import os
import datetime

from bs4 import BeautifulSoup
from urllib.parse import unquote

from Attrap import Attrap


class Attrap_pref35(Attrap):

    # Config
    __HOST = 'https://www.ille-et-vilaine.gouv.fr'
    __RAA_PAGE = [
        f'{__HOST}/Publications/Recueil-des-actes-administratifs/Recueil-des-actes-administratifs-2024',
        f'{__HOST}/Publications/Recueil-des-actes-administratifs/Archives-des-recueils-des-actes-administratifs/Recueil-des-actes-administratifs-2023',
        f'{__HOST}/Publications/Recueil-des-actes-administratifs/Archives-des-recueils-des-actes-administratifs/Recueil-des-actes-administratifs-2022',
        f'{__HOST}/Publications/Recueil-des-actes-administratifs/Archives-des-recueils-des-actes-administratifs/Recueil-des-actes-administratifs-2021',
        f'{__HOST}/Publications/Recueil-des-actes-administratifs/Archives-des-recueils-des-actes-administratifs/Recueil-des-actes-administratifs-2020',
        f'{__HOST}/Publications/Recueil-des-actes-administratifs/Archives-des-recueils-des-actes-administratifs/Recueil-des-actes-administratifs-2019'
    ]
    __USER_AGENT = 'Mozilla/5.0 (Windows NT 10.0; rv:109.0) Gecko/20100101 Firefox/115.0'
    full_name = 'Préfecture d\'Ille-et-Vilaine'
    short_code = 'pref35'

    def __init__(self, data_dir):
        super().__init__(data_dir, self.__USER_AGENT)
        self.enable_tor(10)

    def get_raa(self, keywords):
        elements = []
        for raa_page in self.__RAA_PAGE:
            page_content = self.get_page(raa_page, 'get').content
            for element in self.get_raa_elements(page_content):
                elements.append(element)

        self.parse_raa(elements, keywords)
        self.mailer()

    def get_raa_elements(self, page_content):
        elements = []
        # On charge le parser
        soup = BeautifulSoup(page_content, 'html.parser')

        # Pour chaque balise a, on regarde si c'est un PDF, et si oui on le
        # parse
        for a in soup.find_all('a', href=True, class_='fr-link--download'):
            if a['href'].endswith('.pdf'):
                if a['href'].startswith('/'):
                    url = f"{self.__HOST}{a['href']}"
                else:
                    url = a['href']

                url = unquote(url)
                name = a.find('span').previous_sibling.replace('Télécharger ', '').strip()
                date = datetime.datetime.strptime(a.find('span').get_text().split(' - ')[-1].strip(), '%d/%m/%Y')

                raa = Attrap.RAA(url, date, name)
                elements.append(raa)
        return elements