calendar-bot/scraping.py

from typing import List
from bs4 import BeautifulSoup
import requests
from operator import add
from functools import reduce
import datetime
from event import Event


def get_soup() -> BeautifulSoup:
    r = requests.get("http://www.pvv.ntnu.no/hendelser/")
    soup = BeautifulSoup(r.text, "html.parser")
    return soup 


def process_soup(soup: BeautifulSoup) -> List[Event]:
    zips = []
    events = soup.find_all("ul", "events")

    for event in events:
        times, locations, organizers = zip(*(list(
            map(lambda x: map(lambda y: y.find("strong").text, x), 
                filter(lambda x: x != [], 
                       map(lambda x: x.find_all("li"), 
                           event.find_all("li")))))))
        times = list(map(lambda x: datetime.datetime.strptime(x, "%A %d. %b %H.%M").replace(year=datetime.datetime.now().year), times))
        names = list(map(lambda x: x[0].text, 
                     filter(lambda x: x != [], 
                            map(lambda x: x.find_all("a"), 
                                event.find_all("li")))))
        descriptions = list(map(lambda x: x.text, 
                     filter(lambda x: x is not None, 
                            map(lambda x: x.find("p"), 
                                event.find_all("li")))))
        ids = []
        for a in event.find_all("a", href=True):
            if a['href'][:10] == '/hendelser':
                ids.append(int(a['href'][a['href'].find("=")+1:]))

        zips.append(list(zip(ids, times, names, organizers, locations, descriptions)))

    events = reduce(add, zips)
    events = list(map(lambda x: Event(*x), events))
    return events


def get_events_today(events: List[Event]):
    return list(filter(lambda e: e.start.date() == datetime.datetime.today().date(), events))


if __name__ == "__main__":
    print(get_events_today(process_soup(get_soup())))