pref13: détecte l'URL de l'année voulue automatiquement

Attrap_pref13.py

@@ -12,12 +12,8 @@ class Attrap_pref13(Attrap):
     # Config
     __HOST = 'https://www.bouches-du-rhone.gouv.fr'
     __RAA_PAGE = [
-        f'{__HOST}/Publications/RAA-et-Archives/RAA-2024',
-        f'{__HOST}/Publications/RAA-et-Archives/RAA-2023',
-        f'{__HOST}/Publications/RAA-et-Archives/Archives-RAA-des-Bouches-du-Rhone/RAA-2022',
-        f'{__HOST}/Publications/RAA-et-Archives/Archives-RAA-des-Bouches-du-Rhone/RAA-2021',
-        f'{__HOST}/Publications/RAA-et-Archives/Archives-RAA-des-Bouches-du-Rhone/RAA-2020',
-        f'{__HOST}/Publications/RAA-et-Archives/Archives-RAA-des-Bouches-du-Rhone/RAA-2019'
+        f'{__HOST}/Publications/RAA-et-Archives',
+        f'{__HOST}/Publications/RAA-et-Archives/Archives-RAA-des-Bouches-du-Rhone'
     ]
     __USER_AGENT = 'Mozilla/5.0 (Windows NT 10.0; rv:109.0) Gecko/20100101 Firefox/115.0'
     full_name = 'Préfecture des Bouches-du-Rhône'
@@ -28,9 +24,23 @@ class Attrap_pref13(Attrap):
         self.set_sleep_time(30)
 
     def get_raa(self, keywords):
-        elements = []
+        pages_to_parse = []
         for raa_page in self.__RAA_PAGE:
             page_content = self.get_page(raa_page, 'get').content
+            year_pages = self.get_sub_pages(
+                page_content,
+                '.fr-card.fr-card--sm.fr-card--grey.fr-enlarge-link div.fr-card__body div.fr-card__content h2.fr-card__title a',
+                self.__HOST,
+                False
+            )
+            for year_page in year_pages:
+                year = Attrap.guess_date(year_page['name'], '.*([0-9]{4})').year
+                if year >= self.not_before.year and year < 9999:
+                    pages_to_parse.append(year_page['url'])
+
+        elements = []
+        for page_to_parse in pages_to_parse:
+            page_content = self.get_page(page_to_parse, 'get').content
             for element in self.get_raa_elements(page_content):
                 elements.append(element)