Newer
Older
import os
import datetime
from bs4 import BeautifulSoup
from urllib.parse import unquote
from Attrap import Attrap
class Attrap_pref09(Attrap):
# Config
__HOST = 'https://www.ariege.gouv.fr'
__RAA_PAGE = f'{__HOST}/Publications/Recueil-des-actes-administratifs/Recueils-des-Actes-Administratifs-de-l-Ariege-a-partir-du-28-avril-2015'
__USER_AGENT = 'Mozilla/5.0 (Windows NT 10.0; rv:109.0) Gecko/20100101 Firefox/115.0'
full_name = 'Préfecture de l\'Ariège'
short_code = 'pref09'
def __init__(self, data_dir):
super().__init__(data_dir, self.__USER_AGENT)
self.set_sleep_time(30)
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
def get_raa(self, keywords):
pages_to_parse = []
# Les RAA de l'Ariège sont éparpillés sur des sous-pages par mois.
# Donc on parse la page principale à la recherche des sous-pages.
sub_pages = self.get_sub_pages_with_pager(
self.__RAA_PAGE,
'div.fr-card__body div.fr-card__content h2.fr-card__title a.fr-card__link',
'ul.fr-pagination__list li a.fr-pagination__link.fr-pagination__link--next',
'div.fr-card__body div.fr-card__content div.fr-card__end p.fr-card__detail',
self.__HOST
)[::-1]
# On filtre par date les sous-pages pour limiter les requêtes
for sub_page in sub_pages:
guessed_date = datetime.datetime.strptime(sub_page['details'].replace('Publié le ', '').strip(), '%d/%m/%Y')
guessed_date.replace(day=1)
if guessed_date >= self.not_before:
pages_to_parse.append(sub_page['url'])
# On parse les pages contenant des RAA
elements = []
for page in pages_to_parse:
page_content = self.get_page(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')
# On récupère chaque balise a
for a in soup.select('div.fr-downloads-group.fr-downloads-group--bordered ul li a'):
if a.get('href') and 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