Googlenews_v2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 安裝套件模組"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "```\n",
    "pip install pandas\n",
    "pip install jieba\n",
    "pip install GoogleNews\n",
    "pip install matplotlib\n",
    "pip install Pillow\n",
    "pip install numpy\n",
    "pip install scipy\n",
    "pip install wordcloud\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "PYPI 對應套件簡介\n",
    "- [GoogleNews](https://pypi.org/project/GoogleNews/)\n",
    "- [jieba](https://pypi.org/project/jieba/)\n",
    "- [wordcloud](https://pypi.org/project/wordcloud/)\n",
    "- [matplotlib](https://pypi.org/project/matplotlib/)\n",
    "- [PIL](https://pypi.org/project/Pillow/)\n",
    "- [pandas](https://pypi.org/project/pandas/)\n",
    "- [numpy](https://pypi.org/project/numpy/)\n",
    "- [scipy](https://pypi.org/project/scipy/)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 匯入模組\n",
    "import pandas as pd\n",
    "import jieba\n",
    "from GoogleNews import GoogleNews\n",
    "from collections import Counter\n",
    "import matplotlib.pyplot as plt\n",
    "from datetime import datetime\n",
    "from PIL import Image\n",
    "import numpy as np\n",
    "from scipy.ndimage import gaussian_gradient_magnitude\n",
    "from wordcloud import WordCloud, ImageColorGenerator\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "需要知道的觀念和套件：\n",
    "\n",
    "1. Pandas 資料處理\n",
    "2. GoogleNews 爬蟲套件\n",
    "3. WordCloud 文字雲套件\n",
    "4. Image.open 讀取圖片\n",
    "5. numpy 陣列操作\n",
    "6. jieba 斷詞套件\n",
    "7. datetime 日期處理\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 初始化 Google News 物件"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Google News初始化設定\n",
    "googlenews = GoogleNews(lang='zh-tw', region='TW', encode='utf-8')\n",
    "\n",
    "# 啟用除錯模式 try except\n",
    "googlenews.enableException(True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<GoogleNews.GoogleNews object at 0x14b6d68f0>\n",
      "<class 'GoogleNews.GoogleNews'>\n",
      "object 可以使用的屬性跟方法\n",
      "['_GoogleNews__encode', '_GoogleNews__end', '_GoogleNews__exception', '_GoogleNews__lang', '_GoogleNews__links', '_GoogleNews__period', '_GoogleNews__results', '_GoogleNews__section', '_GoogleNews__start', '_GoogleNews__texts', '_GoogleNews__topic', '_GoogleNews__totalcount', '_GoogleNews__version', '__class__', '__delattr__', '__dict__', '__dir__', '__doc__', '__eq__', '__format__', '__ge__', '__getattribute__', '__gt__', '__hash__', '__init__', '__init_subclass__', '__le__', '__lt__', '__module__', '__ne__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__setattr__', '__sizeof__', '__str__', '__subclasshook__', '__weakref__', 'build_response', 'clear', 'enableException', 'getVersion', 'get_links', 'get_news', 'get_page', 'get_texts', 'getpage', 'gettext', 'headers', 'page_at', 'remove_after_last_fullstop', 'result', 'results', 'search', 'setTimeRange', 'set_encode', 'set_lang', 'set_period', 'set_section', 'set_time_range', 'set_topic', 'setencode', 'setlang', 'setperiod', 'total_count', 'user_agent']\n"
     ]
    }
   ],
   "source": [
    "object = GoogleNews()\n",
    "\n",
    "print(object)\n",
    "print(type(object))\n",
    "print(\"object 可以使用的屬性跟方法\")\n",
    "print(dir(object))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "utf-8\n",
      "{'_GoogleNews__texts': [], '_GoogleNews__links': [], '_GoogleNews__results': [], '_GoogleNews__totalcount': 0, 'user_agent': 'Mozilla/5.0 (X11; Ubuntu; Linux i686; rv:64.0) Gecko/20100101 Firefox/64.0', '_GoogleNews__lang': 'en', 'headers': {'User-Agent': 'Mozilla/5.0 (X11; Ubuntu; Linux i686; rv:64.0) Gecko/20100101 Firefox/64.0'}, '_GoogleNews__period': '', '_GoogleNews__start': '', '_GoogleNews__end': '', '_GoogleNews__encode': 'utf-8', '_GoogleNews__exception': False, '_GoogleNews__version': '1.6.15', '_GoogleNews__topic': None, '_GoogleNews__section': None}\n"
     ]
    }
   ],
   "source": [
    "object.set_encode('utf-8')\n",
    "print(object._GoogleNews__encode)\n",
    "print(object.__dict__)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<GoogleNews.GoogleNews object at 0x123a83d70> 物件 記憶體位置\n",
    "\n",
    "<class 'GoogleNews.GoogleNews'> 哪個類別 資料型態\n",
    "\n",
    "\n",
    "{'_GoogleNews__texts': [], '_GoogleNews__links': [], '_GoogleNews__results': [], '_GoogleNews__totalcount': 0, 'user_agent': 'Mozilla/5.0 (X11; Ubuntu; Linux i686; rv:64.0) Gecko/20100101 Firefox/64.0', '_GoogleNews__lang': 'en', 'headers': {'User-Agent': 'Mozilla/5.0 (X11; Ubuntu; Linux i686; rv:64.0) Gecko/20100101 Firefox/64.0'}, '_GoogleNews__period': '', '_GoogleNews__start': '', '_GoogleNews__end': '', '_GoogleNews__encode': 'utf-8', '_GoogleNews__exception': False, '_GoogleNews__version': '1.6.15', '_GoogleNews__topic': None, '_GoogleNews__section': None} 物件的屬性\n",
    "\n",
    "_ 底線開頭的，類似私有屬性，不希望被外部直接訪問，用方法存取"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "物件的方法\n",
    "\n",
    "build_response', 'clear', 'enableException', 'getVersion', 'get_links', 'get_news', 'get_page', 'get_texts', 'getpage', 'gettext', 'headers', 'page_at', 'remove_after_last_fullstop', 'result', 'results', 'search', 'setTimeRange', 'set_encode', 'set_lang', 'set_period', 'set_section', 'set_time_range', 'set_topic', 'setencode', 'setlang', 'setperiod', 'total_count', 'user_agent'\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 使用者輸入關鍵字"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "# , period='1d'\n",
    "set_time = input(\"請輸入要搜尋的時間 (1d, 7d, 30d): \")\n",
    "googlenews.set_period(set_time)\n",
    "# 使用者輸入關鍵字\n",
    "keyword = input(\"請輸入要搜尋的關鍵字: \")\n",
    "# keyword = \"柯文哲\"\n",
    "googlenews.search(keyword)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "標題: 颱風生成偏少?至今僅18個 氣象專家:正常範圍偏少 ｜十點不一樣20241002 @TVBSNEWS01\n"
     ]
    }
   ],
   "source": [
    "# test\n",
    "result = googlenews.result()\n",
    "print(f\"標題: {result[0]['title']}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "# 顯示搜尋結果"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "標題: 颱風生成偏少?至今僅18個 氣象專家:正常範圍偏少 ｜十點不一樣20241002 @TVBSNEWS01\n",
      "連結: https://www.youtube.com/watch%3Fv%3DX_0gECybLbk&ved=2ahUKEwikoYyR9--IAxWYkq8BHdAiPPcQxfQBegQIBBAC&usg=AOvVaw1UIrEyr8dS0R581v4p5UZG\n",
      "---\n",
      "標題: 颱風天物資搶光！這牌鮮奶滯銷 網：感謝台灣人沒忘記過去\n",
      "連結: https://news.ltn.com.tw/news/life/breakingnews/4818510&ved=2ahUKEwikoYyR9--IAxWYkq8BHdAiPPcQxfQBegQICRAC&usg=AOvVaw0jYAhx4Dti8hvHEGiXKWGr\n",
      "---\n",
      "標題: 【吳宇舒晚報】 【山陀兒颱風】颱風擾郵輪行！說好會員免費搭旅客晚報到控遭丟包2024／10／02@newsebc\n",
      "連結: https://www.youtube.com/watch%3Fv%3DDusuLXJpp-0&ved=2ahUKEwikoYyR9--IAxWYkq8BHdAiPPcQxfQBegQIAxAC&usg=AOvVaw3_bgpYQOBJYOr4v7d405Xn\n",
      "---\n",
      "標題: 台中明颱風假！多數醫院正常門診 中榮開颱風特別門診\n",
      "連結: https://health.ltn.com.tw/article/breakingnews/4818602&ved=2ahUKEwikoYyR9--IAxWYkq8BHdAiPPcQxfQBegQIBRAC&usg=AOvVaw0_8oFFcBp95HZNsOq50L3t\n",
      "---\n",
      "標題: 山陀兒颱風3日又要轉彎？日本預測「回馬槍再襲南部」，民間傳說恐破功\n",
      "連結: https://www.storm.mg/lifestyle/5248147&ved=2ahUKEwikoYyR9--IAxWYkq8BHdAiPPcQxfQBegQIBhAC&usg=AOvVaw3PMf4Sz_hCkBjCcv2fUz1e\n",
      "---\n",
      "標題: 【山陀兒颱風】綠扯颱風假救樑 氣象署曝原估有風雨 4縣市顧及市民安全 @newsebc\n",
      "連結: https://www.youtube.com/watch%3Fv%3D-vuHQm9Hczw&ved=2ahUKEwikoYyR9--IAxWYkq8BHdAiPPcQxfQBegQIARAC&usg=AOvVaw2-zAT6apC3N0juIsjBXxwn\n",
      "---\n",
      "標題: 山陀兒襲台「連放颱風假」 恐衝擊薪水發放！\n",
      "連結: https://www.ettoday.net/news/20241002/2828162.htm&ved=2ahUKEwikoYyR9--IAxWYkq8BHdAiPPcQxfQBegQIABAC&usg=AOvVaw3Xp7JrUnwFRV3Ow5RFckwD\n",
      "---\n",
      "標題: 【吳宇舒晚報】 【山陀兒颱風】獨家》「山陀兒」牛步進逼蘭嶼擱淺貨輪慘斷兩截2024／10／02@newsebc\n",
      "連結: https://www.youtube.com/watch%3Fv%3DkTiF8fM32fc&ved=2ahUKEwikoYyR9--IAxWYkq8BHdAiPPcQxfQBegQIAhAC&usg=AOvVaw3lhlz8hkp54f9YGltRQ57C\n",
      "---\n",
      "標題: 山陀兒準備「進家門」！颱風核心遭陸地摩擦、持續變弱 專家急籲：不可輕忽\n",
      "連結: https://tw.news.yahoo.com/%25E5%25B1%25B1%25E9%2599%2580%25E5%2585%2592%25E6%25BA%2596%25E5%2582%2599-%25E9%2580%25B2%25E5%25AE%25B6%25E9%2596%2580-%25E9%25A2%25B1%25E9%25A2%25A8%25E6%25A0%25B8%25E5%25BF%2583%25E9%2581%25AD%25E9%2599%25B8%25E5%259C%25B0%25E6%2591%25A9%25E6%2593%25A6-%25E6%258C%2581%25E7%25BA%258C%25E8%25AE%258A%25E5%25BC%25B1-%25E5%25B0%2588%25E5%25AE%25B6%25E6%2580%25A5%25E7%25B1%25B2-134502022.html&ved=2ahUKEwikoYyR9--IAxWYkq8BHdAiPPcQxfQBegQICBAC&usg=AOvVaw2lcQb2O9hPgagkWcIeCrYb\n",
      "---\n",
      "標題: 東港迎王颱風攪局 千歲爺指示王船法會等科儀按表進行\n",
      "連結: https://udn.com/news/story/7327/8266662&ved=2ahUKEwikoYyR9--IAxWYkq8BHdAiPPcQxfQBegQIBxAC&usg=AOvVaw27fVFwUiyclgksJylxWdmg\n",
      "---\n"
     ]
    }
   ],
   "source": [
    "# 獲取搜尋結果\n",
    "result = googlenews.result()\n",
    "\n",
    "# 顯示新聞標題和連結\n",
    "for item in result:\n",
    "    print(f\"標題: {item['title']}\")\n",
    "    print(f\"連結: {item['link']}\")\n",
    "    print('---')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 儲存成 Json 格式檔案"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "搜尋結果已儲存為: ./json_output/20241002_颱風_7d.json\n"
     ]
    }
   ],
   "source": [
    "# <!-- 儲存至 Json -->\n",
    "import json\n",
    "\n",
    "from datetime import datetime\n",
    "\n",
    "# 獲取當前日期\n",
    "today = datetime.now().strftime('%Y%m%d')\n",
    "\n",
    "# 生成文件名\n",
    "filename = f\"./json_output/{today}_{keyword}_{set_time}.json\"\n",
    "\n",
    "# 將 result 寫入 JSON 文件\n",
    "with open(filename, 'w', encoding='utf-8') as f:\n",
    "    json.dump(result, f, ensure_ascii=False, indent=4)\n",
    "\n",
    "print(f\"搜尋結果已儲存為: {filename}\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 讀取靜態 Json 格式檔案"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[{'title': '颱風生成偏少?至今僅18個 氣象專家:正常範圍偏少 ｜十點不一樣20241002 @TVBSNEWS01', 'media': 'YouTube', 'date': '2 分鐘前', 'datetime': nan, 'desc': '歷史紀錄的平均值，西北太平洋生成颱風數26.5個，但今年都已經來到10月份了，山陀兒算是第18個，似乎有點偏少，引起各方關注，而且颱風更是從8月份開始像在趕進度似的，...', 'link': 'https://www.youtube.com/watch%3Fv%3DX_0gECybLbk&ved=2ahUKEwikoYyR9--IAxWYkq8BHdAiPPcQxfQBegQIBBAC&usg=AOvVaw1UIrEyr8dS0R581v4p5UZG', 'img': 'data:image/gif;base64,R0lGODlhAQABAIAAAP///////yH5BAEKAAEALAAAAAABAAEAAAICTAEAOw=='}, {'title': '颱風天物資搶光！這牌鮮奶滯銷 網：感謝台灣人沒忘記過去', 'media': '自由時報', 'date': '8 分鐘前', 'datetime': nan, 'desc': '吳裕堯／核稿編輯社群平台Threads一則貼文掀起熱烈討論，文中附上的超市鮮奶區照片顯示，颱風天前夕物資被搶購一空，味全鮮奶依然滯銷，引發眾多網友留言感謝大家沒有忘記...', 'link': 'https://news.ltn.com.tw/news/life/breakingnews/4818510&ved=2ahUKEwikoYyR9--IAxWYkq8BHdAiPPcQxfQBegQICRAC&usg=AOvVaw0jYAhx4Dti8hvHEGiXKWGr', 'img': 'data:image/gif;base64,R0lGODlhAQABAIAAAP///////yH5BAEKAAEALAAAAAABAAEAAAICTAEAOw=='}, {'title': '【吳宇舒晚報】 【山陀兒颱風】颱風擾郵輪行！說好會員免費搭旅客晚報到控遭丟包2024／10／02@newsebc', 'media': 'YouTube', 'date': '2 分鐘前', 'datetime': nan, 'desc': '【山陀兒颱風】旗津風大雨大浪大民眾穿越封鎖線冒險觀浪【山陀兒颱風】抗戰防淹水！銀行水閘門增高ATM被迫暫停服務【山陀兒颱風】颱風擾郵輪行！說好會員免費搭旅客晚報...', 'link': 'https://www.youtube.com/watch%3Fv%3DDusuLXJpp-0&ved=2ahUKEwikoYyR9--IAxWYkq8BHdAiPPcQxfQBegQIAxAC&usg=AOvVaw3_bgpYQOBJYOr4v7d405Xn', 'img': 'data:image/gif;base64,R0lGODlhAQABAIAAAP///////yH5BAEKAAEALAAAAAABAAEAAAICTAEAOw=='}, {'title': '台中明颱風假！多數醫院正常門診 中榮開颱風特別門診', 'media': '自由健康網', 'date': '6 分鐘前', 'datetime': nan, 'desc': '山陀兒颱風逼近，因氣象署預估台中市明日符合停班課標準，台中市晚間8點宣佈明天停止上班上課，台中將連放2天颱風假，包括中國醫藥大學附設醫院及台中醫院均門診正常服務，...', 'link': 'https://health.ltn.com.tw/article/breakingnews/4818602&ved=2ahUKEwikoYyR9--IAxWYkq8BHdAiPPcQxfQBegQIBRAC&usg=AOvVaw0_8oFFcBp95HZNsOq50L3t', 'img': 'data:image/gif;base64,R0lGODlhAQABAIAAAP///////yH5BAEKAAEALAAAAAABAAEAAAICTAEAOw=='}, {'title': '山陀兒颱風3日又要轉彎？日本預測「回馬槍再襲南部」，民間傳說恐破功', 'media': '風傳媒', 'date': '8 分鐘前', 'datetime': nan, 'desc': '山陀兒颱風滯留台灣西南部許久，今（2）日終於動起來北上，雖然中心還是有些許滯留狀況，不過暴風圈已進入雲林以南及臺東陸地，對新竹以南、南投、宜蘭、花蓮、臺東地區及...', 'link': 'https://www.storm.mg/lifestyle/5248147&ved=2ahUKEwikoYyR9--IAxWYkq8BHdAiPPcQxfQBegQIBhAC&usg=AOvVaw3PMf4Sz_hCkBjCcv2fUz1e', 'img': 'data:image/gif;base64,R0lGODlhAQABAIAAAP///////yH5BAEKAAEALAAAAAABAAEAAAICTAEAOw=='}, {'title': '【山陀兒颱風】綠扯颱風假救樑 氣象署曝原估有風雨 4縣市顧及市民安全 @newsebc', 'media': 'YouTube', 'date': '0 分鐘前', 'datetime': nan, 'desc': '颱風山陀兒影響持續，基北北桃宣布10月2號放颱風假，但這個決定，是否達「放假條件」引發討論，也讓綠營質疑是不是趁機救謝國樑？對此，台北市長蔣萬安還原決策過程，...', 'link': 'https://www.youtube.com/watch%3Fv%3D-vuHQm9Hczw&ved=2ahUKEwikoYyR9--IAxWYkq8BHdAiPPcQxfQBegQIARAC&usg=AOvVaw2-zAT6apC3N0juIsjBXxwn', 'img': 'data:image/gif;base64,R0lGODlhAQABAIAAAP///////yH5BAEKAAEALAAAAAABAAEAAAICTAEAOw=='}, {'title': '山陀兒襲台「連放颱風假」 恐衝擊薪水發放！', 'media': 'ETtoday新聞雲', 'date': '4 分鐘前', 'datetime': nan, 'desc': '颱風「山陀兒」怠速前進，目前會不會連續放2到3天以上颱風假還很難說，但有民間單位苦惱，這樣下去5日薪資可能無法如期發放。位於高雄的「方舟就業服務協會」理事長王姿琦...', 'link': 'https://www.ettoday.net/news/20241002/2828162.htm&ved=2ahUKEwikoYyR9--IAxWYkq8BHdAiPPcQxfQBegQIABAC&usg=AOvVaw3Xp7JrUnwFRV3Ow5RFckwD', 'img': 'data:image/gif;base64,R0lGODlhAQABAIAAAP///////yH5BAEKAAEALAAAAAABAAEAAAICTAEAOw=='}, {'title': '【吳宇舒晚報】 【山陀兒颱風】獨家》「山陀兒」牛步進逼蘭嶼擱淺貨輪慘斷兩截2024／10／02@newsebc', 'media': 'YouTube', 'date': '7 分鐘前', 'datetime': nan, 'desc': '【山陀兒颱風】獨家》「山陀兒」牛步進逼蘭嶼擱淺貨輪慘斷兩截【山陀兒颱風】急搶修！枋寮狂浪沖垮堤防30尺缺口釀海水倒灌【山陀兒颱風】山陀兒挾強風豪雨襲台東卑南...', 'link': 'https://www.youtube.com/watch%3Fv%3DkTiF8fM32fc&ved=2ahUKEwikoYyR9--IAxWYkq8BHdAiPPcQxfQBegQIAhAC&usg=AOvVaw3lhlz8hkp54f9YGltRQ57C', 'img': 'data:image/gif;base64,R0lGODlhAQABAIAAAP///////yH5BAEKAAEALAAAAAABAAEAAAICTAEAOw=='}, {'title': '山陀兒準備「進家門」！颱風核心遭陸地摩擦、持續變弱 專家急籲：不可輕忽', 'media': 'Yahoo新聞', 'date': '7 分鐘前', 'datetime': nan, 'desc': '記者高珞曦／綜合報導. 中颱山陀兒即將登陸，2日傍晚起風雨逐漸明顯，氣象粉專「台灣颱風論壇｜天氣特急」表示，山陀兒因移動速度緩慢，東側核心被陸地摩擦、強度減弱，但其...', 'link': 'https://tw.news.yahoo.com/%25E5%25B1%25B1%25E9%2599%2580%25E5%2585%2592%25E6%25BA%2596%25E5%2582%2599-%25E9%2580%25B2%25E5%25AE%25B6%25E9%2596%2580-%25E9%25A2%25B1%25E9%25A2%25A8%25E6%25A0%25B8%25E5%25BF%2583%25E9%2581%25AD%25E9%2599%25B8%25E5%259C%25B0%25E6%2591%25A9%25E6%2593%25A6-%25E6%258C%2581%25E7%25BA%258C%25E8%25AE%258A%25E5%25BC%25B1-%25E5%25B0%2588%25E5%25AE%25B6%25E6%2580%25A5%25E7%25B1%25B2-134502022.html&ved=2ahUKEwikoYyR9--IAxWYkq8BHdAiPPcQxfQBegQICBAC&usg=AOvVaw2lcQb2O9hPgagkWcIeCrYb', 'img': 'data:image/gif;base64,R0lGODlhAQABAIAAAP///////yH5BAEKAAEALAAAAAABAAEAAAICTAEAOw=='}, {'title': '東港迎王颱風攪局 千歲爺指示王船法會等科儀按表進行', 'media': '聯合新聞網', 'date': '1 小時前', 'datetime': nan, 'desc': '三年一科的東港迎王平安祭典遇到山陀兒颱風攪局，東港東隆宮今晚聲明，經向代天巡狩千歲爺擲杯請示，代天府內各項科儀將依祭典順...', 'link': 'https://udn.com/news/story/7327/8266662&ved=2ahUKEwikoYyR9--IAxWYkq8BHdAiPPcQxfQBegQIBxAC&usg=AOvVaw27fVFwUiyclgksJylxWdmg', 'img': 'data:image/gif;base64,R0lGODlhAQABAIAAAP///////yH5BAEKAAEALAAAAAABAAEAAAICTAEAOw=='}]\n",
      "<class 'list'>\n"
     ]
    }
   ],
   "source": [
    "import json\n",
    "# filename = \"./json_output/20240919_柯文哲_15d.json\"\n",
    "# 讀取 JSON 文件\n",
    "with open(filename, 'r', encoding='utf-8') as f:\n",
    "    loaded_result = json.load(f)\n",
    "\n",
    "print(loaded_result)\n",
    "print(type(loaded_result))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['颱風生成偏少?至今僅18個 氣象專家:正常範圍偏少 ｜十點不一樣20241002 @TVBSNEWS01', '颱風天物資搶光！這牌鮮奶滯銷 網：感謝台灣人沒忘記過去', '【吳宇舒晚報】 【山陀兒颱風】颱風擾郵輪行！說好會員免費搭旅客晚報到控遭丟包2024／10／02@newsebc', '台中明颱風假！多數醫院正常門診 中榮開颱風特別門診', '山陀兒颱風3日又要轉彎？日本預測「回馬槍再襲南部」，民間傳說恐破功', '【山陀兒颱風】綠扯颱風假救樑 氣象署曝原估有風雨 4縣市顧及市民安全 @newsebc', '山陀兒襲台「連放颱風假」 恐衝擊薪水發放！', '【吳宇舒晚報】 【山陀兒颱風】獨家》「山陀兒」牛步進逼蘭嶼擱淺貨輪慘斷兩截2024／10／02@newsebc', '山陀兒準備「進家門」！颱風核心遭陸地摩擦、持續變弱 專家急籲：不可輕忽', '東港迎王颱風攪局 千歲爺指示王船法會等科儀按表進行']\n",
      "['歷史紀錄的平均值，西北太平洋生成颱風數26.5個，但今年都已經來到10月份了，山陀兒算是第18個，似乎有點偏少，引起各方關注，而且颱風更是從8月份開始像在趕進度似的，...', '吳裕堯／核稿編輯社群平台Threads一則貼文掀起熱烈討論，文中附上的超市鮮奶區照片顯示，颱風天前夕物資被搶購一空，味全鮮奶依然滯銷，引發眾多網友留言感謝大家沒有忘記...', '【山陀兒颱風】旗津風大雨大浪大民眾穿越封鎖線冒險觀浪【山陀兒颱風】抗戰防淹水！銀行水閘門增高ATM被迫暫停服務【山陀兒颱風】颱風擾郵輪行！說好會員免費搭旅客晚報...', '山陀兒颱風逼近，因氣象署預估台中市明日符合停班課標準，台中市晚間8點宣佈明天停止上班上課，台中將連放2天颱風假，包括中國醫藥大學附設醫院及台中醫院均門診正常服務，...', '山陀兒颱風滯留台灣西南部許久，今（2）日終於動起來北上，雖然中心還是有些許滯留狀況，不過暴風圈已進入雲林以南及臺東陸地，對新竹以南、南投、宜蘭、花蓮、臺東地區及...', '颱風山陀兒影響持續，基北北桃宣布10月2號放颱風假，但這個決定，是否達「放假條件」引發討論，也讓綠營質疑是不是趁機救謝國樑？對此，台北市長蔣萬安還原決策過程，...', '颱風「山陀兒」怠速前進，目前會不會連續放2到3天以上颱風假還很難說，但有民間單位苦惱，這樣下去5日薪資可能無法如期發放。位於高雄的「方舟就業服務協會」理事長王姿琦...', '【山陀兒颱風】獨家》「山陀兒」牛步進逼蘭嶼擱淺貨輪慘斷兩截【山陀兒颱風】急搶修！枋寮狂浪沖垮堤防30尺缺口釀海水倒灌【山陀兒颱風】山陀兒挾強風豪雨襲台東卑南...', '記者高珞曦／綜合報導. 中颱山陀兒即將登陸，2日傍晚起風雨逐漸明顯，氣象粉專「台灣颱風論壇｜天氣特急」表示，山陀兒因移動速度緩慢，東側核心被陸地摩擦、強度減弱，但其...', '三年一科的東港迎王平安祭典遇到山陀兒颱風攪局，東港東隆宮今晚聲明，經向代天巡狩千歲爺擲杯請示，代天府內各項科儀將依祭典順...']\n",
      "['颱風生成偏少?至今僅18個 氣象專家:正常範圍偏少 ｜十點不一樣20241002 @TVBSNEWS01 歷史紀錄的平均值，西北太平洋生成颱風數26.5個，但今年都已經來到10月份了，山陀兒算是第18個，似乎有點偏少，引起各方關注，而且颱風更是從8月份開始像在趕進度似的，...', '颱風天物資搶光！這牌鮮奶滯銷 網：感謝台灣人沒忘記過去 吳裕堯／核稿編輯社群平台Threads一則貼文掀起熱烈討論，文中附上的超市鮮奶區照片顯示，颱風天前夕物資被搶購一空，味全鮮奶依然滯銷，引發眾多網友留言感謝大家沒有忘記...', '【吳宇舒晚報】 【山陀兒颱風】颱風擾郵輪行！說好會員免費搭旅客晚報到控遭丟包2024／10／02@newsebc 【山陀兒颱風】旗津風大雨大浪大民眾穿越封鎖線冒險觀浪【山陀兒颱風】抗戰防淹水！銀行水閘門增高ATM被迫暫停服務【山陀兒颱風】颱風擾郵輪行！說好會員免費搭旅客晚報...', '台中明颱風假！多數醫院正常門診 中榮開颱風特別門診 山陀兒颱風逼近，因氣象署預估台中市明日符合停班課標準，台中市晚間8點宣佈明天停止上班上課，台中將連放2天颱風假，包括中國醫藥大學附設醫院及台中醫院均門診正常服務，...', '山陀兒颱風3日又要轉彎？日本預測「回馬槍再襲南部」，民間傳說恐破功 山陀兒颱風滯留台灣西南部許久，今（2）日終於動起來北上，雖然中心還是有些許滯留狀況，不過暴風圈已進入雲林以南及臺東陸地，對新竹以南、南投、宜蘭、花蓮、臺東地區及...', '【山陀兒颱風】綠扯颱風假救樑 氣象署曝原估有風雨 4縣市顧及市民安全 @newsebc 颱風山陀兒影響持續，基北北桃宣布10月2號放颱風假，但這個決定，是否達「放假條件」引發討論，也讓綠營質疑是不是趁機救謝國樑？對此，台北市長蔣萬安還原決策過程，...', '山陀兒襲台「連放颱風假」 恐衝擊薪水發放！ 颱風「山陀兒」怠速前進，目前會不會連續放2到3天以上颱風假還很難說，但有民間單位苦惱，這樣下去5日薪資可能無法如期發放。位於高雄的「方舟就業服務協會」理事長王姿琦...', '【吳宇舒晚報】 【山陀兒颱風】獨家》「山陀兒」牛步進逼蘭嶼擱淺貨輪慘斷兩截2024／10／02@newsebc 【山陀兒颱風】獨家》「山陀兒」牛步進逼蘭嶼擱淺貨輪慘斷兩截【山陀兒颱風】急搶修！枋寮狂浪沖垮堤防30尺缺口釀海水倒灌【山陀兒颱風】山陀兒挾強風豪雨襲台東卑南...', '山陀兒準備「進家門」！颱風核心遭陸地摩擦、持續變弱 專家急籲：不可輕忽 記者高珞曦／綜合報導. 中颱山陀兒即將登陸，2日傍晚起風雨逐漸明顯，氣象粉專「台灣颱風論壇｜天氣特急」表示，山陀兒因移動速度緩慢，東側核心被陸地摩擦、強度減弱，但其...', '東港迎王颱風攪局 千歲爺指示王船法會等科儀按表進行 三年一科的東港迎王平安祭典遇到山陀兒颱風攪局，東港東隆宮今晚聲明，經向代天巡狩千歲爺擲杯請示，代天府內各項科儀將依祭典順...']\n"
     ]
    }
   ],
   "source": [
    "# 查看資料 資料分類 資料合併 練習\n",
    "\n",
    "title_tokens = []\n",
    "desc_tokens = []\n",
    "\n",
    "all_content = []\n",
    "\n",
    "for n in range(len(loaded_result)):\n",
    "\n",
    "    title = loaded_result[n]['title']\n",
    "    desc = loaded_result[n]['desc']  \n",
    "    title_tokens.append(loaded_result[n]['title'])\n",
    "    desc_tokens.append(loaded_result[n]['desc'])\n",
    "    all_content.append(title + \" \" + desc)\n",
    "\n",
    "print(title_tokens)\n",
    "print(desc_tokens)\n",
    "print(all_content)\n",
    "# for item in all_content:\n",
    "#     print(\"\".join(item))\n",
    "    \n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 停用詞過濾"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "# ckip-transformers  0.3.4 "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Initializing drivers ... WS\n",
      "Initializing drivers ... POS\n",
      "Initializing drivers ... NER\n",
      "Initializing drivers ... done\n",
      "颱風生成偏少至今僅18個 氣象專家正常範圍偏少 十點不一樣20241002 TVBSNEWS01 歷史紀錄的平均值西北太平洋生成颱風數265個但今年都已經來到10月份了山陀兒算是第18個似乎有點偏少引起各方關注而且颱風更是從8月份開始像在趕進度似的\n",
      "颱風天物資搶光這牌鮮奶滯銷 網感謝台灣人沒忘記過去 吳裕堯核稿編輯社群平台Threads一則貼文掀起熱烈討論文中附上的超市鮮奶區照片顯示颱風天前夕物資被搶購一空味全鮮奶依然滯銷引發眾多網友留言感謝大家沒有忘記\n",
      "吳宇舒晚報 山陀兒颱風颱風擾郵輪行說好會員免費搭旅客晚報到控遭丟包20241002newsebc 山陀兒颱風旗津風大雨大浪大民眾穿越封鎖線冒險觀浪山陀兒颱風抗戰防淹水銀行水閘門增高ATM被迫暫停服務山陀兒颱風颱風擾郵輪行說好會員免費搭旅客晚報\n",
      "台中明颱風假多數醫院正常門診 中榮開颱風特別門診 山陀兒颱風逼近因氣象署預估台中市明日符合停班課標準台中市晚間8點宣佈明天停止上班上課台中將連放2天颱風假包括中國醫藥大學附設醫院及台中醫院均門診正常服務\n",
      "山陀兒颱風3日又要轉彎日本預測回馬槍再襲南部民間傳說恐破功 山陀兒颱風滯留台灣西南部許久今2日終於動起來北上雖然中心還是有些許滯留狀況不過暴風圈已進入雲林以南及臺東陸地對新竹以南南投宜蘭花蓮臺東地區及\n",
      "山陀兒颱風綠扯颱風假救樑 氣象署曝原估有風雨 4縣市顧及市民安全 newsebc 颱風山陀兒影響持續基北北桃宣布10月2號放颱風假但這個決定是否達放假條件引發討論也讓綠營質疑是不是趁機救謝國樑對此台北市長蔣萬安還原決策過程\n",
      "山陀兒襲台連放颱風假 恐衝擊薪水發放 颱風山陀兒怠速前進目前會不會連續放2到3天以上颱風假還很難說但有民間單位苦惱這樣下去5日薪資可能無法如期發放位於高雄的方舟就業服務協會理事長王姿琦\n",
      "吳宇舒晚報 山陀兒颱風獨家山陀兒牛步進逼蘭嶼擱淺貨輪慘斷兩截20241002newsebc 山陀兒颱風獨家山陀兒牛步進逼蘭嶼擱淺貨輪慘斷兩截山陀兒颱風急搶修枋寮狂浪沖垮堤防30尺缺口釀海水倒灌山陀兒颱風山陀兒挾強風豪雨襲台東卑南\n",
      "山陀兒準備進家門颱風核心遭陸地摩擦持續變弱 專家急籲不可輕忽 記者高珞曦綜合報導 中颱山陀兒即將登陸2日傍晚起風雨逐漸明顯氣象粉專台灣颱風論壇天氣特急表示山陀兒因移動速度緩慢東側核心被陸地摩擦強度減弱但其\n",
      "東港迎王颱風攪局 千歲爺指示王船法會等科儀按表進行 三年一科的東港迎王平安祭典遇到山陀兒颱風攪局東港東隆宮今晚聲明經向代天巡狩千歲爺擲杯請示代天府內各項科儀將依祭典順\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Tokenization: 100%|██████████| 10/10 [00:00<00:00, 10637.34it/s]\n",
      "Inference: 100%|██████████| 1/1 [00:00<00:00,  1.63it/s]\n",
      "Tokenization: 100%|██████████| 10/10 [00:00<00:00, 19608.71it/s]\n",
      "Inference: 100%|██████████| 1/1 [00:00<00:00,  2.17it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "總詞性數: 547\n",
      "句子 0:['颱風', '生成', '偏', '少', '至今', '僅', '18', '個', ' ', '氣象', '專家', '正常', '範圍', '偏', '少', ' ', '十', '點', '不', '一樣', '20241002', ' TVBSNEWS01 ', '歷史', '紀錄', '的', '平均值', '西北', '太平洋', '生成', '颱風', '數', '265', '個', '但', '今年', '都', '已經', '來到', '10月份', '了', '山陀兒', '算是', '第18', '個', '似乎', '有點', '偏', '少', '引起', '各', '方', '關注', '而且', '颱風', '更', '是', '從', '8月份', '開始', '像', '在', '趕', '進度', '似的']\n",
      "句子 1:['颱風天', '物資', '搶光', '這', '牌', '鮮奶', '滯銷', ' 網', '感謝', '台灣人', '沒', '忘記', '過去', ' ', '吳裕堯', '核稿', '編輯', '社群', '平台', 'Threads', '一', '則', '貼文', '掀起', '熱烈', '討論', '文', '中', '附上', '的', '超市', '鮮奶區', '照片', '顯示', '颱風天', '前夕', '物資', '被', '搶購', '一', '空', '味全', '鮮奶', '依然', '滯銷', '引發', '眾多', '網友', '留言', '感謝', '大家', '沒有', '忘記']\n",
      "句子 2:['吳宇舒', '晚報', ' ', '山陀兒', '颱風', '颱風', '擾', '郵輪行', '說好', '會員', '免費', '搭', '旅客', '晚報', '到', '控', '遭', '丟包', '20241002', 'newsebc ', '山陀兒', '颱風', '旗津', '風', '大雨', '大浪', '大', '民眾', '穿越', '封鎖線', '冒險', '觀浪', '山陀兒', '颱風', '抗戰', '防', '淹水', '銀行', '水閘門', '增高', 'ATM', '被迫', '暫停', '服務', '山陀兒', '颱風', '颱風', '擾', '郵輪行', '說好', '會員', '免費', '搭', '旅客', '晚報']\n",
      "句子 3:['台中', '明', '颱風假', '多數', '醫院', '正常', '門診', ' ', '中榮開', '颱風', '特別', '門診', ' ', '山陀兒', '颱風', '逼近', '因', '氣象署', '預估', '台中市', '明日', '符合', '停班', '課', '標準', '台中市', '晚間', '8點', '宣佈', '明天', '停止', '上班', '上課', '台中', '將', '連', '放', '2', '天', '颱風假', '包括', '中國', '醫藥', '大學', '附設', '醫院', '及', '台中', '醫院', '均', '門診', '正常', '服務']\n",
      "句子 4:['山陀兒', '颱風', '3日', '又', '要', '轉彎', '日本', '預測', '回馬槍', '再', '襲', '南部', '民間', '傳說', '恐', '破功', ' ', '山陀兒', '颱風', '滯留', '台灣', '西南部', '許久', '今', '2日', '終於', '動', '起來', '北上', '雖然', '中心', '還是', '有', '些許', '滯留', '狀況', '不過', '暴風圈', '已', '進入', '雲林', '以南', '及', '臺東', '陸地', '對', '新竹', '以南', '南投', '宜蘭', '花蓮', '臺東', '地區', '及']\n",
      "句子 5:['山陀兒', '颱風', '綠', '扯', '颱風假', '救', '樑', ' ', '氣象署', '曝', '原', '估', '有', '風雨', ' 4', '縣市', '顧及', '市民', '安全', ' newsebc ', '颱風', '山陀兒', '影響', '持續', '基', '北北桃', '宣布', '10月', '2號', '放', '颱風假', '但', '這', '個', '決定', '是否', '達', '放假', '條件', '引發', '討論', '也', '讓', '綠營', '質疑', '是', '不', '是', '趁機', '救', '謝國樑', '對', '此', '台北市長', '蔣萬安', '還原', '決策', '過程']\n",
      "句子 6:['山陀兒', '襲', '台', '連', '放', '颱風假', ' ', '恐', '衝擊', '薪水', '發放', ' ', '颱風', '山陀兒', '怠速', '前進', '目前', '會不會', '連續', '放', '2', '到', '3', '天', '以上', '颱風假', '還', '很', '難說', '但', '有', '民間', '單位', '苦惱', '這樣', '下去', '5', '日', '薪資', '可能', '無法', '如期', '發放', '位於', '高雄', '的', '方舟', '就業', '服務', '協會', '理事長', '王姿琦']\n",
      "句子 7:['吳宇舒', '晚報', ' ', '山陀兒', '颱風', '獨家', '山陀兒', '牛步', '進逼', '蘭嶼', '擱淺', '貨輪', '慘斷', '兩', '截', '20241002', 'newsebc ', '山陀兒', '颱風', '獨家', '山陀兒', '牛', '步', '進逼', '蘭嶼', '擱淺', '貨輪', '慘', '斷', '兩', '截', '山陀兒', '颱風', '急', '搶修', '枋寮', '狂浪', '沖垮', '堤防', '30尺', '缺口', '釀', '海水', '倒灌', '山陀兒', '颱風', '山陀兒', '挾', '強風', '豪雨', '襲', '台東', '卑南']\n",
      "句子 8:['山陀兒', '準備', '進', '家門', '颱風', '核心', '遭', '陸地', '摩擦', '持續', '變弱', ' ', '專家', '急', '籲', '不可', '輕忽', ' ', '記者', '高珞曦', '綜合', '報導', ' ', '中颱', '山陀兒', '即將', '登陸', '2日', '傍晚', '起', '風雨', '逐漸', '明顯', '氣象', '粉專', '台灣', '颱風', '論壇', '天氣', '特急', '表示', '山陀兒', '因', '移動', '速度', '緩慢', '東側', '核心', '被', '陸地', '摩擦', '強度', '減弱', '但', '其']\n",
      "句子 9:['東港', '迎', '王', '颱風', '攪局', ' ', '千歲爺', '指示', '王船', '法會', '等', '科儀', '按', '表', '進行', ' ', '三', '年', '一', '科', '的', '東港', '迎', '王', '平安', '祭典', '遇到', '山陀兒', '颱風', '攪局', '東港', '東隆宮', '今晚', '聲明', '經', '向', '代天巡狩', '千歲爺', '擲', '杯', '請示', '代天府', '內', '各', '項', '科儀', '將', '依', '祭典', '順']\n",
      "\n",
      "所有詞的列表:\n",
      "['颱風', '生成', '18', '氣象', '專家', '正常', '範圍', '一樣', '20241002', '歷史', '紀錄', '平均值', '西北', '太平洋', '生成', '颱風', '265', '今年', '來到', '10月份', '山陀兒', '算是', '第18', '引起', '關注', '颱風', '8月份', '開始', '進度', '似的', '颱風天', '物資', '搶光', '鮮奶', '滯銷', '感謝', '台灣人', '忘記', '過去', '吳裕堯', '核稿', '編輯', '社群', '平台', '貼文', '掀起', '熱烈', '討論', '附上', '超市', '鮮奶區', '照片', '顯示', '颱風天', '前夕', '物資', '搶購', '味全', '鮮奶', '滯銷', '引發', '眾多', '網友', '留言', '感謝', '大家', '忘記', '吳宇舒', '晚報', '山陀兒', '颱風', '颱風', '郵輪行', '說好', '會員', '免費', '旅客', '晚報', '丟包', '20241002', '山陀兒', '颱風', '旗津', '大雨', '大浪', '民眾', '穿越', '封鎖線', '冒險', '觀浪', '山陀兒', '颱風', '抗戰', '淹水', '銀行', '水閘門', '增高', '被迫', '暫停', '服務', '山陀兒', '颱風', '颱風', '郵輪行', '說好', '會員', '免費', '旅客', '晚報', '台中', '颱風假', '多數', '醫院', '正常', '門診', '中榮開', '颱風', '特別', '門診', '山陀兒', '颱風', '逼近', '氣象署', '預估', '台中市', '明日', '符合', '停班', '標準', '台中市', '晚間', '8點', '宣佈', '明天', '停止', '上班', '上課', '台中', '颱風假', '包括', '中國', '醫藥', '大學', '附設', '醫院', '台中', '醫院', '門診', '正常', '服務', '山陀兒', '颱風', '3日', '轉彎', '日本', '預測', '回馬槍', '南部', '民間', '傳說', '破功', '山陀兒', '颱風', '滯留', '台灣', '西南部', '2日', '北上', '中心', '些許', '滯留', '狀況', '暴風圈', '進入', '雲林', '以南', '臺東', '陸地', '新竹', '以南', '南投', '宜蘭', '花蓮', '臺東', '地區', '山陀兒', '颱風', '颱風假', '氣象署', '風雨', '縣市', '顧及', '市民', '安全', '颱風', '山陀兒', '影響', '持續', '北北桃', '宣布', '10月', '2號', '颱風假', '決定', '放假', '條件', '引發', '討論', '綠營', '質疑', '謝國樑', '台北市長', '蔣萬安', '還原', '決策', '過程', '山陀兒', '颱風假', '衝擊', '薪水', '發放', '颱風', '山陀兒', '怠速', '前進', '目前', '以上', '颱風假', '難說', '民間', '單位', '苦惱', '這樣', '薪資', '發放', '位於', '高雄', '方舟', '就業', '服務', '協會', '理事長', '王姿琦', '吳宇舒', '晚報', '山陀兒', '颱風', '山陀兒', '牛步', '進逼', '蘭嶼', '擱淺', '貨輪', '慘斷', '20241002', '山陀兒', '颱風', '山陀兒', '進逼', '蘭嶼', '擱淺', '貨輪', '山陀兒', '颱風', '搶修', '枋寮', '狂浪', '沖垮', '堤防', '30尺', '缺口', '海水', '倒灌', '山陀兒', '颱風', '山陀兒', '強風', '豪雨', '台東', '卑南', '山陀兒', '準備', '家門', '颱風', '核心', '陸地', '摩擦', '持續', '變弱', '專家', '輕忽', '記者', '高珞曦', '綜合', '報導', '中颱', '山陀兒', '登陸', '2日', '傍晚', '風雨', '明顯', '氣象', '粉專', '台灣', '颱風', '論壇', '天氣', '特急', '表示', '山陀兒', '移動', '速度', '緩慢', '東側', '核心', '陸地', '摩擦', '強度', '減弱', '東港', '颱風', '攪局', '千歲爺', '指示', '王船', '法會', '科儀', '進行', '東港', '平安', '祭典', '遇到', '山陀兒', '颱風', '攪局', '東港', '東隆宮', '今晚', '聲明', '代天巡狩', '千歲爺', '請示', '代天府', '科儀', '祭典']\n",
      "總詞數: 346\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "import re\n",
    "import time\n",
    "from ckip_transformers import __version__\n",
    "from ckip_transformers.nlp import CkipWordSegmenter, CkipPosTagger, CkipNerChunker\n",
    "\n",
    "\n",
    "# Initialize drivers\n",
    "# 初始化斷詞WS、詞性標記POS、命名實體辨識NER\n",
    "print(\"Initializing drivers ... WS\")\n",
    "ws_driver = CkipWordSegmenter(model=\"bert-base\")\n",
    "print(\"Initializing drivers ... POS\")\n",
    "pos_driver = CkipPosTagger(model=\"bert-base\")\n",
    "print(\"Initializing drivers ... NER\")\n",
    "ner_driver = CkipNerChunker(model=\"bert-base\")\n",
    "print(\"Initializing drivers ... done\")\n",
    "\n",
    "def process_text(text):\n",
    "    # 選取所有 “不是 ＾” 中文（\\u4e00-\\u9fff）、英文、數字及空格\n",
    "    pattern = re.compile(r'[^\\u4e00-\\u9fffA-Za-z0-9\\s]')\n",
    "    # 用空字元取代符合 pattern 的字串\n",
    "    return pattern.sub('', text)\n",
    "\n",
    "sentences = []\n",
    "\n",
    "for item in all_content:\n",
    "    cleaned_text = process_text(item)\n",
    "    print(cleaned_text)\n",
    "    sentences.append(cleaned_text)\n",
    "    \n",
    "\n",
    "\n",
    "# 應該用有限制條件去剃掉那些多種不想要的\n",
    "\n",
    "# - 初始化斷詞（WS）、命名實體辨識（NER）輸入格式 ----> 必須是句子的串列（list of sentences）\n",
    "# - 詞性標記（POS）輸入格式 ----> 必須是詞串列的串列（list of list of words，來自斷詞的輸出）\n",
    "\n",
    "all_words = []\n",
    "\n",
    "\n",
    "# print(\"\\n斷詞結果：\")\n",
    "ws_results = ws_driver(sentences)\n",
    "\n",
    "all_pos = []\n",
    "\n",
    "# 執行詞性標記\n",
    "# print(\"\\n詞性標記結果：\")\n",
    "pos_results = pos_driver(ws_results)\n",
    "for sentence in pos_results:\n",
    "    # print(sentence)\n",
    "    # print(type(sentence))\n",
    "    \n",
    "    for pos in sentence:\n",
    "        all_pos.append(pos)\n",
    "\n",
    "print(f\"總詞性數: {len(all_pos)}\")\n",
    "\n",
    "\n",
    "# print(ws_results)\n",
    "# print(type(ws_results[0]))\n",
    "for n, sentence in enumerate(ws_results):\n",
    "    print(f\"句子 {n}:{sentence}\")\n",
    "    # print(len(sentence))\n",
    "    # print(f\"資料型態: {type(sentence)}\")\n",
    "    # print(type(sentence))\n",
    "    \n",
    "    # 遍歷每個句子的詞與對應的詞性標註\n",
    "    for word, pos in zip(sentence, pos_results[n]):\n",
    "        # 判斷條件：非空字串、詞長度大於1，且詞性為名詞(N)或動詞(V)\n",
    "        if word.strip() and len(word) > 1 and (pos.startswith('N') or pos.startswith('V')):\n",
    "            all_words.append(word)\n",
    "            \n",
    "    # or all_words.extend(sentence) 把可迭代對象的元素逐個添加到列表的末尾\n",
    "\n",
    "print(\"\\n所有詞的列表:\")\n",
    "print(all_words)\n",
    "print(f\"總詞數: {len(all_words)}\")\n",
    "\n",
    "\n",
    "# 學習資源附上的的文章教學有比較乾淨的寫法 用 double zip \n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "斷詞結果:\n",
      "['颱', '風', '生成', '偏少', '?', '至今', '僅', '18', '個', ' ', '氣象', '專家', ':', '正常', '範圍', '偏少', ' ', '｜', '十點', '不', '一樣', '20241002', ' ', '@', 'TVBSNEWS01', '\\n', '歷史', '紀錄', '的', '平均值', '，', '西北', '太平洋', '生成', '颱', '風數', '26.5', '個', '，', '但', '今年', '都', '已', '經來', '到', '10', '月份', '了', '，', '山陀兒', '算是', '第', '18', '個', '，', '似乎', '有點', '偏少', '，', '引起', '各方', '關注', '，', '而且', '颱', '風', '更是', '從', '8', '月份', '開始', '像', '在', '趕', '進度', '似的', '，', '...', '\\n', '\\n', '颱', '風天物資', '搶光', '！', '這牌', '鮮奶', '滯銷', ' ', '網', '：', '感謝', '台灣', '人', '沒', '忘', '記過', '去', '\\n', '吳裕堯', '／', '核稿', '編輯', '社群', '平台', 'Threads', '一則', '貼文', '掀起', '熱烈', '討論', '，', '文中', '附上', '的', '超市', '鮮奶區', '照片', '顯示', '，', '颱', '風天', '前夕', '物資', '被', '搶購', '一空', '，', '味全', '鮮奶', '依然', '滯銷', '，', '引發眾', '多', '網友', '留言', '感謝', '大家', '沒有', '忘記', '...', '\\n', '\\n', '【', '吳宇', '舒晚', '報', '】', ' ', '【', '山陀兒', '颱', '風', '】', '颱', '風擾', '郵輪行', '！', '說', '好會員', '免費', '搭', '旅客', '晚報', '到', '控遭', '丟包', '2024', '／', '10', '／', '02', '@', 'newsebc', '\\n', '【', '山陀兒', '颱', '風', '】', '旗津風', '大雨', '大浪', '大民眾', '穿越', '封鎖線', '冒險', '觀浪', '【', '山陀兒', '颱', '風', '】', '抗戰防', '淹水', '！', '銀行', '水閘門', '增高', 'ATM', '被迫', '暫停', '服務', '【', '山陀兒', '颱', '風', '】', '颱', '風擾', '郵輪行', '！', '說', '好會員', '免費', '搭', '旅客', '晚報', '...', '\\n', '\\n', '台中明', '颱', '風假', '！', '多數', '醫院', '正常', '門診', ' ', '中榮開', '颱', '風特別', '門診', '\\n', '山陀兒', '颱', '風', '逼近', '，', '因氣', '象署', '預估', '台中市', '明日', '符合', '停班課', '標準', '，', '台中市', '晚間', '8', '點宣', '佈', '明天', '停止', '上班', '上課', '，', '台中將', '連放', '2', '天', '颱', '風假', '，', '包括', '中國', '醫藥', '大學', '附設', '醫院', '及', '台', '中', '醫院', '均', '門診', '正常', '服務', '，', '...', '\\n', '\\n', '山陀兒', '颱', '風', '3', '日', '又', '要', '轉彎', '？', '日本', '預測', '「', '回馬槍', '再襲', '南部', '」', '，', '民間', '傳說', '恐破功', '\\n', '山陀兒', '颱', '風滯', '留台灣', '西南部', '許久', '，', '今', '（', '2', '）', '日終', '於', '動起', '來', '北上', '，', '雖然', '中心', '還是', '有些', '許滯', '留狀況', '，', '不過', '暴風圈', '已', '進入', '雲林', '以南', '及', '臺', '東陸', '地', '，', '對', '新竹', '以南', '、', '南投', '、', '宜蘭', '、', '花蓮', '、', '臺', '東地區', '及', '...', '\\n', '\\n', '【', '山陀兒', '颱', '風', '】', '綠扯', '颱', '風假', '救', '樑', ' ', '氣象署', '曝原', '估有', '風雨', ' ', '4', '縣市', '顧及', '市民', '安全', ' ', '@', 'newsebc', '\\n', '颱', '風山陀兒', '影響', '持續', '，', '基北北桃', '宣布', '10', '月', '2', '號放', '颱', '風假', '，', '但', '這個', '決定', '，', '是否', '達', '「', '放假', '條件', '」', '引發', '討論', '，', '也', '讓', '綠營', '質疑', '是不是', '趁機', '救', '謝國', '樑', '？', '對此', '，', '台北市', '長', '蔣萬安', '還原', '決策', '過程', '，', '...', '\\n', '\\n', '山陀兒', '襲台', '「', '連放', '颱', '風假', '」', ' ', '恐衝擊', '薪水', '發放', '！', '\\n', '颱', '風', '「', '山陀兒', '」', '怠速', '前進', '，', '目前', '會不會', '連續', '放', '2', '到', '3', '天', '以上', '颱', '風假還', '很', '難', '說', '，', '但', '有民間', '單位', '苦惱', '，', '這樣', '下去', '5', '日薪', '資', '可能', '無法', '如期', '發放', '。', '位', '於', '高雄', '的', '「', '方舟', '就業', '服務協會', '」', '理事', '長', '王姿琦', '...', '\\n', '\\n', '【', '吳宇', '舒晚', '報', '】', ' ', '【', '山陀兒', '颱', '風', '】', '獨家', '》', '「', '山陀兒', '」', '牛步', '進逼蘭嶼', '擱淺', '貨輪', '慘斷', '兩截', '2024', '／', '10', '／', '02', '@', 'newsebc', '\\n', '【', '山陀兒', '颱', '風', '】', '獨家', '》', '「', '山陀兒', '」', '牛步', '進逼蘭嶼', '擱淺', '貨輪', '慘斷', '兩截', '【', '山陀兒', '颱', '風', '】', '急', '搶', '修', '！', '枋寮', '狂浪', '沖', '垮', '堤防', '30', '尺', '缺口', '釀', '海水', '倒灌', '【', '山陀兒', '颱', '風', '】', '山陀兒', '挾強', '風', '豪雨', '襲台東', '卑南', '...', '\\n', '\\n', '山陀兒', '準備', '「', '進家門', '」', '！', '颱', '風', '核心', '遭陸', '地', '摩擦', '、', '持續', '變弱', ' ', '專家', '急籲', '：', '不可', '輕忽', '\\n', '記者', '高', '珞曦', '／', '綜合', '報導', '.', ' ', '中', '颱', '山陀兒', '即將', '登陸', '，', '2', '日', '傍晚', '起風雨', '逐漸', '明顯', '，', '氣象', '粉專', '「', '台灣', '颱', '風論壇', '｜', '天氣', '特急', '」', '表示', '，', '山陀兒', '因移動', '速度', '緩慢', '，', '東側', '核心', '被', '陸地', '摩擦', '、', '強度', '減弱', '，', '但', '其', '...', '\\n', '\\n', '東港', '迎王', '颱', '風', '攪局', ' ', '千歲爺', '指示', '王船', '法會', '等', '科儀', '按表', '進行', '\\n', '三年', '一科', '的', '東港', '迎', '王平安', '祭典', '遇到', '山陀兒', '颱', '風', '攪局', '，', '東港', '東隆宮', '今晚', '聲明', '，', '經向代天', '巡', '狩千歲爺', '擲', '杯', '請示', '，', '代', '天府', '內', '各項', '科儀', '將依', '祭典', '順', '...', '\\n', '\\n']\n",
      "------------------------------------------------------------------------------------------------------------------------\n",
      "過濾後的結果: \n",
      "508\n",
      "------------------------------------------------------------------------------------------------------------------------\n",
      "\n",
      "TF-IDF 關鍵詞及其權重:\n",
      "山陀兒: 4548\n",
      "風假: 1624\n",
      "醫院: 974\n",
      "風山陀兒: 974\n",
      "東港: 974\n",
      "偏少: 850\n",
      "牛步: 695\n",
      "18: 649\n",
      "專家: 649\n",
      "10: 649\n",
      "\n",
      "TextRank 關鍵詞及其權重:\n",
      "山陀: 1.0000\n",
      "引發: 0.4896\n",
      "摩擦: 0.3790\n",
      "服務: 0.3789\n",
      "台中市: 0.3670\n",
      "山陀兒: 0.3669\n",
      "討論: 0.3422\n",
      "專家: 0.3349\n",
      "核心: 0.3149\n",
      "暴風圈: 0.3000\n",
      "------------------------------------------------------------------------------------------------------------------------\n",
      "第一種過濾後的結果 (TF-IDF): \n",
      "['偏少', '18', '專家', '偏少', '10', '山陀兒', '18', '偏少', '山陀兒', '10', '山陀兒', '山陀兒', '山陀兒', '風假', '醫院', '山陀兒', '風假', '醫院', '醫院', '山陀兒', '山陀兒', '山陀兒', '風假', '風山陀兒', '10', '風假', '山陀兒', '風假', '山陀兒', '山陀兒', '山陀兒', '牛步', '10', '山陀兒', '山陀兒', '牛步', '山陀兒', '山陀兒', '山陀兒', '山陀兒', '專家', '山陀兒', '山陀兒', '東港', '東港', '山陀兒', '東港']\n",
      "47\n",
      "------------------------------------------------------------------------------------------------------------------------\n",
      "第二種過濾後的結果 (TextRank): \n",
      "['專家', '山陀兒', '討論', '山陀兒', '山陀兒', '山陀兒', '服務', '山陀兒', '山陀兒', '台中市', '台中市', '服務', '山陀兒', '山陀兒', '暴風圈', '山陀兒', '引發', '討論', '山陀兒', '山陀兒', '山陀兒', '山陀兒', '山陀兒', '山陀兒', '山陀兒', '山陀兒', '山陀兒', '山陀兒', '核心', '摩擦', '專家', '山陀兒', '山陀兒', '核心', '摩擦', '山陀兒']\n",
      "36\n"
     ]
    }
   ],
   "source": [
    "import re\n",
    "import jieba.analyse\n",
    "\n",
    "# 讀取停用詞\n",
    "with open(r'./stopword.txt', 'r', encoding='utf-8') as file:\n",
    "    stopword = file.read().splitlines() # BIF splitlines\n",
    "    # print(stopword)\n",
    "\n",
    "# 合併標題和描述，\\n 是方便閱讀，可以不用加\n",
    "all_content = \"\"\n",
    "for item in loaded_result:\n",
    "    all_content += item['title'] + \"\\n\"\n",
    "    if 'desc' in item:\n",
    "        all_content += item['desc'] + \"\\n\"\n",
    "    all_content += \"\\n\"\n",
    "\n",
    "# print(\"原始內容:\")\n",
    "# print(all_content)\n",
    "\n",
    "print(\"斷詞結果:\")\n",
    "seg_list = jieba.lcut(all_content)\n",
    "print(seg_list)\n",
    "print(\"--\"*60)\n",
    "\n",
    "# 原本的正則表達式\n",
    "# english_pattern = re.compile(r'^[a-zA-Z]+$')\n",
    "\n",
    "# def is_english(text):\n",
    "#     return bool(english_pattern.match(text))\n",
    "\n",
    "# 過濾停用詞和空白字符\n",
    "filtered_words = []\n",
    "for word in seg_list:\n",
    "    # 去除詞的前後空白 跟上面產生的 \\n\n",
    "    cleaned_word = word.strip()\n",
    "    \n",
    "    cleaned_word = re.sub(r'[^\\u4e00-\\u9fffA-Za-z0-9\\s]', '', cleaned_word)\n",
    "    # re.sub(pattern, repl, string, count=0)\n",
    "    # 檢查詞是否為空且不在停用詞列表中\n",
    "    if cleaned_word and cleaned_word not in stopword:\n",
    "        filtered_words.append(cleaned_word)\n",
    "\n",
    "print('過濾後的結果: ')\n",
    "print(len(filtered_words))\n",
    "print(\"--\"*60)\n",
    "\n",
    "# 使用 jieba TF-IDF 算法提取關鍵詞\n",
    "tags = jieba.analyse.extract_tags(\"\".join(filtered_words), topK=10, withWeight=True)\n",
    "\n",
    "# print(\"\\n提取的關鍵詞:\")\n",
    "# print(\"Tags type:\", type(tags))\n",
    "# print(\"Tags content:\", tags)\n",
    "\n",
    "top_50_tags = [tag for tag, _ in tags]\n",
    "\n",
    "second_filtered_words = []\n",
    "print(\"\\nTF-IDF 關鍵詞及其權重:\")\n",
    "for tag_tuple in tags:\n",
    "    tag = tag_tuple[0]\n",
    "    weight = tag_tuple[1]\n",
    "    print(f\"{tag}: {int(weight * 10000)}\")\n",
    "    \n",
    "for word in filtered_words:\n",
    "    if word in top_50_tags:\n",
    "        second_filtered_words.append(word)\n",
    "\n",
    "# 使用 TextRank 算法提取關鍵詞\n",
    "textrank_tags = jieba.analyse.textrank(\"\".join(filtered_words), topK=10, withWeight=True)\n",
    "\n",
    "# print(\"\\nTextRank 提取的關鍵詞:\")\n",
    "# print(\"TextRank Tags type:\", type(textrank_tags))\n",
    "# print(\"TextRank Tags content:\", textrank_tags)\n",
    "\n",
    "top_50_textrank_tags = [tag for tag, _ in textrank_tags]\n",
    "\n",
    "third_filtered_words = []\n",
    "print(\"\\nTextRank 關鍵詞及其權重:\")\n",
    "for tag_tuple in textrank_tags:\n",
    "    tag = tag_tuple[0]\n",
    "    weight = tag_tuple[1]\n",
    "    print(f\"{tag}: {weight:.4f}\")\n",
    "    \n",
    "for word in filtered_words:\n",
    "    if word in top_50_textrank_tags:\n",
    "        third_filtered_words.append(word)\n",
    "\n",
    "print(\"--\"*60)\n",
    "print(\"第一種過濾後的結果 (TF-IDF): \")\n",
    "print(second_filtered_words)\n",
    "print(len(second_filtered_words))\n",
    "\n",
    "print(\"--\"*60)\n",
    "print(\"第二種過濾後的結果 (TextRank): \")\n",
    "print(third_filtered_words)\n",
    "print(len(third_filtered_words))\n",
    "        \n",
    "# You can choose which filtered words to use for further processing\n",
    "# filtered_words = second_filtered_words  # Using TF-IDF results\n",
    "# filtered_words = third_filtered_words  # Using TextRank results\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "   - `^` 表示字符串的開始\n",
    "   - `[a-zA-Z]` 表示任何英文字母（大寫或小寫）\n",
    "   - `+` 表示前面的字符至少出現一次\n",
    "   - `$` 表示字符串的結束\n",
    "   -  `.*`：匹配任意字符（除換行符外）零次或多次"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "import re\n",
    "\n",
    "english_pattern = re.compile(r'^[a-zA-Z]+$')\n",
    "\n",
    "def process_text(text):\n",
    "    return bool(english_pattern.match(text))\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 計算文字出現次數"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Counter({'颱': 33, '山陀兒': 22, '風': 18, '風假': 5, '但': 4, '10': 4, '偏少': 3, '正常': 3, '到': 3, '說': 3, 'newsebc': 3, '醫院': 3, '門診': 3, '及': 3, '東港': 3, '生成': 2, '18': 2, '氣象': 2, '專家': 2, '已': 2, '月份': 2, '鮮奶': 2, '滯銷': 2, '感謝': 2, '台灣': 2, '討論': 2, '被': 2, '吳宇': 2, '舒晚': 2, '報': 2, '風擾': 2, '郵輪行': 2, '好會員': 2, '免費': 2, '搭': 2, '旅客': 2, '晚報': 2, '2024': 2, '服務': 2, '台中市': 2, '連放': 2, '天': 2, '中': 2, '日': 2, '於': 2, '以南': 2, '臺': 2, '地': 2, '救': 2, '樑': 2, '持續': 2, '長': 2, '發放': 2, '獨家': 2, '牛步': 2, '進逼蘭嶼': 2, '擱淺': 2, '貨輪': 2, '慘斷': 2, '兩截': 2, '核心': 2, '摩擦': 2, '攪局': 2, '科儀': 2, '祭典': 2, '至今': 1, '僅': 1, '範圍': 1, '十點': 1, '不': 1, '一樣': 1, '20241002': 1, 'TVBSNEWS01': 1, '歷史': 1, '紀錄': 1, '平均值': 1, '西北': 1, '太平洋': 1, '風數': 1, '265': 1, '今年': 1, '都': 1, '經來': 1, '算是': 1, '第': 1, '似乎': 1, '有點': 1, '引起': 1, '各方': 1, '關注': 1, '而且': 1, '更是': 1, '從': 1, '開始': 1, '像': 1, '趕': 1, '進度': 1, '似的': 1, '風天物資': 1, '搶光': 1, '這牌': 1, '網': 1, '人': 1, '沒': 1, '忘': 1, '記過': 1, '去': 1, '吳裕堯': 1, '核稿': 1, '編輯': 1, '社群': 1, '平台': 1, 'Threads': 1, '一則': 1, '貼文': 1, '掀起': 1, '熱烈': 1, '文中': 1, '附上': 1, '超市': 1, '鮮奶區': 1, '照片': 1, '顯示': 1, '風天': 1, '前夕': 1, '物資': 1, '搶購': 1, '一空': 1, '味全': 1, '依然': 1, '引發眾': 1, '多': 1, '網友': 1, '留言': 1, '大家': 1, '沒有': 1, '忘記': 1, '控遭': 1, '丟包': 1, '旗津風': 1, '大雨': 1, '大浪': 1, '大民眾': 1, '穿越': 1, '封鎖線': 1, '冒險': 1, '觀浪': 1, '抗戰防': 1, '淹水': 1, '銀行': 1, '水閘門': 1, '增高': 1, 'ATM': 1, '被迫': 1, '暫停': 1, '台中明': 1, '多數': 1, '中榮開': 1, '風特別': 1, '逼近': 1, '因氣': 1, '象署': 1, '預估': 1, '明日': 1, '符合': 1, '停班課': 1, '標準': 1, '晚間': 1, '點宣': 1, '佈': 1, '明天': 1, '停止': 1, '上班': 1, '上課': 1, '台中將': 1, '包括': 1, '中國': 1, '醫藥': 1, '大學': 1, '附設': 1, '台': 1, '均': 1, '又': 1, '要': 1, '轉彎': 1, '日本': 1, '預測': 1, '回馬槍': 1, '再襲': 1, '南部': 1, '民間': 1, '傳說': 1, '恐破功': 1, '風滯': 1, '留台灣': 1, '西南部': 1, '許久': 1, '今': 1, '日終': 1, '動起': 1, '來': 1, '北上': 1, '雖然': 1, '中心': 1, '還是': 1, '有些': 1, '許滯': 1, '留狀況': 1, '不過': 1, '暴風圈': 1, '進入': 1, '雲林': 1, '東陸': 1, '對': 1, '新竹': 1, '南投': 1, '宜蘭': 1, '花蓮': 1, '東地區': 1, '綠扯': 1, '氣象署': 1, '曝原': 1, '估有': 1, '風雨': 1, '縣市': 1, '顧及': 1, '市民': 1, '安全': 1, '風山陀兒': 1, '影響': 1, '基北北桃': 1, '宣布': 1, '月': 1, '號放': 1, '這個': 1, '決定': 1, '是否': 1, '達': 1, '放假': 1, '條件': 1, '引發': 1, '讓': 1, '綠營': 1, '質疑': 1, '是不是': 1, '趁機': 1, '謝國': 1, '對此': 1, '台北市': 1, '蔣萬安': 1, '還原': 1, '決策': 1, '過程': 1, '襲台': 1, '恐衝擊': 1, '薪水': 1, '怠速': 1, '前進': 1, '目前': 1, '會不會': 1, '連續': 1, '放': 1, '以上': 1, '風假還': 1, '很': 1, '難': 1, '有民間': 1, '單位': 1, '苦惱': 1, '這樣': 1, '下去': 1, '日薪': 1, '資': 1, '可能': 1, '無法': 1, '如期': 1, '位': 1, '高雄': 1, '方舟': 1, '就業': 1, '服務協會': 1, '理事': 1, '王姿琦': 1, '急': 1, '搶': 1, '修': 1, '枋寮': 1, '狂浪': 1, '沖': 1, '垮': 1, '堤防': 1, '30': 1, '尺': 1, '缺口': 1, '釀': 1, '海水': 1, '倒灌': 1, '挾強': 1, '豪雨': 1, '襲台東': 1, '卑南': 1, '準備': 1, '進家門': 1, '遭陸': 1, '變弱': 1, '急籲': 1, '不可': 1, '輕忽': 1, '記者': 1, '高': 1, '珞曦': 1, '綜合': 1, '報導': 1, '即將': 1, '登陸': 1, '傍晚': 1, '起風雨': 1, '逐漸': 1, '明顯': 1, '粉專': 1, '風論壇': 1, '天氣': 1, '特急': 1, '表示': 1, '因移動': 1, '速度': 1, '緩慢': 1, '東側': 1, '陸地': 1, '強度': 1, '減弱': 1, '其': 1, '迎王': 1, '千歲爺': 1, '指示': 1, '王船': 1, '法會': 1, '等': 1, '按表': 1, '進行': 1, '三年': 1, '一科': 1, '迎': 1, '王平安': 1, '遇到': 1, '東隆宮': 1, '今晚': 1, '聲明': 1, '經向代天': 1, '巡': 1, '狩千歲爺': 1, '擲': 1, '杯': 1, '請示': 1, '代': 1, '天府': 1, '內': 1, '各項': 1, '將依': 1, '順': 1})\n",
      "<class 'collections.Counter'>\n",
      "['__add__', '__and__', '__class__', '__class_getitem__', '__contains__', '__delattr__', '__delitem__', '__dict__', '__dir__', '__doc__', '__eq__', '__format__', '__ge__', '__getattribute__', '__getitem__', '__gt__', '__hash__', '__iadd__', '__iand__', '__init__', '__init_subclass__', '__ior__', '__isub__', '__iter__', '__le__', '__len__', '__lt__', '__missing__', '__module__', '__ne__', '__neg__', '__new__', '__or__', '__pos__', '__reduce__', '__reduce_ex__', '__repr__', '__reversed__', '__ror__', '__setattr__', '__setitem__', '__sizeof__', '__str__', '__sub__', '__subclasshook__', '__weakref__', '_keep_positive', 'clear', 'copy', 'elements', 'fromkeys', 'get', 'items', 'keys', 'most_common', 'pop', 'popitem', 'setdefault', 'subtract', 'total', 'update', 'values']\n",
      "------------------------------------------------------------------------------------------------------------------------\n",
      "word_count.items():\n",
      "dict_items([('颱', 33), ('風', 18), ('生成', 2), ('偏少', 3), ('至今', 1), ('僅', 1), ('18', 2), ('氣象', 2), ('專家', 2), ('正常', 3), ('範圍', 1), ('十點', 1), ('不', 1), ('一樣', 1), ('20241002', 1), ('TVBSNEWS01', 1), ('歷史', 1), ('紀錄', 1), ('平均值', 1), ('西北', 1), ('太平洋', 1), ('風數', 1), ('265', 1), ('但', 4), ('今年', 1), ('都', 1), ('已', 2), ('經來', 1), ('到', 3), ('10', 4), ('月份', 2), ('山陀兒', 22), ('算是', 1), ('第', 1), ('似乎', 1), ('有點', 1), ('引起', 1), ('各方', 1), ('關注', 1), ('而且', 1), ('更是', 1), ('從', 1), ('開始', 1), ('像', 1), ('趕', 1), ('進度', 1), ('似的', 1), ('風天物資', 1), ('搶光', 1), ('這牌', 1), ('鮮奶', 2), ('滯銷', 2), ('網', 1), ('感謝', 2), ('台灣', 2), ('人', 1), ('沒', 1), ('忘', 1), ('記過', 1), ('去', 1), ('吳裕堯', 1), ('核稿', 1), ('編輯', 1), ('社群', 1), ('平台', 1), ('Threads', 1), ('一則', 1), ('貼文', 1), ('掀起', 1), ('熱烈', 1), ('討論', 2), ('文中', 1), ('附上', 1), ('超市', 1), ('鮮奶區', 1), ('照片', 1), ('顯示', 1), ('風天', 1), ('前夕', 1), ('物資', 1), ('被', 2), ('搶購', 1), ('一空', 1), ('味全', 1), ('依然', 1), ('引發眾', 1), ('多', 1), ('網友', 1), ('留言', 1), ('大家', 1), ('沒有', 1), ('忘記', 1), ('吳宇', 2), ('舒晚', 2), ('報', 2), ('風擾', 2), ('郵輪行', 2), ('說', 3), ('好會員', 2), ('免費', 2), ('搭', 2), ('旅客', 2), ('晚報', 2), ('控遭', 1), ('丟包', 1), ('2024', 2), ('newsebc', 3), ('旗津風', 1), ('大雨', 1), ('大浪', 1), ('大民眾', 1), ('穿越', 1), ('封鎖線', 1), ('冒險', 1), ('觀浪', 1), ('抗戰防', 1), ('淹水', 1), ('銀行', 1), ('水閘門', 1), ('增高', 1), ('ATM', 1), ('被迫', 1), ('暫停', 1), ('服務', 2), ('台中明', 1), ('風假', 5), ('多數', 1), ('醫院', 3), ('門診', 3), ('中榮開', 1), ('風特別', 1), ('逼近', 1), ('因氣', 1), ('象署', 1), ('預估', 1), ('台中市', 2), ('明日', 1), ('符合', 1), ('停班課', 1), ('標準', 1), ('晚間', 1), ('點宣', 1), ('佈', 1), ('明天', 1), ('停止', 1), ('上班', 1), ('上課', 1), ('台中將', 1), ('連放', 2), ('天', 2), ('包括', 1), ('中國', 1), ('醫藥', 1), ('大學', 1), ('附設', 1), ('及', 3), ('台', 1), ('中', 2), ('均', 1), ('日', 2), ('又', 1), ('要', 1), ('轉彎', 1), ('日本', 1), ('預測', 1), ('回馬槍', 1), ('再襲', 1), ('南部', 1), ('民間', 1), ('傳說', 1), ('恐破功', 1), ('風滯', 1), ('留台灣', 1), ('西南部', 1), ('許久', 1), ('今', 1), ('日終', 1), ('於', 2), ('動起', 1), ('來', 1), ('北上', 1), ('雖然', 1), ('中心', 1), ('還是', 1), ('有些', 1), ('許滯', 1), ('留狀況', 1), ('不過', 1), ('暴風圈', 1), ('進入', 1), ('雲林', 1), ('以南', 2), ('臺', 2), ('東陸', 1), ('地', 2), ('對', 1), ('新竹', 1), ('南投', 1), ('宜蘭', 1), ('花蓮', 1), ('東地區', 1), ('綠扯', 1), ('救', 2), ('樑', 2), ('氣象署', 1), ('曝原', 1), ('估有', 1), ('風雨', 1), ('縣市', 1), ('顧及', 1), ('市民', 1), ('安全', 1), ('風山陀兒', 1), ('影響', 1), ('持續', 2), ('基北北桃', 1), ('宣布', 1), ('月', 1), ('號放', 1), ('這個', 1), ('決定', 1), ('是否', 1), ('達', 1), ('放假', 1), ('條件', 1), ('引發', 1), ('讓', 1), ('綠營', 1), ('質疑', 1), ('是不是', 1), ('趁機', 1), ('謝國', 1), ('對此', 1), ('台北市', 1), ('長', 2), ('蔣萬安', 1), ('還原', 1), ('決策', 1), ('過程', 1), ('襲台', 1), ('恐衝擊', 1), ('薪水', 1), ('發放', 2), ('怠速', 1), ('前進', 1), ('目前', 1), ('會不會', 1), ('連續', 1), ('放', 1), ('以上', 1), ('風假還', 1), ('很', 1), ('難', 1), ('有民間', 1), ('單位', 1), ('苦惱', 1), ('這樣', 1), ('下去', 1), ('日薪', 1), ('資', 1), ('可能', 1), ('無法', 1), ('如期', 1), ('位', 1), ('高雄', 1), ('方舟', 1), ('就業', 1), ('服務協會', 1), ('理事', 1), ('王姿琦', 1), ('獨家', 2), ('牛步', 2), ('進逼蘭嶼', 2), ('擱淺', 2), ('貨輪', 2), ('慘斷', 2), ('兩截', 2), ('急', 1), ('搶', 1), ('修', 1), ('枋寮', 1), ('狂浪', 1), ('沖', 1), ('垮', 1), ('堤防', 1), ('30', 1), ('尺', 1), ('缺口', 1), ('釀', 1), ('海水', 1), ('倒灌', 1), ('挾強', 1), ('豪雨', 1), ('襲台東', 1), ('卑南', 1), ('準備', 1), ('進家門', 1), ('核心', 2), ('遭陸', 1), ('摩擦', 2), ('變弱', 1), ('急籲', 1), ('不可', 1), ('輕忽', 1), ('記者', 1), ('高', 1), ('珞曦', 1), ('綜合', 1), ('報導', 1), ('即將', 1), ('登陸', 1), ('傍晚', 1), ('起風雨', 1), ('逐漸', 1), ('明顯', 1), ('粉專', 1), ('風論壇', 1), ('天氣', 1), ('特急', 1), ('表示', 1), ('因移動', 1), ('速度', 1), ('緩慢', 1), ('東側', 1), ('陸地', 1), ('強度', 1), ('減弱', 1), ('其', 1), ('東港', 3), ('迎王', 1), ('攪局', 2), ('千歲爺', 1), ('指示', 1), ('王船', 1), ('法會', 1), ('等', 1), ('科儀', 2), ('按表', 1), ('進行', 1), ('三年', 1), ('一科', 1), ('迎', 1), ('王平安', 1), ('祭典', 2), ('遇到', 1), ('東隆宮', 1), ('今晚', 1), ('聲明', 1), ('經向代天', 1), ('巡', 1), ('狩千歲爺', 1), ('擲', 1), ('杯', 1), ('請示', 1), ('代', 1), ('天府', 1), ('內', 1), ('各項', 1), ('將依', 1), ('順', 1)])\n",
      "<class 'dict_items'>\n",
      "------------------------------------------------------------------------------------------------------------------------\n",
      "word_count_dict:\n",
      "OrderedDict([('颱', 33), ('山陀兒', 22), ('風', 18), ('風假', 5), ('但', 4), ('10', 4), ('偏少', 3), ('正常', 3), ('到', 3), ('說', 3), ('newsebc', 3), ('醫院', 3), ('門診', 3), ('及', 3), ('東港', 3), ('生成', 2), ('18', 2), ('氣象', 2), ('專家', 2), ('已', 2), ('月份', 2), ('鮮奶', 2), ('滯銷', 2), ('感謝', 2), ('台灣', 2), ('討論', 2), ('被', 2), ('吳宇', 2), ('舒晚', 2), ('報', 2), ('風擾', 2), ('郵輪行', 2), ('好會員', 2), ('免費', 2), ('搭', 2), ('旅客', 2), ('晚報', 2), ('2024', 2), ('服務', 2), ('台中市', 2), ('連放', 2), ('天', 2), ('中', 2), ('日', 2), ('於', 2), ('以南', 2), ('臺', 2), ('地', 2), ('救', 2), ('樑', 2), ('持續', 2), ('長', 2), ('發放', 2), ('獨家', 2), ('牛步', 2), ('進逼蘭嶼', 2), ('擱淺', 2), ('貨輪', 2), ('慘斷', 2), ('兩截', 2), ('核心', 2), ('摩擦', 2), ('攪局', 2), ('科儀', 2), ('祭典', 2), ('至今', 1), ('僅', 1), ('範圍', 1), ('十點', 1), ('不', 1), ('一樣', 1), ('20241002', 1), ('TVBSNEWS01', 1), ('歷史', 1), ('紀錄', 1), ('平均值', 1), ('西北', 1), ('太平洋', 1), ('風數', 1), ('265', 1), ('今年', 1), ('都', 1), ('經來', 1), ('算是', 1), ('第', 1), ('似乎', 1), ('有點', 1), ('引起', 1), ('各方', 1), ('關注', 1), ('而且', 1), ('更是', 1), ('從', 1), ('開始', 1), ('像', 1), ('趕', 1), ('進度', 1), ('似的', 1), ('風天物資', 1), ('搶光', 1), ('這牌', 1), ('網', 1), ('人', 1), ('沒', 1), ('忘', 1), ('記過', 1), ('去', 1), ('吳裕堯', 1), ('核稿', 1), ('編輯', 1), ('社群', 1), ('平台', 1), ('Threads', 1), ('一則', 1), ('貼文', 1), ('掀起', 1), ('熱烈', 1), ('文中', 1), ('附上', 1), ('超市', 1), ('鮮奶區', 1), ('照片', 1), ('顯示', 1), ('風天', 1), ('前夕', 1), ('物資', 1), ('搶購', 1), ('一空', 1), ('味全', 1), ('依然', 1), ('引發眾', 1), ('多', 1), ('網友', 1), ('留言', 1), ('大家', 1), ('沒有', 1), ('忘記', 1), ('控遭', 1), ('丟包', 1), ('旗津風', 1), ('大雨', 1), ('大浪', 1), ('大民眾', 1), ('穿越', 1), ('封鎖線', 1), ('冒險', 1), ('觀浪', 1), ('抗戰防', 1), ('淹水', 1), ('銀行', 1), ('水閘門', 1), ('增高', 1), ('ATM', 1), ('被迫', 1), ('暫停', 1), ('台中明', 1), ('多數', 1), ('中榮開', 1), ('風特別', 1), ('逼近', 1), ('因氣', 1), ('象署', 1), ('預估', 1), ('明日', 1), ('符合', 1), ('停班課', 1), ('標準', 1), ('晚間', 1), ('點宣', 1), ('佈', 1), ('明天', 1), ('停止', 1), ('上班', 1), ('上課', 1), ('台中將', 1), ('包括', 1), ('中國', 1), ('醫藥', 1), ('大學', 1), ('附設', 1), ('台', 1), ('均', 1), ('又', 1), ('要', 1), ('轉彎', 1), ('日本', 1), ('預測', 1), ('回馬槍', 1), ('再襲', 1), ('南部', 1), ('民間', 1), ('傳說', 1), ('恐破功', 1), ('風滯', 1), ('留台灣', 1), ('西南部', 1), ('許久', 1), ('今', 1), ('日終', 1), ('動起', 1), ('來', 1), ('北上', 1), ('雖然', 1), ('中心', 1), ('還是', 1), ('有些', 1), ('許滯', 1), ('留狀況', 1), ('不過', 1), ('暴風圈', 1), ('進入', 1), ('雲林', 1), ('東陸', 1), ('對', 1), ('新竹', 1), ('南投', 1), ('宜蘭', 1), ('花蓮', 1), ('東地區', 1), ('綠扯', 1), ('氣象署', 1), ('曝原', 1), ('估有', 1), ('風雨', 1), ('縣市', 1), ('顧及', 1), ('市民', 1), ('安全', 1), ('風山陀兒', 1), ('影響', 1), ('基北北桃', 1), ('宣布', 1), ('月', 1), ('號放', 1), ('這個', 1), ('決定', 1), ('是否', 1), ('達', 1), ('放假', 1), ('條件', 1), ('引發', 1), ('讓', 1), ('綠營', 1), ('質疑', 1), ('是不是', 1), ('趁機', 1), ('謝國', 1), ('對此', 1), ('台北市', 1), ('蔣萬安', 1), ('還原', 1), ('決策', 1), ('過程', 1), ('襲台', 1), ('恐衝擊', 1), ('薪水', 1), ('怠速', 1), ('前進', 1), ('目前', 1), ('會不會', 1), ('連續', 1), ('放', 1), ('以上', 1), ('風假還', 1), ('很', 1), ('難', 1), ('有民間', 1), ('單位', 1), ('苦惱', 1), ('這樣', 1), ('下去', 1), ('日薪', 1), ('資', 1), ('可能', 1), ('無法', 1), ('如期', 1), ('位', 1), ('高雄', 1), ('方舟', 1), ('就業', 1), ('服務協會', 1), ('理事', 1), ('王姿琦', 1), ('急', 1), ('搶', 1), ('修', 1), ('枋寮', 1), ('狂浪', 1), ('沖', 1), ('垮', 1), ('堤防', 1), ('30', 1), ('尺', 1), ('缺口', 1), ('釀', 1), ('海水', 1), ('倒灌', 1), ('挾強', 1), ('豪雨', 1), ('襲台東', 1), ('卑南', 1), ('準備', 1), ('進家門', 1), ('遭陸', 1), ('變弱', 1), ('急籲', 1), ('不可', 1), ('輕忽', 1), ('記者', 1), ('高', 1), ('珞曦', 1), ('綜合', 1), ('報導', 1), ('即將', 1), ('登陸', 1), ('傍晚', 1), ('起風雨', 1), ('逐漸', 1), ('明顯', 1), ('粉專', 1), ('風論壇', 1), ('天氣', 1), ('特急', 1), ('表示', 1), ('因移動', 1), ('速度', 1), ('緩慢', 1), ('東側', 1), ('陸地', 1), ('強度', 1), ('減弱', 1), ('其', 1), ('迎王', 1), ('千歲爺', 1), ('指示', 1), ('王船', 1), ('法會', 1), ('等', 1), ('按表', 1), ('進行', 1), ('三年', 1), ('一科', 1), ('迎', 1), ('王平安', 1), ('遇到', 1), ('東隆宮', 1), ('今晚', 1), ('聲明', 1), ('經向代天', 1), ('巡', 1), ('狩千歲爺', 1), ('擲', 1), ('杯', 1), ('請示', 1), ('代', 1), ('天府', 1), ('內', 1), ('各項', 1), ('將依', 1), ('順', 1)])\n",
      "<class 'collections.OrderedDict'>\n",
      "字典數量: 360\n",
      "------------------------------------------------------------------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "# 使用 BIF Counter 計算詞頻\n",
    "word_count = Counter(filtered_words)\n",
    "print(word_count)\n",
    "print(type(word_count))\n",
    "print(dir(word_count))\n",
    "\n",
    "print(\"--\"*60)\n",
    "\n",
    "print(\"word_count.items():\")\n",
    "print(word_count.items())\n",
    "print(type(word_count.items()))\n",
    "\n",
    "print(\"--\"*60)\n",
    "\n",
    "# 將 word_count 轉換為 dict\n",
    "# word_count_dict = dict(word_count)\n",
    "# 欸？ counter 轉 dict 後，key 的順序改變了\n",
    "\n",
    "from collections import OrderedDict\n",
    "word_count_dict = OrderedDict(word_count.most_common())\n",
    "\n",
    "print(\"word_count_dict:\")\n",
    "print(word_count_dict)\n",
    "print(type(word_count_dict))\n",
    "print(f\"字典數量: {len(word_count_dict)}\")\n",
    "print(\"--\"*60)\n",
    "\n",
    " "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 從字典取出資料"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "字典可以使用的方法\n",
    "['clear', 'copy', 'fromkeys', 'get', 'items', 'keys', \n",
    "\n",
    "'pop', 'popitem', 'setdefault', 'update', 'values']\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "------------------------------------------------------------------------------------------------------------------------\n",
      "颱 33\n",
      "山陀兒 22\n",
      "風 18\n",
      "風假 5\n",
      "但 4\n",
      "10 4\n",
      "偏少 3\n",
      "正常 3\n",
      "到 3\n",
      "說 3\n",
      "newsebc 3\n",
      "醫院 3\n",
      "門診 3\n",
      "及 3\n",
      "東港 3\n",
      "生成 2\n",
      "18 2\n",
      "氣象 2\n",
      "專家 2\n",
      "已 2\n",
      "月份 2\n",
      "鮮奶 2\n",
      "滯銷 2\n",
      "感謝 2\n",
      "台灣 2\n",
      "討論 2\n",
      "被 2\n",
      "吳宇 2\n",
      "舒晚 2\n",
      "報 2\n",
      "風擾 2\n",
      "郵輪行 2\n",
      "好會員 2\n",
      "免費 2\n",
      "搭 2\n",
      "旅客 2\n",
      "晚報 2\n",
      "2024 2\n",
      "服務 2\n",
      "台中市 2\n",
      "連放 2\n",
      "天 2\n",
      "中 2\n",
      "日 2\n",
      "於 2\n",
      "以南 2\n",
      "臺 2\n",
      "地 2\n",
      "救 2\n",
      "樑 2\n",
      "持續 2\n",
      "長 2\n",
      "發放 2\n",
      "獨家 2\n",
      "牛步 2\n",
      "進逼蘭嶼 2\n",
      "擱淺 2\n",
      "貨輪 2\n",
      "慘斷 2\n",
      "兩截 2\n",
      "核心 2\n",
      "摩擦 2\n",
      "攪局 2\n",
      "科儀 2\n",
      "祭典 2\n",
      "至今 1\n",
      "僅 1\n",
      "範圍 1\n",
      "十點 1\n",
      "不 1\n",
      "一樣 1\n",
      "20241002 1\n",
      "TVBSNEWS01 1\n",
      "歷史 1\n",
      "紀錄 1\n",
      "平均值 1\n",
      "西北 1\n",
      "太平洋 1\n",
      "風數 1\n",
      "265 1\n",
      "今年 1\n",
      "都 1\n",
      "經來 1\n",
      "算是 1\n",
      "第 1\n",
      "似乎 1\n",
      "有點 1\n",
      "引起 1\n",
      "各方 1\n",
      "關注 1\n",
      "而且 1\n",
      "更是 1\n",
      "從 1\n",
      "開始 1\n",
      "像 1\n",
      "趕 1\n",
      "進度 1\n",
      "似的 1\n",
      "風天物資 1\n",
      "搶光 1\n",
      "這牌 1\n",
      "網 1\n",
      "人 1\n",
      "沒 1\n",
      "忘 1\n",
      "記過 1\n",
      "去 1\n",
      "吳裕堯 1\n",
      "核稿 1\n",
      "編輯 1\n",
      "社群 1\n",
      "平台 1\n",
      "Threads 1\n",
      "一則 1\n",
      "貼文 1\n",
      "掀起 1\n",
      "熱烈 1\n",
      "文中 1\n",
      "附上 1\n",
      "超市 1\n",
      "鮮奶區 1\n",
      "照片 1\n",
      "顯示 1\n",
      "風天 1\n",
      "前夕 1\n",
      "物資 1\n",
      "搶購 1\n",
      "一空 1\n",
      "味全 1\n",
      "依然 1\n",
      "引發眾 1\n",
      "多 1\n",
      "網友 1\n",
      "留言 1\n",
      "大家 1\n",
      "沒有 1\n",
      "忘記 1\n",
      "控遭 1\n",
      "丟包 1\n",
      "旗津風 1\n",
      "大雨 1\n",
      "大浪 1\n",
      "大民眾 1\n",
      "穿越 1\n",
      "封鎖線 1\n",
      "冒險 1\n",
      "觀浪 1\n",
      "抗戰防 1\n",
      "淹水 1\n",
      "銀行 1\n",
      "水閘門 1\n",
      "增高 1\n",
      "ATM 1\n",
      "被迫 1\n",
      "暫停 1\n",
      "台中明 1\n",
      "多數 1\n",
      "中榮開 1\n",
      "風特別 1\n",
      "逼近 1\n",
      "因氣 1\n",
      "象署 1\n",
      "預估 1\n",
      "明日 1\n",
      "符合 1\n",
      "停班課 1\n",
      "標準 1\n",
      "晚間 1\n",
      "點宣 1\n",
      "佈 1\n",
      "明天 1\n",
      "停止 1\n",
      "上班 1\n",
      "上課 1\n",
      "台中將 1\n",
      "包括 1\n",
      "中國 1\n",
      "醫藥 1\n",
      "大學 1\n",
      "附設 1\n",
      "台 1\n",
      "均 1\n",
      "又 1\n",
      "要 1\n",
      "轉彎 1\n",
      "日本 1\n",
      "預測 1\n",
      "回馬槍 1\n",
      "再襲 1\n",
      "南部 1\n",
      "民間 1\n",
      "傳說 1\n",
      "恐破功 1\n",
      "風滯 1\n",
      "留台灣 1\n",
      "西南部 1\n",
      "許久 1\n",
      "今 1\n",
      "日終 1\n",
      "動起 1\n",
      "來 1\n",
      "北上 1\n",
      "雖然 1\n",
      "中心 1\n",
      "還是 1\n",
      "有些 1\n",
      "許滯 1\n",
      "留狀況 1\n",
      "不過 1\n",
      "暴風圈 1\n",
      "進入 1\n",
      "雲林 1\n",
      "東陸 1\n",
      "對 1\n",
      "新竹 1\n",
      "南投 1\n",
      "宜蘭 1\n",
      "花蓮 1\n",
      "東地區 1\n",
      "綠扯 1\n",
      "氣象署 1\n",
      "曝原 1\n",
      "估有 1\n",
      "風雨 1\n",
      "縣市 1\n",
      "顧及 1\n",
      "市民 1\n",
      "安全 1\n",
      "風山陀兒 1\n",
      "影響 1\n",
      "基北北桃 1\n",
      "宣布 1\n",
      "月 1\n",
      "號放 1\n",
      "這個 1\n",
      "決定 1\n",
      "是否 1\n",
      "達 1\n",
      "放假 1\n",
      "條件 1\n",
      "引發 1\n",
      "讓 1\n",
      "綠營 1\n",
      "質疑 1\n",
      "是不是 1\n",
      "趁機 1\n",
      "謝國 1\n",
      "對此 1\n",
      "台北市 1\n",
      "蔣萬安 1\n",
      "還原 1\n",
      "決策 1\n",
      "過程 1\n",
      "襲台 1\n",
      "恐衝擊 1\n",
      "薪水 1\n",
      "怠速 1\n",
      "前進 1\n",
      "目前 1\n",
      "會不會 1\n",
      "連續 1\n",
      "放 1\n",
      "以上 1\n",
      "風假還 1\n",
      "很 1\n",
      "難 1\n",
      "有民間 1\n",
      "單位 1\n",
      "苦惱 1\n",
      "這樣 1\n",
      "下去 1\n",
      "日薪 1\n",
      "資 1\n",
      "可能 1\n",
      "無法 1\n",
      "如期 1\n",
      "位 1\n",
      "高雄 1\n",
      "方舟 1\n",
      "就業 1\n",
      "服務協會 1\n",
      "理事 1\n",
      "王姿琦 1\n",
      "急 1\n",
      "搶 1\n",
      "修 1\n",
      "枋寮 1\n",
      "狂浪 1\n",
      "沖 1\n",
      "垮 1\n",
      "堤防 1\n",
      "30 1\n",
      "尺 1\n",
      "缺口 1\n",
      "釀 1\n",
      "海水 1\n",
      "倒灌 1\n",
      "挾強 1\n",
      "豪雨 1\n",
      "襲台東 1\n",
      "卑南 1\n",
      "準備 1\n",
      "進家門 1\n",
      "遭陸 1\n",
      "變弱 1\n",
      "急籲 1\n",
      "不可 1\n",
      "輕忽 1\n",
      "記者 1\n",
      "高 1\n",
      "珞曦 1\n",
      "綜合 1\n",
      "報導 1\n",
      "即將 1\n",
      "登陸 1\n",
      "傍晚 1\n",
      "起風雨 1\n",
      "逐漸 1\n",
      "明顯 1\n",
      "粉專 1\n",
      "風論壇 1\n",
      "天氣 1\n",
      "特急 1\n",
      "表示 1\n",
      "因移動 1\n",
      "速度 1\n",
      "緩慢 1\n",
      "東側 1\n",
      "陸地 1\n",
      "強度 1\n",
      "減弱 1\n",
      "其 1\n",
      "迎王 1\n",
      "千歲爺 1\n",
      "指示 1\n",
      "王船 1\n",
      "法會 1\n",
      "等 1\n",
      "按表 1\n",
      "進行 1\n",
      "三年 1\n",
      "一科 1\n",
      "迎 1\n",
      "王平安 1\n",
      "遇到 1\n",
      "東隆宮 1\n",
      "今晚 1\n",
      "聲明 1\n",
      "經向代天 1\n",
      "巡 1\n",
      "狩千歲爺 1\n",
      "擲 1\n",
      "杯 1\n",
      "請示 1\n",
      "代 1\n",
      "天府 1\n",
      "內 1\n",
      "各項 1\n",
      "將依 1\n",
      "順 1\n",
      "['颱', '山陀兒', '風', '風假', '但', '10', '偏少', '正常', '到', '說', 'newsebc', '醫院', '門診', '及', '東港', '生成', '18', '氣象', '專家', '已', '月份', '鮮奶', '滯銷', '感謝', '台灣', '討論', '被', '吳宇', '舒晚', '報', '風擾', '郵輪行', '好會員', '免費', '搭', '旅客', '晚報', '2024', '服務', '台中市', '連放', '天', '中', '日', '於', '以南', '臺', '地', '救', '樑', '持續', '長', '發放', '獨家', '牛步', '進逼蘭嶼', '擱淺', '貨輪', '慘斷', '兩截', '核心', '摩擦', '攪局', '科儀', '祭典', '至今', '僅', '範圍', '十點', '不', '一樣', '20241002', 'TVBSNEWS01', '歷史', '紀錄', '平均值', '西北', '太平洋', '風數', '265', '今年', '都', '經來', '算是', '第', '似乎', '有點', '引起', '各方', '關注', '而且', '更是', '從', '開始', '像', '趕', '進度', '似的', '風天物資', '搶光', '這牌', '網', '人', '沒', '忘', '記過', '去', '吳裕堯', '核稿', '編輯', '社群', '平台', 'Threads', '一則', '貼文', '掀起', '熱烈', '文中', '附上', '超市', '鮮奶區', '照片', '顯示', '風天', '前夕', '物資', '搶購', '一空', '味全', '依然', '引發眾', '多', '網友', '留言', '大家', '沒有', '忘記', '控遭', '丟包', '旗津風', '大雨', '大浪', '大民眾', '穿越', '封鎖線', '冒險', '觀浪', '抗戰防', '淹水', '銀行', '水閘門', '增高', 'ATM', '被迫', '暫停', '台中明', '多數', '中榮開', '風特別', '逼近', '因氣', '象署', '預估', '明日', '符合', '停班課', '標準', '晚間', '點宣', '佈', '明天', '停止', '上班', '上課', '台中將', '包括', '中國', '醫藥', '大學', '附設', '台', '均', '又', '要', '轉彎', '日本', '預測', '回馬槍', '再襲', '南部', '民間', '傳說', '恐破功', '風滯', '留台灣', '西南部', '許久', '今', '日終', '動起', '來', '北上', '雖然', '中心', '還是', '有些', '許滯', '留狀況', '不過', '暴風圈', '進入', '雲林', '東陸', '對', '新竹', '南投', '宜蘭', '花蓮', '東地區', '綠扯', '氣象署', '曝原', '估有', '風雨', '縣市', '顧及', '市民', '安全', '風山陀兒', '影響', '基北北桃', '宣布', '月', '號放', '這個', '決定', '是否', '達', '放假', '條件', '引發', '讓', '綠營', '質疑', '是不是', '趁機', '謝國', '對此', '台北市', '蔣萬安', '還原', '決策', '過程', '襲台', '恐衝擊', '薪水', '怠速', '前進', '目前', '會不會', '連續', '放', '以上', '風假還', '很', '難', '有民間', '單位', '苦惱', '這樣', '下去', '日薪', '資', '可能', '無法', '如期', '位', '高雄', '方舟', '就業', '服務協會', '理事', '王姿琦', '急', '搶', '修', '枋寮', '狂浪', '沖', '垮', '堤防', '30', '尺', '缺口', '釀', '海水', '倒灌', '挾強', '豪雨', '襲台東', '卑南', '準備', '進家門', '遭陸', '變弱', '急籲', '不可', '輕忽', '記者', '高', '珞曦', '綜合', '報導', '即將', '登陸', '傍晚', '起風雨', '逐漸', '明顯', '粉專', '風論壇', '天氣', '特急', '表示', '因移動', '速度', '緩慢', '東側', '陸地', '強度', '減弱', '其', '迎王', '千歲爺', '指示', '王船', '法會', '等', '按表', '進行', '三年', '一科', '迎', '王平安', '遇到', '東隆宮', '今晚', '聲明', '經向代天', '巡', '狩千歲爺', '擲', '杯', '請示', '代', '天府', '內', '各項', '將依', '順']\n",
      "[33, 22, 18, 5, 4, 4, 3, 3, 3, 3, 3, 3, 3, 3, 3, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]\n",
      "------------------------------------------------------------------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "# print(dir(word_count_dict))\n",
    "# print(word_count_dict.items())\n",
    "# print(type(word_count_dict.items()))\n",
    "# print(word_count_dict.keys())\n",
    "# words = list(word_count_dict.keys())\n",
    "# counts = list(word_count_dict.values())\n",
    "\n",
    "# items ===> 同時取出 key 和 value 就可以直接加入 dataframe\n",
    "\n",
    "words = []\n",
    "counts = []\n",
    "\n",
    "print(\"--\"*60)\n",
    "\n",
    "for key in word_count_dict:\n",
    "    print(key, word_count_dict[key])\n",
    "    words.append(key)\n",
    "    counts.append(word_count_dict[key])\n",
    "\n",
    "# words = [key for key in word_count_dict]\n",
    "# counts = [word_count_dict[key] for key in word_count_dict]\n",
    "\n",
    "print(words)\n",
    "print(counts)\n",
    "\n",
    "print(\"--\"*60)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 使用 Pandas 排序呈現資料"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "練習用 Series 方式加入\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.series.Series'>\n",
      "Index: 360 entries, 颱 to 順\n",
      "Series name: 一維陣列詞頻Series\n",
      "Non-Null Count  Dtype\n",
      "--------------  -----\n",
      "360 non-null    int64\n",
      "dtypes: int64(1)\n",
      "memory usage: 5.6+ KB\n",
      "None\n",
      "<bound method NDFrame.head of 颱      33\n",
      "山陀兒    22\n",
      "風      18\n",
      "風假      5\n",
      "但       4\n",
      "       ..\n",
      "天府      1\n",
      "內       1\n",
      "各項      1\n",
      "將依      1\n",
      "順       1\n",
      "Name: 一維陣列詞頻Series, Length: 360, dtype: int64>\n",
      "------------------------------------------------------------------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "# 創建一個 pandas Series，使用單字作為索引，詞頻作為值\n",
    "word_count_series = pd.Series(data=counts, index=words, name='一維陣列詞頻Series')\n",
    "print(word_count_series.info())\n",
    "print(word_count_series.head)\n",
    "print(\"--\"*60)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.series.Series'>\n",
      "RangeIndex: 360 entries, 0 to 359\n",
      "Series name: 一維陣列文字Series\n",
      "Non-Null Count  Dtype \n",
      "--------------  ----- \n",
      "360 non-null    object\n",
      "dtypes: object(1)\n",
      "memory usage: 2.9+ KB\n",
      "None\n",
      "------------------------------------------------------------------------------------------------------------------------\n",
      "0      颱\n",
      "1    山陀兒\n",
      "2      風\n",
      "3     風假\n",
      "4      但\n",
      "Name: 一維陣列文字Series, dtype: object\n"
     ]
    }
   ],
   "source": [
    "word_text_series = pd.Series(data=words, name='一維陣列文字Series')\n",
    "print(word_text_series.info())\n",
    "\n",
    "print(\"--\"*60)\n",
    "print(word_text_series.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 360 entries, 0 to 359\n",
      "Data columns (total 2 columns):\n",
      " #   Column  Non-Null Count  Dtype \n",
      "---  ------  --------------  ----- \n",
      " 0   word    360 non-null    object\n",
      " 1   count   360 non-null    int64 \n",
      "dtypes: int64(1), object(1)\n",
      "memory usage: 5.8+ KB\n",
      "None\n",
      "  word  count\n",
      "0    颱     33\n",
      "1  山陀兒     22\n",
      "2    風     18\n",
      "3   風假      5\n",
      "4    但      4\n"
     ]
    }
   ],
   "source": [
    "# 將單一 Series 轉換為 DataFrame\n",
    "df = word_count_series.reset_index()\n",
    "df.columns = ['word', 'count']  # 重命名列\n",
    "print(df.info())\n",
    "print(df.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 360 entries, 0 to 359\n",
      "Data columns (total 2 columns):\n",
      " #   Column  Non-Null Count  Dtype \n",
      "---  ------  --------------  ----- \n",
      " 0   word    360 non-null    object\n",
      " 1   count   360 non-null    int64 \n",
      "dtypes: int64(1), object(1)\n",
      "memory usage: 5.8+ KB\n",
      "None\n",
      "DataFrame 預覽：\n",
      "  word  count\n",
      "0    颱     33\n",
      "1  山陀兒     22\n",
      "2    風     18\n",
      "3   風假      5\n",
      "4    但      4\n"
     ]
    }
   ],
   "source": [
    "# 將多個 Series 轉換為 DataFrame\n",
    "\n",
    "# 重置兩個 Series 的索引\n",
    "word_text_series = word_text_series.reset_index(drop=True)\n",
    "word_count_series = word_count_series.reset_index(drop=True)\n",
    "\n",
    "# 使用 concat 創建 DataFrame\n",
    "df = pd.concat([word_text_series, word_count_series], axis=1)\n",
    "\n",
    "# 重命名列\n",
    "df.columns = ['word', 'count']\n",
    "\n",
    "# 顯示 DataFrame 的資訊\n",
    "print(df.info())\n",
    "\n",
    "# 顯示 DataFrame 的前幾行\n",
    "print(\"DataFrame 預覽：\")\n",
    "print(df.head())\n",
    "\n",
    "# # 如果需要，可以按計數降序排序\n",
    "# df_sorted = df.sort_values('count', ascending=False).reset_index(drop=True)\n",
    "\n",
    "# # 顯示排序後的 DataFrame 前幾行\n",
    "# print(\"\\n排序後：\")\n",
    "# print(df_sorted.head())\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "df2:\n",
      "    word  count\n",
      "0      颱     33\n",
      "1    山陀兒     22\n",
      "2      風     18\n",
      "3     風假      5\n",
      "4      但      4\n",
      "..   ...    ...\n",
      "355   天府      1\n",
      "356    內      1\n",
      "357   各項      1\n",
      "358   將依      1\n",
      "359    順      1\n",
      "\n",
      "[360 rows x 2 columns]\n",
      "------------------------------------------------------------------------------------------------------------------------\n",
      "\n",
      "df_sorted 資料結構:\n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 360 entries, 0 to 359\n",
      "Data columns (total 2 columns):\n",
      " #   Column  Non-Null Count  Dtype \n",
      "---  ------  --------------  ----- \n",
      " 0   word    360 non-null    object\n",
      " 1   count   360 non-null    int64 \n",
      "dtypes: int64(1), object(1)\n",
      "memory usage: 5.8+ KB\n",
      "None\n",
      "\n",
      "df_sorted:\n",
      "<bound method NDFrame.head of     word  count\n",
      "0      颱     33\n",
      "1    山陀兒     22\n",
      "2      風     18\n",
      "3     風假      5\n",
      "4      但      4\n",
      "..   ...    ...\n",
      "355   象署      1\n",
      "356   因氣      1\n",
      "357   逼近      1\n",
      "358  風特別      1\n",
      "359    順      1\n",
      "\n",
      "[360 rows x 2 columns]>\n",
      "\n",
      "df_sorted 索引:\n",
      "RangeIndex(start=0, stop=360, step=1)\n"
     ]
    }
   ],
   "source": [
    "# 將字典轉 list 放入 DataFrame\n",
    "df2 = pd.DataFrame(list(word_count_dict.items()), columns=['word', 'count'])\n",
    "\n",
    "\n",
    "# df_sorted_org = df.sort_values(by='count', ascending=False)\n",
    "df_sorted = df2.sort_values(by='count', ascending=False).reset_index(drop=True)\n",
    "\n",
    "print(\"df2:\")\n",
    "print(df2)\n",
    "print(\"--\"*60)\n",
    "print(\"\\ndf_sorted 資料結構:\")\n",
    "print(df_sorted.info())\n",
    "print(\"\\ndf_sorted:\")\n",
    "print(df_sorted.head)\n",
    "print(\"\\ndf_sorted 索引:\")\n",
    "print(df_sorted.index)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 製作文字雲圖片遮罩"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/Users/larry/Github/Python-WordCloud/TaipeiSansTCBeta-Bold.ttf\n",
      "(540, 960)\n",
      "[[0. 0. 0. ... 0. 0. 0.]\n",
      " [0. 0. 0. ... 0. 0. 0.]\n",
      " [0. 0. 0. ... 0. 0. 0.]\n",
      " ...\n",
      " [0. 0. 0. ... 0. 0. 0.]\n",
      " [0. 0. 0. ... 0. 0. 0.]\n",
      " [0. 0. 0. ... 0. 0. 0.]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "原始x色像素數量：144292，總像素數：1555200（佔比：9.28%）\n",
      "處理後x色像素數量：144292，總像素數：1555200（佔比：9.28%）\n"
     ]
    }
   ],
   "source": [
    "# 正常圖片版本\n",
    "\n",
    "# from PIL import ImageOps # 反轉黑白\n",
    "# ImageOps.invert()\n",
    "\n",
    "# 詢問使用者字體檔案路徑\n",
    "# font_path_input = input(\"請輸入中文字體的檔案路徑: \")\n",
    "font_path_input = \"/Users/larry/Github/Python-WordCloud/TaipeiSansTCBeta-Bold.ttf\"\n",
    "print(font_path_input)\n",
    "\n",
    "# 文字雲圖片遮罩檔案路徑\n",
    "img_path = \"./Mask_Image/Messenge_mask.png\"\n",
    "\n",
    "\n",
    "mask_color = np.array((Image.open(img_path).convert(\"RGB\")).resize((960, 540)))\n",
    "# mask_color = mask_color[::2, ::2]\n",
    "\n",
    "total_pixels = mask_color.size\n",
    "# total_pixels = mask_image.shape[0] * mask_image.shape[1]\n",
    "\n",
    "org_filtered_pixels = np.sum(np.all(mask_color == 0, axis=2))\n",
    "\n",
    "mask_image = mask_color.copy()\n",
    "\n",
    "threshold = 20 # 根據圖片調整閾值\n",
    "mask_image[np.all(mask_image > threshold, axis=2)] = 255\n",
    "# 只要是（偏）()色區域，都設為()色\n",
    "filtered_pixels = np.sum(np.all(mask_image == 0, axis=2))\n",
    "\n",
    "# 邊緣檢測 (進階用法，此部分為當初課堂老師提供)\n",
    "# 我的理解：將三個顏色通道做了某個處理，然後將三個通道的結果累加，最後除以 3 得到平均值\n",
    "edges = np.mean([gaussian_gradient_magnitude(mask_image[:, :, i] / 255., 1) for i in range(3)], axis=0)\n",
    "print(edges.shape)\n",
    "print(edges)\n",
    "# 使用平均值作為閾值\n",
    "threshold = edges.mean() * 0.2\n",
    "mask_image[edges > threshold] = 255\n",
    "\n",
    "\n",
    "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(10, 5))\n",
    "# 顯示原始遮罩圖片\n",
    "ax1.imshow(mask_color)\n",
    "ax1.set_title(\"Original Mask\")\n",
    "ax1.axis('off')\n",
    "\n",
    "# 顯示處理後的遮罩圖片\n",
    "ax2.imshow(mask_image)\n",
    "ax2.set_title(\"Processed Mask\")\n",
    "ax2.axis('off')\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.show()\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "print(f\"原始x色像素數量：{org_filtered_pixels}，總像素數：{total_pixels}（佔比：{org_filtered_pixels/total_pixels*100:.2f}%）\")\n",
    "print(f\"處理後x色像素數量：{filtered_pixels}，總像素數：{total_pixels}（佔比：{filtered_pixels/total_pixels*100:.2f}%）\")\n",
    "\n",
    "# 將高於閾值的邊緣像素設為白色\n",
    "\n",
    "# 顏色生成\n",
    "image_colors = ImageColorGenerator(mask_image)\n",
    "# image_colors.default_color = [0.9,0.9,0.9]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 選擇黑色遮罩圖案練習處理圖片\n",
    "\n",
    "# # from PIL import ImageOps # 反轉黑白\n",
    "# # ImageOps.invert()\n",
    "\n",
    "# # 詢問使用者字體檔案路徑\n",
    "# # font_path_input = input(\"請輸入中文字體的檔案路徑: \")\n",
    "# font_path_input = \"/Users/larry/Github/Python-WordCloud/TaipeiSansTCBeta-Bold.ttf\"\n",
    "# print(font_path_input)\n",
    "# # 文字雲圖片遮罩檔案路徑\n",
    "# img_path = \"./Mask_Image/LeafDrak_mask.png\"\n",
    "\n",
    "# # 讀取遮罩圖片，轉換為 RGB，調整大小，然後轉換為 numpy 陣列\n",
    "# # mask_color = np.array(Image.open(img_path).convert(\"RGB\").resize((960, 540)))\n",
    "\n",
    "# # 轉灰階\n",
    "# mask_color = np.array((Image.open(img_path).convert(\"L\")).resize((960, 540)))\n",
    "# # 查看 mask_color 的類型和屬性\n",
    "# print(\"Type of mask_color:\", type(mask_color))\n",
    "# print(\"Shape of mask_color:\", mask_color.shape)\n",
    "# print(\"Data type of mask_color:\", mask_color.dtype)\n",
    "# # print(\"Attributes and methods of array:\", dir(mask_color))\n",
    "# total_pixels = mask_color.size\n",
    "# # org_filtered_pixels = np.sum(np.all(mask_image == 255, axis=2))\n",
    "# org_pixels = np.sum(mask_color == 255)\n",
    "\n",
    "# # 縮小\n",
    "# # # 每隔 n 個像素取一個像素\n",
    "# # # mask_color = mask_color[::4, ::4]\n",
    "\n",
    "# # # 放大（重複像素）\n",
    "# # mask_color_large = np.repeat(np.repeat(mask_color, 4, axis=0), 4, axis=1)\n",
    "\n",
    "# # 進行二次處理，保留灰階\n",
    "# mask_image = mask_color.copy()\n",
    "# # 不使用模組，numpy 反轉處理方式\n",
    "# # mask_image = 255 - mask_color\n",
    "# # 去噪應用\n",
    "# threshold1 = 190\n",
    "# threshold2 = 20\n",
    "# # mask_image[np.all(mask_image < threshold, axis=2)] = 255\n",
    "# mask_image[mask_image < threshold2] = 0\n",
    "# mask_image[mask_image > threshold1] = 255\n",
    "\n",
    "# # 只要是（偏）()色區域，都設為()色\n",
    "\n",
    "# # 高級用法，感謝 GPT 的解釋，真的是很厲害的做法....\n",
    "# # 0-255 變成 0-1 的數據，然後把 某個閾值以上的數據變成 1，剩下的變成 0\n",
    "# # 然後再次合成上去\n",
    "\n",
    "# # 對灰階圖像進行邊緣檢測\n",
    "\n",
    "# edges = gaussian_gradient_magnitude(mask_image, sigma=0.4)\n",
    "# # 設置邊緣閾值\n",
    "# edge_threshold = edges.mean()\n",
    "\n",
    "# # 創建邊緣遮罩 True 為 1，False 為 0 ---> 1 (資料格式)＊ 255 白色\n",
    "# edge_mask = (edges > edge_threshold).astype(np.uint8) * 255\n",
    "\n",
    "# # 將邊緣與原始遮罩結合\n",
    "# mask_image = np.maximum(mask_image, edge_mask)\n",
    "\n",
    "\n",
    "# # 三維陣列圖片邊緣檢測 (進階用法，此部分為當初課堂老師提供，要有一定的 numpy 操作陣列觀念後，才比較了解這個做法)\n",
    "\n",
    "# # 我的理解：將三個顏色通道做了某個處理，然後將三個通道的結果累加，最後除以 3 得到平均值\n",
    "# # edges = np.mean([gaussian_gradient_magnitude(mask_color[:, :, i] / 255., 2) for i in range(3)], axis=0)\n",
    "\n",
    "\n",
    "# fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(10, 5))\n",
    "# # 顯示原始遮罩圖片\n",
    "# ax1.imshow(mask_color, cmap='gray')\n",
    "# ax1.set_title(\"Original Mask\")\n",
    "# ax1.axis('off')\n",
    "\n",
    "# # 顯示處理後的遮罩圖片\n",
    "# ax2.imshow(mask_image, cmap='gray')\n",
    "# ax2.set_title(\"Processed Mask\")\n",
    "# ax2.axis('off')\n",
    "\n",
    "# plt.tight_layout()\n",
    "# plt.show()\n",
    "\n",
    "\n",
    "\n",
    "# # filtered_pixels = np.sum(np.all(mask_image == 255, axis=2))\n",
    "# filtered_pixels = np.sum(mask_image == 255)\n",
    "# # total_pixels = mask_image.shape[0] * mask_image.shape[1]\n",
    "\n",
    "\n",
    "# print(f\"原始x色像素數量：{org_pixels}，總像素數：{total_pixels}（佔比：{org_pixels/total_pixels*100:.2f}%）\")\n",
    "# print(f\"處理後x色像素數量：{filtered_pixels}，總像素數：{total_pixels}（佔比：{filtered_pixels/total_pixels*100:.2f}%）\")\n",
    "\n",
    "# # 將高於閾值的邊緣像素設為白色\n",
    "\n",
    "# # 顏色生成\n",
    "# # image_colors = ImageColorGenerator(mask_image)\n",
    "# # image_colors.default_color = [0.9,0.9,0.9]\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\"\"\"\n",
    "\n",
    "AI 回覆的內容，關於邊緣檢測：\n",
    "\n",
    "這個版本的代碼做了以下事情：\n",
    "\n",
    "1. 創建一個與單個顏色通道相同大小的零矩陣。\n",
    "2. 遍歷三個顏色通道（RGB）。\n",
    "3. 對每個通道計算高斯梯度幅值。\n",
    "4. 將每個通道的梯度累加到 `edges` 矩陣中。\n",
    "5. 最後，將累加的結果除以 3 來獲得平均值。\n",
    "\n",
    "這個方法實現了與原始代碼相同的功能，但避免了使用列表推導式和 `np.mean` 的 `axis` 參數，可能更容易理解每一步在做什麼。最後，仍然將高於閾值的邊緣像素設為白色，以在遮罩中突出顯示邊緣。\n",
    "\n",
    "\n",
    "- 高斯梯度幅值越低 → 越可能是平滑區域（非邊緣）\n",
    "\n",
    "- 高斯梯度幅值越高 → 越可能是邊緣\n",
    "\n",
    "我：邊緣設為白色不就是不明顯了嗎？\n",
    "\n",
    "您提出了一個很好的觀察。讓我解釋一下這個看似矛盾的做法的原因：\n",
    "\n",
    "1. 在生成詞雲的上下文中，白色（255）通常代表可以放置文字的區域。\n",
    "\n",
    "2. 將邊緣設為白色的目的不是為了在最終的詞雲中使邊緣更明顯，而是為了指導詞雲生成算法在哪裡可以放置文字。\n",
    "\n",
    "3. 通過將邊緣設為白色，我們實際上是告訴詞雲生成器：\"這些是圖像的重要特徵，你可以在這些地方放置文字\"。\n",
    "\n",
    "4. 這樣做的結果是，文字會傾向於沿著原始圖像的輪廓和主要特徵分布，而不是隨機分布或覆蓋重要的視覺元素。\n",
    "\n",
    "5. 在最終的詞雲中，這些邊緣不會直接可見，但它們會影響文字的排列方式，使得詞雲的整體形狀能夠反映原始圖像的輪廓。\n",
    "\n",
    "所以，雖然在遮罩圖像中邊緣被設為白色看起來不明顯，但這實際上是為了在生成詞雲時保留原始圖像的結構和形狀。最終的效果是，生成的詞雲會呈現出原始圖像的大致輪廓，而不是完全忽視圖像的特徵。\n",
    "\n",
    "如果您想要在最終的詞雲中使邊緣更加明顯，可以考慮相反的方法：將邊緣設為黑色或其他深色，這樣詞雲生成器就會避免在這些區域放置文字，從而在視覺上保留原始圖像的輪廓。\n",
    "\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "AI 回覆，關於 Pillow 的 Mode 參數\n",
    "\n",
    "\n",
    "1. \"L\" (Luminance) 模式：\n",
    "   - 每個像素用 0 到 255 的整數值表示，其中 0 是黑色，255 是白色。\n",
    "2. \"RGB\" (Red, Green, Blue) 模式：\n",
    "3. \"RGBA\" (Red, Green, Blue, Alpha) 模式：\n",
    "4. \"1\" (1-bit pixels, black and white) 模式：\n",
    "   - 每個像素只能是 0（黑）或 1（白）。\n",
    "\n",
    "<!-- 使用 Python Teseeract OCR OpenCV 模組處理圖片，有用過 -->"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(540, 960, 3)\n"
     ]
    }
   ],
   "source": [
    "# 因為轉灰階變成二維，再轉回去三維\n",
    "print(mask_image.shape)\n",
    "if len(mask_image.shape) == 2:\n",
    "    mask_image = np.stack((mask_image) * 3, axis=-1)\n",
    "    print(mask_image.shape)\n",
    "    print(mask_image)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mask shape: (540, 960, 3)\n",
      "Mask dtype: uint8\n",
      "Mask min value: 0\n",
      "Mask max value: 255\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1500x500 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# 檢查轉換後的遮罩圖片\n",
    "plt.figure(figsize=(10, 6))\n",
    "plt.imshow(mask_image)\n",
    "plt.title(\"Mask Image\")\n",
    "plt.axis('off')\n",
    "plt.show()\n",
    "\n",
    "# 打印遮罩圖片的形狀和一些基本統計信息\n",
    "print(\"Mask shape:\", mask_image.shape)\n",
    "print(\"Mask dtype:\", mask_image.dtype)\n",
    "print(\"Mask min value:\", mask_image.min())\n",
    "print(\"Mask max value:\", mask_image.max())\n",
    "\n",
    "# 如果需要查看 RGB 通道\n",
    "fig, (ax1, ax2, ax3, ax4) = plt.subplots(1, 4, figsize=(15, 5))\n",
    "# 顯示完整的 RGB 圖像\n",
    "ax1.imshow(mask_image)\n",
    "ax1.set_title('Full RGB Image')\n",
    "ax1.axis('off')\n",
    "\n",
    "# 顯示紅色通道\n",
    "ax2.imshow(mask_image[:,:,0], cmap='Reds')\n",
    "ax2.set_title('Red Channel')\n",
    "ax2.axis('off')\n",
    "\n",
    "# 顯示綠色通道\n",
    "ax3.imshow(mask_image[:,:,1], cmap='Greens')\n",
    "ax3.set_title('Green Channel')\n",
    "ax3.axis('off')\n",
    "\n",
    "# 顯示藍色通道\n",
    "ax4.imshow(mask_image[:,:,2], cmap='Blues')\n",
    "ax4.set_title('Blue Channel')\n",
    "ax4.axis('off')\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 生成文字雲"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "文字雲圖片已儲存為: ./img/20241002_颱風_title+desc_relative_scaling50%.png\n"
     ]
    }
   ],
   "source": [
    "# 生成文字雲\n",
    "\n",
    "# wordcloud = WordCloud(font_path=font_path_input, width=320, height=320,max_font_size=50,max_words=4000,color_func=image_colors).generate_from_frequencies(word_count)\n",
    "wordcloud = WordCloud(\n",
    "    font_path=font_path_input,\n",
    "    width=960,  # 增加寬度以提高畫質\n",
    "    height=540,  # 增加高度以提高畫質\n",
    "    mask=mask_image,\n",
    "    max_font_size=40,  # 增加最大字體大小\n",
    "    max_words=400,\n",
    "    # color_func=image_colors,\n",
    "    prefer_horizontal=0.9,  # 90% 的詞水平放置\n",
    "    scale=2,  # 增加縮放比例以提高畫質\n",
    "    min_font_size=10,  # 設置最小字體大小\n",
    "    background_color=\"black\",  # 設置背景顏色\n",
    "    mode=\"RGB\",\n",
    "    relative_scaling=0.3,  # 平衡詞頻和詞排名的重要性\n",
    "    collocations=True,  # 包含常見詞組\n",
    "    colormap=\"Spectral\",\n",
    "    contour_width=0.001,  # 添加輪廓以增強視覺效果\n",
    "    contour_color=\"white\"# 設置輪廓顏色\n",
    ").generate_from_frequencies(word_count_dict)\n",
    "plt.imshow(wordcloud)\n",
    "plt.axis(\"off\")\n",
    "plt.show()\n",
    "\n",
    "\n",
    "# 儲存文字雲圖片 檔案名稱格式化\n",
    "today = datetime.now().strftime('%Y%m%d')\n",
    "image_filename = f\"./img/{today}_{keyword}_title+desc_relative_scaling50%.png\"\n",
    "wordcloud.to_file(image_filename)\n",
    "print(f\"文字雲圖片已儲存為: {image_filename}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## WordCloud 參數釐清\n",
    "\n",
    "AI 回覆的內容，關於參數說明：\n",
    "\n",
    "您提出了一個很好的問題，這確實需要進一步澄清。您是對的，排名通常是基於詞頻的，但它們在詞雲生成中的作用略有不同。讓我解釋得更清楚：\n",
    "\n",
    "1. 排名 vs 詞頻：\n",
    "   - 排名：詞的相對位置（第1常見、第2常見等）\n",
    "   - 詞頻：詞出現的實際次數\n",
    "\n",
    "2. 差異的重要性：\n",
    "   - 排名只反映相對順序，不顯示頻率差距\n",
    "   - 詞頻反映實際出現次數，顯示詞與詞之間的差距\n",
    "\n",
    "3. 舉例說明：\n",
    "   假設有三個詞：\n",
    "   - A：出現100次（排名1）\n",
    "   - B：出現99次（排名2）\n",
    "   - C：出現10次（排名3）\n",
    "\n",
    "   - 純基於排名（0.0）：A、B、C的大小差異很小\n",
    "   - 純基於詞頻（1.0）：A和B幾乎一樣大，C明顯小得多\n",
    "\n",
    "4. `relative_scaling` 的作用：\n",
    "   - 0.0：主要考慮排名，A、B、C大小差異不大\n",
    "   - 1.0：完全按詞頻，A、B很大且相近，C很小\n",
    "   - 0.5：平衡兩者，反映排名，也部分反映頻率差異\n",
    "\n",
    "所以，雖然排名確實基於詞頻，但 `relative_scaling` 參數允許您決定是更注重詞的相對順序（排名），還是更注重詞之間的實際頻率差異。這提供了在視覺呈現上的靈活性。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 其他資訊確認"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Current working directory: /Users/larry/Github/Python-WordCloud\n",
      "Font file exists: True\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "\n",
    "print(\"Current working directory:\", os.getcwd())\n",
    "print(\"Font file exists:\", os.path.exists(font_path_input))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "(2024/09/17更新)\n",
    "\n",
    "- `get_news()` 函數目前返回的是 Google 頁面轉址連結\n",
    "- 參考 [SuYenTing/Python-web-crawler](https://github.com/SuYenTing/Python-web-crawler/blob/main/google_real_time_news.py) 專案，使用 google-news-url-decoder 模組解決轉址問題\n",
    "- `search()` 函數可以獲取真正的新聞網址\n",
    "- 非英文語言搜尋時，無法返回正確的 datetime 時間格式"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "aee8b7b246df8f9039afb4144a1f6fd8d2ca17a180786b69acc140d282b71a49"
  },
  "kernelspec": {
   "display_name": "Python 3.9.7 64-bit",
   "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": "3.10.15"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}