Merge branch 'main' into pages

src/HowBadIsMyBatch.ipynb

@@ -24,36 +24,58 @@
     "from bs4 import BeautifulSoup\n",
     "import requests\n",
     "import re\n",
-    "from dateutil.parser import parse\n",
+    "from datetime import datetime\n",
     "\n",
-    "def needsUpdate():\n",
+    "class DateProvider:\n",
-    "    lastUpdated = _getLastUpdated()\n",
-    "    print('          lastUpdated:', lastUpdated)\n",
     "    \n",
-    "    lastUpdatedDataSource = _getLastUpdatedDataSource()\n",
+    "    DATE_FORMAT = \"%B %d, %Y\"\n",
-    "    print('lastUpdatedDataSource:', lastUpdatedDataSource)\n",
     "\n",
-    "    return lastUpdated < lastUpdatedDataSource\n",
+    "    def __init__(self):\n",
+    "        self.lastUpdated = None\n",
+    "        self.lastUpdatedDataSource = None\n",
     "\n",
-    "def _getLastUpdated():\n",
+    "    def needsUpdate(self):\n",
-    "    return __getLastUpdated(\n",
+    "        return self.getLastUpdated() < self.getLastUpdatedDataSource()\n",
+    "        \n",
+    "    def getLastUpdated(self):\n",
+    "        if self.lastUpdated is None:\n",
+    "            self.lastUpdated = self.__getLastUpdated(\n",
     "                url = \"https://knollfrank.github.io/HowBadIsMyBatch/batchCodeTable.html\",\n",
     "                getDateStr = lambda soup: soup.find(id = \"last_updated\").text)\n",
     "        \n",
-    "def _getLastUpdatedDataSource():\n",
+    "        return self.lastUpdated\n",
+    "\n",
+    "    def getLastUpdatedDataSource(self):\n",
+    "        if self.lastUpdatedDataSource is None:\n",
     "            def getDateStr(soup):\n",
     "                lastUpdated = soup.find(string = re.compile(\"Last updated\"))\n",
     "                return re.search('Last updated: (.+).', lastUpdated).group(1)\n",
     "\n",
-    "    return __getLastUpdated(url = \"https://vaers.hhs.gov/data/datasets.html\", getDateStr = getDateStr)\n",
+    "            self.lastUpdatedDataSource = self.__getLastUpdated(\n",
+    "                url = \"https://vaers.hhs.gov/data/datasets.html\",\n",
+    "                getDateStr = getDateStr)\n",
     "\n",
-    "def __getLastUpdated(url, getDateStr):\n",
+    "        return self.lastUpdatedDataSource\n",
+    "\n",
+    "    def __getLastUpdated(self, url, getDateStr):\n",
     "        htmlContent = requests.get(url).text\n",
     "        soup = BeautifulSoup(htmlContent, \"lxml\")\n",
     "        dateStr = getDateStr(soup)\n",
-    "    return parse(dateStr).date()\n",
+    "        return datetime.strptime(dateStr, DateProvider.DATE_FORMAT)"
-    "\n",
+   ]
-    "print('needsUpdate:', needsUpdate())"
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "ffad1c04",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "dateProvider = DateProvider()\n",
+    "print('          lastUpdated:', dateProvider.getLastUpdated())\n",
+    "print('lastUpdatedDataSource:', dateProvider.getLastUpdatedDataSource())        \n",
+    "needsUpdate = dateProvider.needsUpdate()\n",
+    "print('needsUpdate:', needsUpdate)"
    ]
   },
   {
@@ -439,6 +461,23 @@
     "                    'lxml'))\n"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "f02dddfe",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def saveLastUpdatedBatchCodeTable(lastUpdated):\n",
+    "    def setLastUpdated(soup):\n",
+    "        soup.find(id = \"last_updated\").string.replace_with(lastUpdated.strftime(DateProvider.DATE_FORMAT))\n",
+    "        return soup\n",
+    "\n",
+    "    HtmlTransformerUtil().applySoupTransformerToFile(\n",
+    "        file = \"../docs/batchCodeTable.html\",\n",
+    "        soupTransformer = setLastUpdated)"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -933,16 +972,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "countries = sorted(internationalVaersCovid19['COUNTRY'].unique())"
+    "countries = sorted(internationalVaersCovid19['COUNTRY'].unique())\n",
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "1c518028",
-   "metadata": {},
-   "outputs": [],
-   "source": [
     "countryOptions = ['<option value=\"Global\" selected>Global</option>']  + getCountryOptions(countries)"
    ]
   },
@@ -956,6 +986,16 @@
     "saveCountryOptions(countryOptions)"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "9c7485b5",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "saveLastUpdatedBatchCodeTable(dateProvider.getLastUpdatedDataSource())"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,