Initial commit

custom_components/home_assistant_covid19_augsburg/__init__.py

@@ -0,0 +1,80 @@
+"""The corona_hessen component."""
+
+import asyncio
+import logging
+import re
+from datetime import timedelta
+
+from homeassistant.config_entries import ConfigEntry
+from homeassistant.core import HomeAssistant
+from homeassistant.helpers.update_coordinator import DataUpdateCoordinator
+
+from .const import DOMAIN
+from .crawler import CovidCrawler, IncidenceData
+
+_LOGGER = logging.getLogger(__name__)
+
+__version__ = "0.1.0"
+
+PLATFORMS = ["sensor"]
+
+HYPHEN_PATTERN = re.compile(r"- (.)")
+
+
+async def async_setup(hass: HomeAssistant, config: dict):
+    """Set up the Coronavirus Augsburg component."""
+    # Make sure coordinator is initialized.
+    coordinator = await get_coordinator(hass)
+
+    async def handle_refresh(call):
+        _LOGGER.info("Refreshing Coronavirus Augsburg data...")
+        await coordinator.async_refresh()
+
+    hass.services.async_register(DOMAIN, "refresh", handle_refresh)
+
+    return True
+
+
+async def async_setup_entry(hass: HomeAssistant, entry: ConfigEntry):
+    """Set up Coronavirus Augsburg from a config entry."""
+
+    for component in PLATFORMS:
+        hass.async_create_task(
+            hass.config_entries.async_forward_entry_setup(entry, component)
+        )
+
+    return True
+
+
+async def async_unload_entry(hass: HomeAssistant, entry: ConfigEntry):
+    """Unload a config entry."""
+    unload_ok = all(
+        await asyncio.gather(
+            *[
+                hass.config_entries.async_forward_entry_unload(entry, cmp)
+                for cmp in PLATFORMS
+            ]
+        )
+    )
+
+    return unload_ok
+
+
+async def get_coordinator(hass):
+    """Get the data update coordinator."""
+    if DOMAIN in hass.data:
+        return hass.data[DOMAIN]
+
+    async def async_get_data() -> IncidenceData:
+        crawler = CovidCrawler()
+        return crawler.crawl()
+
+    hass.data[DOMAIN] = DataUpdateCoordinator(
+        hass,
+        logging.getLogger(__name__),
+        name=DOMAIN,
+        update_method=async_get_data,
+        update_interval=timedelta(hours=6),
+    )
+    await hass.data[DOMAIN].async_refresh()
+    return hass.data[DOMAIN]

custom_components/home_assistant_covid19_augsburg/const.py

@@ -0,0 +1 @@
+DOMAIN = "covid19_augsburg"

custom_components/home_assistant_covid19_augsburg/crawler.py

@@ -0,0 +1,94 @@
+import datetime
+import locale
+import logging
+import re
+from abc import ABC, abstractmethod
+from dataclasses import dataclass
+
+import requests
+from bs4 import BeautifulSoup
+
+_log = logging.getLogger(__name__)
+
+
+@dataclass
+class IncidenceData:
+    location: str
+    date: datetime.date
+    incidence: float
+    total_cases: int = 0
+    num_infected: int = 0
+    num_recovered: int = 0
+    num_dead: int = 0
+
+
+class CovidCrawlerBase(ABC):
+    @abstractmethod
+    def crawl(self) -> IncidenceData:
+        pass
+
+
+class CovidCrawler(CovidCrawlerBase):
+    def __init__(self) -> None:
+        self.url = (
+            "https://www.augsburg.de/umwelt-soziales/gesundheit/coronavirus/fallzahlen"
+        )
+
+    def crawl(self) -> IncidenceData:
+        """
+        Fetch COVID-19 infection data from the target website.
+        """
+
+        _log.info("Fetching COVID-19 data update")
+
+        locale.setlocale(locale.LC_ALL, "de_DE.utf8")
+
+        result = requests.get(self.url)
+        if not result.ok:
+            result.raise_for_status()
+
+        soup = BeautifulSoup(result.text, features="html.parser")
+
+        match = soup.find(class_="frame--type-textpic")
+        text = match.p.text
+        _log.debug(f"Infection data text: {text}")
+
+        matches = re.search(r"(\d+,\d+) Neuinfektion", text)
+        if not matches:
+            raise ValueError("Could not extract incidence from scraped web page")
+
+        incidence = locale.atof(matches.group(1))
+        _log.debug(f"Parsed incidence: {incidence}")
+
+        text = match.h2.text
+        matches = re.search(r"\((\d+\. \w+)\)", text)
+        if not matches:
+            raise ValueError("Could not extract date from scraped web page")
+
+        date = datetime.datetime.strptime(matches.group(1), "%d. %B")
+        date = date.replace(year=datetime.datetime.now().year).date()
+        _log.debug(f"Parsed date: {date}")
+
+        match = match.find_next_sibling(class_="frame--type-textpic")
+        text = match.text
+        _log.debug(f"Infection counts text: {text}")
+
+        regexes = [
+            r"Insgesamt: (?P<total_cases>[0-9.]+)",
+            r"genesen: (?P<num_recovered>[0-9.]+)",
+            r"infiziert: (?P<num_infected>[0-9.]+)",
+            r"verstorben: (?P<num_dead>[0-9.]+)",
+        ]
+        cases = {}
+        for r in regexes:
+            matches = re.search(r, text)
+            if not matches:
+                continue
+            cases.update(
+                {k: int(v.replace(".", "")) for k, v in matches.groupdict().items()}
+            )
+
+        result = IncidenceData("Augsburg", incidence, date, **cases)
+        _log.debug(f"Result data: {result}")
+
+        return result

custom_components/home_assistant_covid19_augsburg/main.py

@@ -0,0 +1,11 @@
+from home_assistant_covid19_augsburg.crawler import CovidCrawler
+
+
+def main():
+    crawler = CovidCrawler()
+    result = crawler.crawl()
+    print(result)
+
+
+if __name__ == "__main__":
+    main()

custom_components/home_assistant_covid19_augsburg/manifest.json

@@ -0,0 +1,10 @@
+{
+    "domain": "covid19_augsburg",
+    "name": "COVID-19 Augsburg",
+    "version": "0.1.0",
+    "config_flow": true,
+    "documentation": "https://github.com/AdrianoKF/home-assistant-covid19-augsburg",
+    "requirements": ["beautifulsoup4==4.8.2", "requests==2.25.1"],
+    "dependencies": [],
+    "codeowners": ["@AdrianoKF"]
+  }

custom_components/home_assistant_covid19_augsburg/services.yaml

@@ -0,0 +1,2 @@
+refresh:
+  description: Refreshes data from web