From 071a69f2b2390284f1d1b912a600f78a4b05ec52 Mon Sep 17 00:00:00 2001 From: Hadrien <ketsapiwiq@protonmail.com> Date: Sun, 9 Jun 2024 23:19:12 +0200 Subject: [PATCH] wip: pref971: manque extraction de date --- .gitlab-ci.yml | 5 ++ Attrap_pref971.py | 127 ++++++++++++++++++++++++++++++++++++++++++++++ Makefile | 2 + cli.py | 2 + 4 files changed, 136 insertions(+) create mode 100644 Attrap_pref971.py diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml index 3f2c4f9..66ba727 100644 --- a/.gitlab-ci.yml +++ b/.gitlab-ci.yml @@ -232,3 +232,8 @@ test_pref976: variables: PREF: "pref976" extends: .default_pref + +test_pref971: + variables: + PREF: "pref971" + extends: .default_pref diff --git a/Attrap_pref971.py b/Attrap_pref971.py new file mode 100644 index 0000000..556765f --- /dev/null +++ b/Attrap_pref971.py @@ -0,0 +1,127 @@ +import os +import datetime + +from bs4 import BeautifulSoup +from urllib.parse import unquote + +from Attrap import Attrap + +import locale + +locale.setlocale(locale.LC_TIME, 'fr_FR.UTF-8') + + +# https://www.guadeloupe.gouv.fr/Publications/Le-Recueil-des-actes-administratifs/2021/Decembre + + +class Attrap_pref971(Attrap): + months_fr = [ + "Janvier", + "Fevrier", + "Mars", + "Avril", + "Mai", + "Juin", + "Juillet", + "Aout", + "Septembre", + "Octobre", + "Novembre", + "Decembre", + ] + + # Config + __HOST = "https://www.guadeloupe.gouv.fr" + __RAA_PAGE = { + "2015": [], + "2016": [], + "2017": [], + "2018": [], + "2019": [], + "2020": [], + "2021": [], + "2022": [], + "2023": [], + "2024": [], + } + for year in range(2015, 2024): + for month in range(1, 12): + __RAA_PAGE[str(year)].append( + f"{__HOST}/Publications/Le-Recueil-des-actes-administratifs/{year}/{months_fr[month]}" + ) + __USER_AGENT = ( + "Mozilla/5.0 (X11; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/115.0" + ) + full_name = "Préfecture de la Guadeloupe" + short_code = "pref971" + + def __init__(self, data_dir): + super().__init__(data_dir, self.__USER_AGENT) + self.enable_tor(10) + + def get_raa(self, keywords): + year_pages_to_parse = [] + if self.not_before.year <= 2024: + for year_page in self.__RAA_PAGE["2024"]: + year_pages_to_parse.append(year_page) + if self.not_before.year <= 2023: + for year_page in self.__RAA_PAGE["2023"]: + year_pages_to_parse.append(year_page) + if self.not_before.year <= 2022: + for year_page in self.__RAA_PAGE["2022"]: + year_pages_to_parse.append(year_page) + if self.not_before.year <= 2021: + for year_page in self.__RAA_PAGE["2021"]: + year_pages_to_parse.append(year_page) + if self.not_before.year <= 2020: + for year_page in self.__RAA_PAGE["2020"]: + year_pages_to_parse.append(year_page) + if self.not_before.year <= 2019: + for year_page in self.__RAA_PAGE["2019"]: + year_pages_to_parse.append(year_page) + if self.not_before.year <= 2018: + for year_page in self.__RAA_PAGE["2018"]: + year_pages_to_parse.append(year_page) + if self.not_before.year <= 2017: + for year_page in self.__RAA_PAGE["2017"]: + year_pages_to_parse.append(year_page) + if self.not_before.year <= 2016: + for year_page in self.__RAA_PAGE["2016"]: + year_pages_to_parse.append(year_page) + if self.not_before.year <= 2015: + for year_page in self.__RAA_PAGE["2015"]: + year_pages_to_parse.append(year_page) + + # On parse les pages contenant des RAA + elements = [] + for year in self.__RAA_PAGE: + for page in self.__RAA_PAGE[year]: + page_content = self.get_page(page, "get").content + for raa in self.get_raa_elements(page_content): + elements.append(raa) + + 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("a.fr-link.fr-link--download"): + 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.get_text() + date = datetime.datetime.strptime( + name.split("publié le ")[-1].strip().split('\n')[0], "%d %B %Y" + ) + + raa = Attrap.RAA(url, date, name) + elements.append(raa) + return elements diff --git a/Makefile b/Makefile index 8601dd7..7af30ce 100644 --- a/Makefile +++ b/Makefile @@ -53,6 +53,8 @@ pref93: bin/python3 cli.py pref93 pref94: bin/python3 cli.py pref94 +pref971: + bin/python3 cli.py pref971 pref976: bin/python3 cli.py pref976 lint: diff --git a/cli.py b/cli.py index e40a01f..ea8dec8 100755 --- a/cli.py +++ b/cli.py @@ -52,6 +52,7 @@ available_administrations = [ 'pref38', 'pref42', 'pref44', + 'pref45', 'pref59', 'pref62', 'pref63', @@ -66,6 +67,7 @@ available_administrations = [ 'pref87', 'pref93', 'pref94', + 'pref971', 'pref976' ] -- GitLab