scraping data inipasti.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "#import library\n",
    "import urllib\n",
    "import requests\n",
    "import bs4\n",
    "import pandas as pd\n",
    "import re"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# request a link\n",
    "def request_url(link):\n",
    "    response = requests.get(link)\n",
    "    html = response.text\n",
    "    return html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function to parse html\n",
    "def parse_html(to_parse):\n",
    "    soup = bs4.BeautifulSoup(to_parse, 'html.parser')\n",
    "    return soup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "#take sub menu news form inipasti.com\n",
    "def all_section(main_url):\n",
    "    section_list = []\n",
    "    for i in main_url:\n",
    "        soup = parse_html(request_url(i))\n",
    "        for a in soup.find_all('a', href=True): \n",
    "            if a.text:\n",
    "                section_list.append(a['href'])\n",
    "    return section_list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def find_corona(main_url):\n",
    "    url_list = []\n",
    "    not_news=[]\n",
    "    for i in main_url:\n",
    "        if i.find('corona')!=-1 or i.find('covid')!=-1 or i.find('pandemi')!=-1 or i.find('psbb')!=-1 or i.find('social-distancing')!=-1:\n",
    "            url_list.append(i)\n",
    "    url_list = list(dict.fromkeys(url_list))\n",
    "    return url_list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def writer(main_url):\n",
    "    author = []\n",
    "    for i in main_url:\n",
    "        a = parse_html(request_url(i))\n",
    "        try:\n",
    "            b = a.find(\"div\", class_ = \"td-author-name vcard author\")    \n",
    "            if b:\n",
    "                c = b.find(\"a\")\n",
    "                author.append(c.text)\n",
    "            else:\n",
    "                author.append('None')\n",
    "        except :\n",
    "            author.append('None')\n",
    "            \n",
    "    return author"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def title(main_url):\n",
    "    titles = []\n",
    "    for i in main_url:\n",
    "        soup = parse_html(request_url(i))\n",
    "        try :\n",
    "            h1 = soup.select('h1', {'class' : 'entry-title td-module-title'})[0].text.strip()\n",
    "            titles.append(h1)\n",
    "        except :\n",
    "            titles.append('None')\n",
    "    return titles"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def date(main_url):\n",
    "    dates = []\n",
    "    for i in main_url:\n",
    "        a = parse_html(request_url(i))\n",
    "        try:\n",
    "            b = a.find(\"span\", class_ = \"td-post-date\").text\n",
    "            c = re.sub(\"Inipasti.com - \", \"\", b)\n",
    "            dates.append(c)\n",
    "        except:\n",
    "            dates.append('None')\n",
    "    return dates"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "def collect_text(main_url, titles = [], author = [], dates = []):\n",
    "    paragraf = []\n",
    "    data = []\n",
    "    join =[]\n",
    "    isiteks = []\n",
    "    #collect all text with tag <p> from articles    \n",
    "    for i,j in enumerate(main_url):\n",
    "        a = parse_html(request_url(j))\n",
    "        try:\n",
    "            content = a.find('div', class_='td-post-content')\n",
    "            pragraf = content.find_all('p')\n",
    "            for k in pragraf:\n",
    "                paragraf.append(k.text)\n",
    "            data.append(paragraf)\n",
    "            paragraf = []\n",
    "        except:\n",
    "            data.append('None')\n",
    "    for i in data:\n",
    "        join.append(' '.join(i))\n",
    "        \n",
    "    #collect articles and urls in one variable  \n",
    "    for i, j in enumerate(main_url):\n",
    "        isiteks.append({'news' : 'Inipasti.com', 'link' : j, 'title' : titles[i], 'author' : author[i], 'date_time' : dates[i], 'paragraf' : join[i] })\n",
    "    return isiteks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_news(main_url, file_name = ''):\n",
    "    section = all_section(main_url)\n",
    "    corona_news = find_corona(section)\n",
    "    titles = title(corona_news)\n",
    "    author = writer(corona_news)\n",
    "    datetime = date(corona_news)\n",
    "    text = collect_text(corona_news, titles, author, datetime)\n",
    "    file = save_file(text, file_name)\n",
    "    \n",
    "    return file  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def save_file(file, file_name = ''):\n",
    "    new_file = pd.DataFrame(file, columns=['news','link','title', 'author', 'date_time', 'paragraf'])\n",
    "    new_file.to_csv(file_name + '.csv', index=True, encoding='utf-8', sep = ',')\n",
    "    \n",
    "    return new_file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "main_url=[\"https://inipasti.com/\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "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>news</th>\n",
       "      <th>link</th>\n",
       "      <th>title</th>\n",
       "      <th>author</th>\n",
       "      <th>date_time</th>\n",
       "      <th>paragraf</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/category/kesehatan/corona...</td>\n",
       "      <td>Coronavirus</td>\n",
       "      <td>None</td>\n",
       "      <td>Mei 1, 2020 2:10 pm</td>\n",
       "      <td>N o n e</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/dalam-bayang-bayang-coron...</td>\n",
       "      <td>Dalam Bayang-bayang Coronavirus, Umat Islam di...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>April 21, 2020 10:37 am</td>\n",
       "      <td>INIPASTI.COM, ALGIERS / CAIRO / JAKARTA – Bebe...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/trump-menjanjikan-vaksin-...</td>\n",
       "      <td>Trump Menjanjikan Vaksin Covid-19 pada Akhir T...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>Mei 4, 2020 10:52 am</td>\n",
       "      <td>INIPASTI.COM, Presiden Donald Trump meluncurka...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/partai-gelora-bagi-bantua...</td>\n",
       "      <td>Partai Gelora Bagi Bantuan Sembako ke Warga Te...</td>\n",
       "      <td>Muhammad Seilessy</td>\n",
       "      <td>Mei 3, 2020 5:48 pm</td>\n",
       "      <td>INIPASTI.COM, MAKASSAR, – DPD Partai Gelora In...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/para-pengajarnya-selebrit...</td>\n",
       "      <td>Para Pengajarnya Selebritis, Startup Kelas Onl...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>Mei 5, 2020 9:08 am</td>\n",
       "      <td>INIPASTI.COM, MasterClass, sebuah startup yang...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/tanggap-pandemi-satgas-co...</td>\n",
       "      <td>Tanggap Pandemi, Satgas Covid-19 Unhas Terima ...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>April 23, 2020 5:50 pm</td>\n",
       "      <td>INIPASTI.COM, MAKASSAR – Satuan Tugas Pencegah...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/pasien-eks-covid-19-dikun...</td>\n",
       "      <td>Pasien Eks Covid-19 dikunjungi Tim Gugus Tugas...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>Mei 5, 2020 2:19 pm</td>\n",
       "      <td>INIPASTI.COM, PASANGKAYU– Satu pasien Covid-19...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/terkait-rencana-psbb-di-s...</td>\n",
       "      <td>Terkait Rencana PSBB di Sulbar,Ini Masukan Pem...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>Mei 4, 2020 4:54 am</td>\n",
       "      <td>INIPASTI.COM, PASANGKAYU – Pemprov Sulbar ikut...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/psbb-makassar-diperpanjan...</td>\n",
       "      <td>PSBB Makassar Diperpanjang 14 Hari, Petugas Di...</td>\n",
       "      <td>Muhammad Seilessy</td>\n",
       "      <td>Mei 6, 2020 5:08 pm</td>\n",
       "      <td>INIPASTI.COM, MAKASSAR, — Pembatasan Sosial Be...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/ajiep-surati-pimpinan-dpd...</td>\n",
       "      <td>Ajiep Surati Pimpinan DPD Terkait Postur APBN ...</td>\n",
       "      <td>Muhammad Seilessy</td>\n",
       "      <td>April 10, 2020 2:03 pm</td>\n",
       "      <td>INIPASTI.COM, MAKASSAR, — Anggota Dewan Perwak...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/dampak-corona-penumpang-p...</td>\n",
       "      <td>Dampak Corona, Penumpang Pesawat ke Luar Neger...</td>\n",
       "      <td>Iin Nurfahraeni</td>\n",
       "      <td>April 2, 2020 2:16 pm</td>\n",
       "      <td>INIPASTI.COM, JAKARTA – Badan Pusat Statistik ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/cegah-penyebaran-covid-19...</td>\n",
       "      <td>Cegah Penyebaran Covid-19, WNA Mulai Dilarang ...</td>\n",
       "      <td>Iin Nurfahraeni</td>\n",
       "      <td>April 2, 2020 2:15 pm</td>\n",
       "      <td>INIPASTI.COM, JAKARTA – Penyebaran virus Coron...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/as-dan-inggris-ingatkan-a...</td>\n",
       "      <td>AS dan Inggris Ingatkan, Adanya Peretas Dukung...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>Mei 6, 2020 9:18 am</td>\n",
       "      <td>INIPASTI.COM, LONDON / WASHINGTON – Peretas ya...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/jelang-lebaran-di-tengah-...</td>\n",
       "      <td>Jelang Lebaran di Tengah Pandemi, Muda Mudi To...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>Mei 7, 2020 9:50 am</td>\n",
       "      <td>INIPASTI.COM, MALILI – Melihat pandemi yang su...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/jelang-lebaran-di-tengah-...</td>\n",
       "      <td>Jelang Lebaran di Tengah Pandemi, Muda Mudi To...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>Mei 7, 2020 9:50 am</td>\n",
       "      <td>INIPASTI.COM, MALILI – Melihat pandemi yang su...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/iqbal-suhaeb-psbb-tahap-k...</td>\n",
       "      <td>Iqbal Suhaeb: PSBB Tahap Kedua untuk Bangun So...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>Mei 6, 2020 5:38 pm</td>\n",
       "      <td>INIPASTI.COM, MAKASSAR – Pj Wali Kota Makassar...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/iqbal-suhaeb-psbb-tahap-k...</td>\n",
       "      <td>Iqbal Suhaeb: PSBB Tahap Kedua untuk Bangun So...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>Mei 6, 2020 5:38 pm</td>\n",
       "      <td>INIPASTI.COM, MAKASSAR – Pj Wali Kota Makassar...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/perangi-berita-palsu-fron...</td>\n",
       "      <td>Perangi Berita Palsu: Front Baru dalam Melawan...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>April 14, 2020 8:07 am</td>\n",
       "      <td>INIPASTI.COM, ANALISIS – Jaman yang disesaki d...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/fakta-baru-virus-corona-y...</td>\n",
       "      <td>Fakta Baru Virus Corona yang Mengejutkan</td>\n",
       "      <td>Syakhruddin DN</td>\n",
       "      <td>Februari 8, 2020 11:50 am</td>\n",
       "      <td>INIPASTI.COM, JAKARTA –  Satu pasien, dirawat ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/opini-covid-19-mengubah-l...</td>\n",
       "      <td>[Opini] Covid-19 Mengubah Lanskap Persekolahan...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>Mei 5, 2020 9:28 am</td>\n",
       "      <td>Dr. Naidah Naing, ST., MSi., IAI INIPASTI.COM,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/kenapa-ribut-dengan-satga...</td>\n",
       "      <td>Kenapa Ribut dengan Satgas COVID dan Ramuan HE...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>Mei 4, 2020 11:40 am</td>\n",
       "      <td>Oleh: Haris Rusly Moti INIPASTI.COM, OPINI – S...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/coronavirus-disease-centr...</td>\n",
       "      <td>“Coronavirus Disease Centre” Pemprov Sulsel</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>Mei 2, 2020 8:02 am</td>\n",
       "      <td>Penulis : Muhammad Zaiyani INIPASTI.COM, OPINI...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/ajak-masyarakat-cegah-cor...</td>\n",
       "      <td>Ajak Masyarakat Cegah Corona, Appi Tetap Optim...</td>\n",
       "      <td>Muhammad Seilessy</td>\n",
       "      <td>April 30, 2020 5:30 pm</td>\n",
       "      <td>INIPASTI.COM, MAKASSAR — Bakal calon Wali Kota...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/bagaimana-coronavirus-dap...</td>\n",
       "      <td>Bagaimana Coronavirus Dapat Mengubah Tatanan I...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>Maret 21, 2020 10:31 pm</td>\n",
       "      <td>INIPASTI.COM – Wabah coronavirus COVID-19 tela...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/gratis-di-tengah-pandemi-...</td>\n",
       "      <td>Gratis di Tengah Pandemi, Anda Sekarang Dapat ...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>Mei 2, 2020 7:44 am</td>\n",
       "      <td>INIPASTI.COM, TIPS – Google memotong harga apl...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/para-pengajarnya-selebrit...</td>\n",
       "      <td>Para Pengajarnya Selebritis, Startup Kelas Onl...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>Mei 5, 2020 9:08 am</td>\n",
       "      <td>INIPASTI.COM, MasterClass, sebuah startup yang...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/pasien-eks-covid-19-dikun...</td>\n",
       "      <td>Pasien Eks Covid-19 dikunjungi Tim Gugus Tugas...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>Mei 5, 2020 2:19 pm</td>\n",
       "      <td>INIPASTI.COM, PASANGKAYU– Satu pasien Covid-19...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/terkait-rencana-psbb-di-s...</td>\n",
       "      <td>Terkait Rencana PSBB di Sulbar,Ini Masukan Pem...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>Mei 4, 2020 4:54 am</td>\n",
       "      <td>INIPASTI.COM, PASANGKAYU – Pemprov Sulbar ikut...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/kunjungi-posko-covid-19-w...</td>\n",
       "      <td>Kunjungi Posko Covid 19, Wagub Sulsel Beri Apr...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>Mei 6, 2020 12:13 pm</td>\n",
       "      <td>INIPASTI.COM, MAKASSAR, Wakil Gubernur Sulawes...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/kunjungi-posko-covid-19-w...</td>\n",
       "      <td>Kunjungi Posko Covid 19, Wagub Sulsel Beri Apr...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>Mei 6, 2020 12:13 pm</td>\n",
       "      <td>INIPASTI.COM, MAKASSAR, Wakil Gubernur Sulawes...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/menyakitkan-al-aqsa-ditut...</td>\n",
       "      <td>‘Menyakitkan’: Al-Aqsa Ditutup Selama Ramadan ...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>April 17, 2020 2:23 pm</td>\n",
       "      <td>INIPASTI.COM – Dewan Wakaf Islam Yerusalem mem...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/cegah-meluasnya-wabah-cov...</td>\n",
       "      <td>Cegah Meluasnya Wabah Covid-19, Mudik Lebaran ...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>Maret 25, 2020 9:19 am</td>\n",
       "      <td>INIPASTI.COM, Jakarta –  Mudik lebaran 2020 ak...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/pbb-dunia-harus-mencontoh...</td>\n",
       "      <td>PBB: Dunia harus Mencontoh Korea Selatan dalam...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>Mei 1, 2020 2:10 pm</td>\n",
       "      <td>INIPASTI.COM, Ketua PBB mengatakan bahwasanya ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/hasil-uji-coba-remdesivir...</td>\n",
       "      <td>Ini Obat Covid-19 Paling Berhasil yang Diuji C...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>April 30, 2020 2:11 pm</td>\n",
       "      <td>INIPASTI.COM, Harapan akan terapi obat yang ef...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>Inipasti.com</td>\n",
       "      <td>https://inipasti.com/5-hal-berpenghasilan-yang...</td>\n",
       "      <td>5 Hal Berpenghasilan yang Dapat Anda Pelajari ...</td>\n",
       "      <td>Inipasti</td>\n",
       "      <td>April 21, 2020 7:57 am</td>\n",
       "      <td>INIPASTI.COM, TIPS – Perintah karantina atau k...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            news                                               link  \\\n",
       "0   Inipasti.com  https://inipasti.com/category/kesehatan/corona...   \n",
       "1   Inipasti.com  https://inipasti.com/dalam-bayang-bayang-coron...   \n",
       "2   Inipasti.com  https://inipasti.com/trump-menjanjikan-vaksin-...   \n",
       "3   Inipasti.com  https://inipasti.com/partai-gelora-bagi-bantua...   \n",
       "4   Inipasti.com  https://inipasti.com/para-pengajarnya-selebrit...   \n",
       "5   Inipasti.com  https://inipasti.com/tanggap-pandemi-satgas-co...   \n",
       "6   Inipasti.com  https://inipasti.com/pasien-eks-covid-19-dikun...   \n",
       "7   Inipasti.com  https://inipasti.com/terkait-rencana-psbb-di-s...   \n",
       "8   Inipasti.com  https://inipasti.com/psbb-makassar-diperpanjan...   \n",
       "9   Inipasti.com  https://inipasti.com/ajiep-surati-pimpinan-dpd...   \n",
       "10  Inipasti.com  https://inipasti.com/dampak-corona-penumpang-p...   \n",
       "11  Inipasti.com  https://inipasti.com/cegah-penyebaran-covid-19...   \n",
       "12  Inipasti.com  https://inipasti.com/as-dan-inggris-ingatkan-a...   \n",
       "13  Inipasti.com  https://inipasti.com/jelang-lebaran-di-tengah-...   \n",
       "14  Inipasti.com  https://inipasti.com/jelang-lebaran-di-tengah-...   \n",
       "15  Inipasti.com  https://inipasti.com/iqbal-suhaeb-psbb-tahap-k...   \n",
       "16  Inipasti.com  https://inipasti.com/iqbal-suhaeb-psbb-tahap-k...   \n",
       "17  Inipasti.com  https://inipasti.com/perangi-berita-palsu-fron...   \n",
       "18  Inipasti.com  https://inipasti.com/fakta-baru-virus-corona-y...   \n",
       "19  Inipasti.com  https://inipasti.com/opini-covid-19-mengubah-l...   \n",
       "20  Inipasti.com  https://inipasti.com/kenapa-ribut-dengan-satga...   \n",
       "21  Inipasti.com  https://inipasti.com/coronavirus-disease-centr...   \n",
       "22  Inipasti.com  https://inipasti.com/ajak-masyarakat-cegah-cor...   \n",
       "23  Inipasti.com  https://inipasti.com/bagaimana-coronavirus-dap...   \n",
       "24  Inipasti.com  https://inipasti.com/gratis-di-tengah-pandemi-...   \n",
       "25  Inipasti.com  https://inipasti.com/para-pengajarnya-selebrit...   \n",
       "26  Inipasti.com  https://inipasti.com/pasien-eks-covid-19-dikun...   \n",
       "27  Inipasti.com  https://inipasti.com/terkait-rencana-psbb-di-s...   \n",
       "28  Inipasti.com  https://inipasti.com/kunjungi-posko-covid-19-w...   \n",
       "29  Inipasti.com  https://inipasti.com/kunjungi-posko-covid-19-w...   \n",
       "30  Inipasti.com  https://inipasti.com/menyakitkan-al-aqsa-ditut...   \n",
       "31  Inipasti.com  https://inipasti.com/cegah-meluasnya-wabah-cov...   \n",
       "32  Inipasti.com  https://inipasti.com/pbb-dunia-harus-mencontoh...   \n",
       "33  Inipasti.com  https://inipasti.com/hasil-uji-coba-remdesivir...   \n",
       "34  Inipasti.com  https://inipasti.com/5-hal-berpenghasilan-yang...   \n",
       "\n",
       "                                                title             author  \\\n",
       "0                                         Coronavirus               None   \n",
       "1   Dalam Bayang-bayang Coronavirus, Umat Islam di...           Inipasti   \n",
       "2   Trump Menjanjikan Vaksin Covid-19 pada Akhir T...           Inipasti   \n",
       "3   Partai Gelora Bagi Bantuan Sembako ke Warga Te...  Muhammad Seilessy   \n",
       "4   Para Pengajarnya Selebritis, Startup Kelas Onl...           Inipasti   \n",
       "5   Tanggap Pandemi, Satgas Covid-19 Unhas Terima ...           Inipasti   \n",
       "6   Pasien Eks Covid-19 dikunjungi Tim Gugus Tugas...           Inipasti   \n",
       "7   Terkait Rencana PSBB di Sulbar,Ini Masukan Pem...           Inipasti   \n",
       "8   PSBB Makassar Diperpanjang 14 Hari, Petugas Di...  Muhammad Seilessy   \n",
       "9   Ajiep Surati Pimpinan DPD Terkait Postur APBN ...  Muhammad Seilessy   \n",
       "10  Dampak Corona, Penumpang Pesawat ke Luar Neger...    Iin Nurfahraeni   \n",
       "11  Cegah Penyebaran Covid-19, WNA Mulai Dilarang ...    Iin Nurfahraeni   \n",
       "12  AS dan Inggris Ingatkan, Adanya Peretas Dukung...           Inipasti   \n",
       "13  Jelang Lebaran di Tengah Pandemi, Muda Mudi To...           Inipasti   \n",
       "14  Jelang Lebaran di Tengah Pandemi, Muda Mudi To...           Inipasti   \n",
       "15  Iqbal Suhaeb: PSBB Tahap Kedua untuk Bangun So...           Inipasti   \n",
       "16  Iqbal Suhaeb: PSBB Tahap Kedua untuk Bangun So...           Inipasti   \n",
       "17  Perangi Berita Palsu: Front Baru dalam Melawan...           Inipasti   \n",
       "18           Fakta Baru Virus Corona yang Mengejutkan     Syakhruddin DN   \n",
       "19  [Opini] Covid-19 Mengubah Lanskap Persekolahan...           Inipasti   \n",
       "20  Kenapa Ribut dengan Satgas COVID dan Ramuan HE...           Inipasti   \n",
       "21        “Coronavirus Disease Centre” Pemprov Sulsel           Inipasti   \n",
       "22  Ajak Masyarakat Cegah Corona, Appi Tetap Optim...  Muhammad Seilessy   \n",
       "23  Bagaimana Coronavirus Dapat Mengubah Tatanan I...           Inipasti   \n",
       "24  Gratis di Tengah Pandemi, Anda Sekarang Dapat ...           Inipasti   \n",
       "25  Para Pengajarnya Selebritis, Startup Kelas Onl...           Inipasti   \n",
       "26  Pasien Eks Covid-19 dikunjungi Tim Gugus Tugas...           Inipasti   \n",
       "27  Terkait Rencana PSBB di Sulbar,Ini Masukan Pem...           Inipasti   \n",
       "28  Kunjungi Posko Covid 19, Wagub Sulsel Beri Apr...           Inipasti   \n",
       "29  Kunjungi Posko Covid 19, Wagub Sulsel Beri Apr...           Inipasti   \n",
       "30  ‘Menyakitkan’: Al-Aqsa Ditutup Selama Ramadan ...           Inipasti   \n",
       "31  Cegah Meluasnya Wabah Covid-19, Mudik Lebaran ...           Inipasti   \n",
       "32  PBB: Dunia harus Mencontoh Korea Selatan dalam...           Inipasti   \n",
       "33  Ini Obat Covid-19 Paling Berhasil yang Diuji C...           Inipasti   \n",
       "34  5 Hal Berpenghasilan yang Dapat Anda Pelajari ...           Inipasti   \n",
       "\n",
       "                    date_time  \\\n",
       "0         Mei 1, 2020 2:10 pm   \n",
       "1     April 21, 2020 10:37 am   \n",
       "2        Mei 4, 2020 10:52 am   \n",
       "3         Mei 3, 2020 5:48 pm   \n",
       "4         Mei 5, 2020 9:08 am   \n",
       "5      April 23, 2020 5:50 pm   \n",
       "6         Mei 5, 2020 2:19 pm   \n",
       "7         Mei 4, 2020 4:54 am   \n",
       "8         Mei 6, 2020 5:08 pm   \n",
       "9      April 10, 2020 2:03 pm   \n",
       "10      April 2, 2020 2:16 pm   \n",
       "11      April 2, 2020 2:15 pm   \n",
       "12        Mei 6, 2020 9:18 am   \n",
       "13        Mei 7, 2020 9:50 am   \n",
       "14        Mei 7, 2020 9:50 am   \n",
       "15        Mei 6, 2020 5:38 pm   \n",
       "16        Mei 6, 2020 5:38 pm   \n",
       "17     April 14, 2020 8:07 am   \n",
       "18  Februari 8, 2020 11:50 am   \n",
       "19        Mei 5, 2020 9:28 am   \n",
       "20       Mei 4, 2020 11:40 am   \n",
       "21        Mei 2, 2020 8:02 am   \n",
       "22     April 30, 2020 5:30 pm   \n",
       "23    Maret 21, 2020 10:31 pm   \n",
       "24        Mei 2, 2020 7:44 am   \n",
       "25        Mei 5, 2020 9:08 am   \n",
       "26        Mei 5, 2020 2:19 pm   \n",
       "27        Mei 4, 2020 4:54 am   \n",
       "28       Mei 6, 2020 12:13 pm   \n",
       "29       Mei 6, 2020 12:13 pm   \n",
       "30     April 17, 2020 2:23 pm   \n",
       "31     Maret 25, 2020 9:19 am   \n",
       "32        Mei 1, 2020 2:10 pm   \n",
       "33     April 30, 2020 2:11 pm   \n",
       "34     April 21, 2020 7:57 am   \n",
       "\n",
       "                                             paragraf  \n",
       "0                                             N o n e  \n",
       "1   INIPASTI.COM, ALGIERS / CAIRO / JAKARTA – Bebe...  \n",
       "2   INIPASTI.COM, Presiden Donald Trump meluncurka...  \n",
       "3   INIPASTI.COM, MAKASSAR, – DPD Partai Gelora In...  \n",
       "4   INIPASTI.COM, MasterClass, sebuah startup yang...  \n",
       "5   INIPASTI.COM, MAKASSAR – Satuan Tugas Pencegah...  \n",
       "6   INIPASTI.COM, PASANGKAYU– Satu pasien Covid-19...  \n",
       "7   INIPASTI.COM, PASANGKAYU – Pemprov Sulbar ikut...  \n",
       "8   INIPASTI.COM, MAKASSAR, — Pembatasan Sosial Be...  \n",
       "9   INIPASTI.COM, MAKASSAR, — Anggota Dewan Perwak...  \n",
       "10  INIPASTI.COM, JAKARTA – Badan Pusat Statistik ...  \n",
       "11  INIPASTI.COM, JAKARTA – Penyebaran virus Coron...  \n",
       "12  INIPASTI.COM, LONDON / WASHINGTON – Peretas ya...  \n",
       "13  INIPASTI.COM, MALILI – Melihat pandemi yang su...  \n",
       "14  INIPASTI.COM, MALILI – Melihat pandemi yang su...  \n",
       "15  INIPASTI.COM, MAKASSAR – Pj Wali Kota Makassar...  \n",
       "16  INIPASTI.COM, MAKASSAR – Pj Wali Kota Makassar...  \n",
       "17  INIPASTI.COM, ANALISIS – Jaman yang disesaki d...  \n",
       "18  INIPASTI.COM, JAKARTA –  Satu pasien, dirawat ...  \n",
       "19  Dr. Naidah Naing, ST., MSi., IAI INIPASTI.COM,...  \n",
       "20  Oleh: Haris Rusly Moti INIPASTI.COM, OPINI – S...  \n",
       "21  Penulis : Muhammad Zaiyani INIPASTI.COM, OPINI...  \n",
       "22  INIPASTI.COM, MAKASSAR — Bakal calon Wali Kota...  \n",
       "23  INIPASTI.COM – Wabah coronavirus COVID-19 tela...  \n",
       "24  INIPASTI.COM, TIPS – Google memotong harga apl...  \n",
       "25  INIPASTI.COM, MasterClass, sebuah startup yang...  \n",
       "26  INIPASTI.COM, PASANGKAYU– Satu pasien Covid-19...  \n",
       "27  INIPASTI.COM, PASANGKAYU – Pemprov Sulbar ikut...  \n",
       "28  INIPASTI.COM, MAKASSAR, Wakil Gubernur Sulawes...  \n",
       "29  INIPASTI.COM, MAKASSAR, Wakil Gubernur Sulawes...  \n",
       "30  INIPASTI.COM – Dewan Wakaf Islam Yerusalem mem...  \n",
       "31  INIPASTI.COM, Jakarta –  Mudik lebaran 2020 ak...  \n",
       "32  INIPASTI.COM, Ketua PBB mengatakan bahwasanya ...  \n",
       "33  INIPASTI.COM, Harapan akan terapi obat yang ef...  \n",
       "34  INIPASTI.COM, TIPS – Perintah karantina atau k...  "
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# r = get_news(main_url, \"news_2\")\n",
    "# r\n",
    "\n",
    "inipasti_news=get_news(main_url, file_name = 'inipasti_satu')\n",
    "inipasti_news"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}