pref69: détecte l'URL de l'année voulue automatiquement

Attrap_pref69.py

@@ -11,14 +11,7 @@ class Attrap_pref69(Attrap):
 
     # Config
     hostname = 'https://www.rhone.gouv.fr'
-    raa_page = {
-        '2024': f'{hostname}/Publications/Recueil-des-actes-administratifs-du-Rhone-RAA/Recueils-de-2024',
-        '2023': f'{hostname}/Publications/Recueil-des-actes-administratifs-du-Rhone-RAA/Recueils-de-2023',
-        '2022': f'{hostname}/Publications/Recueil-des-actes-administratifs-du-Rhone-RAA/Recueils-de-2022',
-        '2021': f'{hostname}/Publications/Recueil-des-actes-administratifs-du-Rhone-RAA/Recueils-de-2021',
-        '2020': f'{hostname}/Publications/Recueil-des-actes-administratifs-du-Rhone-RAA/Recueils-de-2020',
-        '2019': f'{hostname}/Publications/Recueil-des-actes-administratifs-du-Rhone-RAA/Recueils-de-2019'
-    }
+    raa_page = f'{hostname}/Publications/Recueil-des-actes-administratifs-du-Rhone-RAA'
     user_agent = 'Mozilla/5.0 (Windows NT 10.0; rv:109.0) Gecko/20100101 Firefox/115.0'
     full_name = 'Préfecture du Rhône'
     short_code = 'pref69'
@@ -29,23 +22,21 @@ class Attrap_pref69(Attrap):
         self.set_sleep_time(30)
 
     def get_raa(self, keywords):
-        pages_to_parse = []
-        if self.not_before.year <= 2024:
-            pages_to_parse.append(self.raa_page['2024'])
-        if self.not_before.year <= 2023:
-            pages_to_parse.append(self.raa_page['2023'])
-        if self.not_before.year <= 2022:
-            pages_to_parse.append(self.raa_page['2022'])
-        if self.not_before.year <= 2021:
-            pages_to_parse.append(self.raa_page['2021'])
-        if self.not_before.year <= 2020:
-            pages_to_parse.append(self.raa_page['2020'])
-        if self.not_before.year <= 2019:
-            pages_to_parse.append(self.raa_page['2019'])
+        # On détermine quelles pages d'année parser
+        year_pages = []
+        page_content = self.get_page(self.raa_page, 'get').content
+        for year_page in self.get_sub_pages(
+            page_content,
+            'div.fr-card__body div.fr-card__content h2.fr-card__title a',
+            self.hostname,
+            False
+        ):
+            year = Attrap.guess_date(year_page['name'].strip(), '.* ([0-9]{4})').year
+            if year < 9999 and year >= self.not_before.year:
+                year_pages.append(year_page['url'])
 
         sub_pages_to_parse = []
-
-        for raa_page in pages_to_parse:
+        for raa_page in year_pages:
             sub_pages = self.get_sub_pages_with_pager(
                 raa_page,
                 'div.fr-card__body div.fr-card__content h2.fr-card__title a.fr-card__link',