import datetime
import logging
import re
import urllib.parse

from bs4 import BeautifulSoup

from Attrap import Attrap

logger = logging.getLogger(__name__)


class Attrap_pref92(Attrap):

    # Config
    __USER_AGENT = 'Mozilla/5.0 (X11; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/115.0'
    __HOST = 'https://www.hauts-de-seine.gouv.fr'
    __FIRST_YEAR = 2016
    __RAA_YEAR_PAGE_TEMPLATE = (
        '/Publications/Annonces-avis/'
        'Le-Recueil-des-actes-administratifs/RAA-{year}/'
    )
    __RAA_PAGER_SELECTOR = 'a.fr-pagination__link.fr-pagination__link--next'
    __RAA_PDF_SELECTOR = '#main .fr-card__content a.fr-card__link.menu-item-link'
    __RAA_PUB_DATE_RE = re.compile(r'(?P<d>\d{2})\/(?P<m>\d{2})\/(?P<y>\d{4})')
    full_name = 'Préfecture des Hauts-de-Seine'
    short_code = 'pref92'

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

    def get_raa(self, keywords):
        # On récupère les pages des années. Pas besoin de fetch la page
        # initiale, l'URL est prévisible pour chaque année.
        year_pages = [
            self.__HOST + self.__RAA_YEAR_PAGE_TEMPLATE.format(year=year)
            for year in self._get_year_range(min_year=self.not_before.year)
        ]
        # On récupère tous les RAA en suivant la navigation.
        elements = self.get_raa_with_pager(
            year_pages,
            self.__RAA_PAGER_SELECTOR,
            self.__HOST
        )

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

    def get_raa_elements(self, page_content):
        soup = BeautifulSoup(page_content, 'html.parser')

        elements = []
        for a in soup.select(self.__RAA_PDF_SELECTOR):
            # On ignore les RAA dont on a pu voir que la date de publication
            # est avant notre limite.
            link_title = a['title']
            pub_date = self._get_pub_date_from_link_title(link_title)
            if not pub_date:
                logger.warning(f'Date non trouvée dans l\'attribut title : {link_title}')
                continue
            if pub_date < self.not_before:
                continue

            if (url := a.get('href')) and url.endswith('.pdf'):
                if url.startswith('/'):
                    url = self.__HOST + urllib.parse.quote(url)
                name = a.get_text().strip()
                raa = Attrap.RAA(url, pub_date, name)
                elements.append(raa)

        return elements

    def _get_pub_date_from_link_title(self, title):
        """C'est sympa il y a la date de publication du RAA dans le title du lien."""
        if match := self.__RAA_PUB_DATE_RE.search(title):
            parts = match.groupdict()
            return datetime.datetime(
                int(parts['y']), int(parts['m']), int(parts['d'])
            )
        return None

    def _get_year_range(self, min_year=0):
        """Renvoie une liste d'années avec des RAA."""
        return list(
            range(
                max(self.__FIRST_YEAR, min_year),
                datetime.datetime.now().year + 1
            )
        )