needsUpdate for DIVI

.gitignore

@@ -1,5 +1,6 @@
 src/VAERS/
 src/intensivstationen/zeitreihe-tagesdaten.csv
+src/intensivstationen/geckodriver.log
 .ipynb_checkpoints/
 .history/
 src/config/

src/HowBadIsMyBatch.ipynb

@@ -27,9 +27,9 @@
     "from dateutil.parser import parse\n",
     "\n",
     "def needsUpdate():\n",
-    "    return _getLastUpdatedHowbadismybatch() < _getLastUpdatedOriginal()\n",
+    "    return _getLastUpdatedBatchCodeTable() < _getLastUpdatedOriginal()\n",
     "    \n",
-    "def _getLastUpdatedHowbadismybatch():\n",
+    "def _getLastUpdatedBatchCodeTable():\n",
     "    return _getLastUpdated(\n",
     "        url = \"https://knollfrank.github.io/HowBadIsMyBatch/batchCodeTable.html\",\n",
     "        getDateStr = lambda soup: soup.find(id = \"last_updated\").text)\n",

src/intensivstationen/Intensivstationen.ipynb

@@ -15,6 +15,54 @@
     "pd.set_option('display.max_columns', None)"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "79de4057",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from bs4 import BeautifulSoup\n",
+    "import requests\n",
+    "from datetime import datetime\n",
+    "from time import sleep\n",
+    "from selenium import webdriver\n",
+    "\n",
+    "def needsUpdate():\n",
+    "    lastUpdatedIntensivstationen = _getLastUpdatedIntensivstationen()\n",
+    "    print('lastUpdatedIntensivstationen:', lastUpdatedIntensivstationen)\n",
+    "    lastUpdatedOriginal = _getLastUpdatedOriginal()\n",
+    "    print('lastUpdatedOriginal:', lastUpdatedOriginal)\n",
+    "    return lastUpdatedIntensivstationen < lastUpdatedOriginal\n",
+    "    \n",
+    "def _getLastUpdatedIntensivstationen():\n",
+    "    htmlContent = requests.get(\"https://knollfrank.github.io/HowBadIsMyBatch/intensivstationen.html\").text\n",
+    "    soup = BeautifulSoup(htmlContent, \"lxml\")\n",
+    "    dateStr = soup.find(id = \"Datenstand\").text\n",
+    "    return datetime.strptime(dateStr, \"%d.%m.%Y, %H:%M Uhr\")\n",
+    "\n",
+    "def _getLastUpdatedOriginal():\n",
+    "    html = _getOriginalHtml()\n",
+    "    dataFrame = _asDataFrame(html)\n",
+    "    return dataFrame.loc['Landkreis-Daten', 'Letzte Änderung'].to_pydatetime()\n",
+    "\n",
+    "def _getOriginalHtml():\n",
+    "    driver = webdriver.Firefox()\n",
+    "    driver.get('https://www.intensivregister.de/#/aktuelle-lage/downloads')\n",
+    "    sleep(5)\n",
+    "    innerHTML = driver.execute_script(\"return document.body.innerHTML\")\n",
+    "    driver.quit()\n",
+    "    return innerHTML\n",
+    "\n",
+    "def _asDataFrame(html):\n",
+    "    dataFrame = pd.read_html(html, parse_dates = ['Letzte Änderung'])[0]\n",
+    "    dataFrame.set_index('Name', inplace = True)\n",
+    "    dataFrame['Letzte Änderung'] = pd.to_datetime(dataFrame['Letzte Änderung'], format = \"%d.%m.%Y %H:%M Uhr\")\n",
+    "    return dataFrame\n",
+    "\n",
+    "print('needsUpdate: ', needsUpdate())"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,