 'poh:type': 3},\n",
       "       {'poh:type': 7}, {'poh:type': 7},\n",
       "       {'no_relation': 3, 'alternate': 2, 'poh:type': 2},\n",
       "       {'poh:type': 6, 'no_relation': 1}, {'poh:type': 7},\n",
       "       {'poh:type': 6, 'alternate': 1}, {'product': 7},\n",
       "       {'related': 6, 'product': 1}, {'poh:end_date': 6, 'related': 1},\n",
       "       {'no_relation': 6, 'poh:end_date': 1},\n",
       "       {'related': 5, 'no_relation': 2}, {'poh:start_date': 7},\n",
       "       {'no_relation': 7}, {'poh:start_date': 7}, {'poh:end_date': 7},\n",
       "       {'org:event': 6, 'no_relation': 1},\n",
       "       {'no_relation': 6, 'related': 1}, {'location': 7}, {'location': 7},\n",
       "       {'org:event': 7}, {'poh:end_date': 7}, {'location': 7},\n",
       "       {'product': 7}, {'related': 7}, {'alternate': 4, 'poh:type': 3},\n",
       "       {'poh:end_date': 7}, {'per:member_of': 6, 'product': 1},\n",
       "       {'location': 7}, {'no_relation': 7}, {'poh:start_date': 7},\n",
       "       {'poh:end_date': 7}, {'related': 7},\n",
       "       {'poh:type': 3, 'org:field': 2, 'product': 2},\n",
       "       {'poh:type': 5, 'related': 2}, {'poh:type': 4, 'alternate': 3},\n",
       "       {'poh:end_date': 7}, {'location': 7}, {'poh:start_date': 7},\n",
       "       {'no_relation': 7}, {'org:members': 6, 'product': 1},\n",
       "       {'poh:start_date': 7}, {'alternate': 7}, {'poh:start_date': 7},\n",
       "       {'alternate': 7}, {'location': 7}, {'org:field': 6, 'poh:type': 1},\n",
       "       {'poh:end_date': 7}, {'no_relation': 5, 'related': 2},\n",
       "       {'per:title': 7}, {'org:event': 7}, {'product': 7}, {'related': 7},\n",
       "       {'no_relation': 7}, {'related': 7}, {'poh:start_date': 7},\n",
       "       {'no_relation': 7}, {'product': 7}, {'poh:start_date': 7},\n",
       "       {'per:title': 7}, {'product': 7}, {'per:member_of': 7},\n",
       "       {'related': 7}, {'related': 4, 'no_relation': 3},\n",
       "       {'no_relation': 7}, {'poh:type': 7},\n",
       "       {'poh:start_date': 6, 'no_relation': 1}, {'poh:type': 7},\n",
       "       {'no_relation': 4, 'poh:type': 3},\n",
       "       {'no_relation': 4, 'poh:start_date': 3},\n",
       "       {'poh:start_date': 4, 'org:founded': 3},\n",
       "       {'org:field': 5, 'poh:type': 2}, {'no_relation': 7},\n",
       "       {'no_relation': 7}, {'no_relation': 7},\n",
       "       {'org:field': 4, 'org:members': 1, 'poh:type': 1, 'no_relation': 1},\n",
       "       {'org:field': 3, 'no_relation': 2, 'org:members': 1, 'poh:type': 1},\n",
       "       {'no_relation': 5, 'product': 2}, {'org:event': 7},\n",
       "       {'org:field': 7}, {'poh:type': 4, 'no_relation': 3},\n",
       "       {'org:founded': 7}, {'product': 7}, {'no_relation': 7}],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "for i in range(len(annotation)):\n",
    "    annotation[i] = {k:v for k,v in sorted(annotation[i].items(),key=lambda x : -x[1])}\n",
    "data['work1'] = annotation.values()\n",
    "data.work1.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 모든 데이터를 하나씩 확인하며 아래와 같이 진행함\n",
    "# 1) 최종 라벨 선정, 2) entity의 타입이 잘못된 데이터 수정 3)에러 데이터 삭제\n",
    "### 1) 7명의 작업자의 라벨링 결과 중, 라벨의 최대 투표수가 4이하인 경우 모두 확인 후 재지정 => e.g) 203번 데이터의 작업자 결과가 {no_relation:4, \"related:3\"}인 경우 회의\n",
    "### 2) 작업자가 작업 중 entity type이 잘못 지정된 경우를 직접 판별하여 회의 후 수정 => e.g) subject : 홍길동, type:org => type:per로 수정\n",
    "### 3) 어색한 문장, 부적절한 문장 등을 회의 후 제거 => 카카오 : 카카오는 69%가 가장 맛있다.(어색한 문장)\n",
    "\n",
    "#101 : <obj:서울대> 출신의 <sbj:적격엔젤투자자> 모임 등 일부 온라인 상에 노출된 적격엔젤투자자가 존재한다.\n",
    "# ~~의 per인 경우 member_of // ~~출신인 경우 현재는 아니므로 no_relation\n",
    "\n",
    "#change label\n",
    "changed_label = {22 : \"no_relation\", 131 : \"poh:type\", 258 : \"no_relation\", 469 :\"poh:type\", 763:\"poh:type\",\n",
    " 800 : \"org:founded\", 79 : \"product\", 99 : \"no_relation\", 103:\"poh:type\", 177 : \"poh:type\",\n",
    " 220 :\"no_relation\", 278 : \"related\", 294 : \"no_relation\", 330 : \"related\", 426 : \"product\", 461 : \"related\",\n",
    " 465 :\"no_relation\", 469 : \"poh:type\", 554: \"location\", 616 : \"poh:type\", 647 : \"poh:type\", 650 : \"poh:type\",\n",
    " 662 :\"per:member_of\", 664 : \"poh:type\", 669:\"org:field\", 673:\"no_relation\",733:\"poh:type\", 755:\"poh:type\"}\n",
    "\n",
    "#change type\n",
    "changed_type = {85 : \"subject poh\", 131 : \"object poh\", 258 : \"object org\", 599 : \"subject org\", 599 : \"object per\",\n",
    " 741 : \"object org\", 763 : \"subject poh\", 800 : \"subject org\", 103:\"subject poh\", 220 : \"subject poh\", 662 : \"object org\"}\n",
    "\n",
    "#delelte data\n",
    "## 문장 내에 : 이 들어간 경우 삭제\n",
    "## 비영리 단체 기관 등 org가 아니라 poh로 구분되는 데이터 삭제\n",
    "del_data = [23,49,80,82,93,124,156,174,183,185,197,201,204,207,211,238,291,396,427,473,494,495,561,595,624,648,651,652,682,718, 732,764]\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "782"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# label 변경\n",
    "for i,v in changed_label.items():\n",
    "    data.loc[i, \"class\"] = v\n",
    "\n",
    "#entity type 변경\n",
    "for i,v in changed_type.items():\n",
    "    entity, type_ = v.split(\" \")\n",
    "    if entity == \"subject\":\n",
    "        data.loc[i, \"subject_entity\"]['type'] = type_\n",
    "    else:\n",
    "        data.loc[i, \"object_entity\"]['type'] = type_\n",
    "\n",
    "#데이터 삭제\n",
    "data.drop(index=del_data,inplace=True)\n",
    "\n",
    "#데이터 제작 최종 결과물을 구글시트로 이동\n",
    "import gspread_dataframe as gd\n",
    "final_data = doc.worksheet(\"final\")\n",
    "gd.set_with_dataframe(final_data, data)\n"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "d4d1e4263499bec80672ea0156c357c1ee493ec2b1c70f0acce89fc37c4a6abe"
  },
  "kernelspec": {
   "display_name": "Python 3.8.5 64-bit ('base': conda)",
   "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": "3.8.5"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}