User:Mibble92/python bot scripts/get featureds.py

import urllib.request as request
import sys
import re

def get_featureds(*args):
    page_name = 'https://en.wikipedia.org/w/index.php?title=Wikipedia:Featured_articles&action=raw';
    if (len(args) > 1):
        page_name = page_name + '&oldid=' + args[1];
    page = request.urlopen(page_name);
    if ((resp_code := page.getcode()) != 200):
        return (resp_code, page.msg);

    regex_list = ['^Image:', '^WP:', '^[Ww]ikipedia:', '^[Tt]emplate', '^[Mm]edia[Ww]iki:', '^[Hh]elp:', 'Main page', 'be bold', '^..:', '^Category:', '^File:', '^Special:', '^\#']
    with open('featureds', 'w') as output_file:
        for line in page.readlines():
            line_str = line.decode('utf-8')
            article = re.search("\[\[.*\]\]", line_str)
            if (article is not None):
                article_title = re.sub('\[\[|\]\]|\|.*', '', article.group(0))
                if (not any (re.match(pattern, article_title) for pattern in regex_list)):
                    print ('FEATURED||' + article_title, file=output_file)
    output_file.close()


if __name__ == '__main__':
  get_featureds(sys.argv)