calibre/recipes/mediapart.recipe

#!/usr/bin/env python
# vim:fileencoding=utf-8
#
# 11 Jan 2021 -  L. Houpert - Major changes in the Mediapart recipe:
#   1) Summary of the article are noow available
#   2) Additional sections  International, France, Economie and Culture have
# been added through custom entries in the function my_parse_index.
#   3) Fix the cover image so it doesn't disappear from the Kindle menu
# ( cover image format is changed to .jpeg)
# 14 Jan 2021 - Add Mediapart Logo url as masthead_url and change cover
#   by overlaying the date on top of the Mediapart cover
# 22 Mar 2023 - Switch to Google feeds

from __future__ import unicode_literals

__license__ = 'GPL v3'
__copyright__ = '2021, Loïc Houpert <houpertloic at gmail .com>. Adapted from: 2016, Daniel Bonnery; 2009, Mathieu Godlewski; 2010-2012, Louis Gesbert'  # noqa
'''
Mediapart
'''

from datetime import datetime, timezone, timedelta
from calibre.ptempfile import PersistentTemporaryFile
from calibre.web.feeds.news import BasicNewsRecipe, classes

class Mediapart(BasicNewsRecipe):
    title = 'Mediapart'
    __author__ = 'Loïc Houpert, unkn0wn'
    description = 'Global news in French from news site Mediapart'
    publication_type = 'newspaper'
    language = 'fr'
    needs_subscription = True

    use_embedded_content = False
    no_stylesheets = True

    keep_only_tags = [
        classes(
            'news__heading__top news__heading__center news__body__center__article'
        )
    ]

    remove_tags = [
        classes('action-links media--rich read-also login-subscribe print-source_url'),
        dict(name='svg'),
    ]

    conversion_options = {'smarten_punctuation': True}

    masthead_url = "https://raw.githubusercontent.com/lhoupert/calibre_contrib/main/mediapart_masthead.png"

    ignore_duplicate_articles = {'title'}
    resolve_internal_links  = True
    remove_empty_feeds = True

    articles_are_obfuscated = True

    def get_obfuscated_article(self, url):
        br = self.get_browser()
        try:
            br.open(url)
        except Exception as e:
            url = e.hdrs.get('location')
        soup = self.index_to_soup(url)
        link = soup.find('a', href=True)
        skip_sections =[ # add sections you want to skip
            '/video/', '/videos/', '/media/'
        ]
        if any(x in link['href'] for x in skip_sections):
            self.log('Aborting Article ', link['href'])
            self.abort_article('skipping video links')

        self.log('Downloading ', link['href'])
        html = br.open(link['href']).read()
        pt = PersistentTemporaryFile('.html')
        pt.write(html)
        pt.close()
        return pt.name

    feeds = []

    sections = [
        'france', 'international', 'economie', 'culture-idees', 'politique', 'ecologie', 'fil-dactualites'
    ]

    for sec in sections:
        a = 'https://news.google.com/rss/search?q=when:27h+allinurl:mediapart.fr%2Fjournal{}&hl=fr-FR&gl=FR&ceid=FR:fr'
        feeds.append((sec.capitalize(), a.format('%2F' + sec + '%2F')))
    feeds.append(('Autres', a.format('')))

    def get_browser(self):
        # -- Handle login

        def is_form_login(form):
            return "id" in form.attrs and form.attrs['id'] == "logFormEl"

        br = BasicNewsRecipe.get_browser(self)
        if self.username is not None and self.password is not None:
            br.open('https://www.mediapart.fr/login')
            br.select_form(predicate=is_form_login)
            br['name'] = self.username
            br['password'] = self.password
            br.submit()
        return br

    def default_cover(self, cover_file):
        '''
        Create a generic cover for recipes that don't have a cover
        '''
        from qt.core import QImage, QPainter, QPen, Qt, QFont, QRect
        from calibre.gui2 import ensure_app, load_builtin_fonts, pixmap_to_data

        def init_environment():
            ensure_app()
            load_builtin_fonts()

        def create_cover_mediapart(date):
            ' Create a cover for mediapart adding the date on Mediapart Cover'
            init_environment()
            # Get data
            image_url = 'https://raw.githubusercontent.com/lhoupert/calibre_contrib/main/mediapart.jpeg'
            data = self.index_to_soup(image_url, raw=True)
            # Get date and hour corresponding to french time zone
            today = datetime.now(timezone.utc) + timedelta(hours=1)
            wkd = today.weekday()
            french_weekday={0:'Mon',1:'Mar',2:'Mer',3:'Jeu',4:'Ven',5:'Sam',6:'Dim'}
            day = french_weekday[wkd]+'.'
            date = day + ' ' + today.strftime('%d %b. %Y')
            edition = today.strftime('Édition de %Hh')

            # Get Cover data
            img  = QImage()
            img.loadFromData(data)

            # Overlay date on cover
            p = QPainter(img)
            pen = QPen(Qt.black)
            pen.setWidth(6)
            p.setPen(pen)
            font = QFont()
            font.setFamily('Times')
            font.setPointSize(72)
            p.setFont(font)
            r = QRect(0, 600, 744,100)
            p.drawText(r, Qt.AlignmentFlag.AlignJustify | Qt.AlignmentFlag.AlignVCenter | Qt.AlignmentFlag.AlignCenter, date)
            p.end()

            # Overlay edition information on cover
            p = QPainter(img)
            pen = QPen(Qt.black)
            pen.setWidth(4)
            p.setPen(pen)
            font = QFont()
            font.setFamily('Times')
            font.setItalic(True)
            font.setPointSize(66)
            p.setFont(font)
            # Add date
            r = QRect(0, 720, 744,100)
            p.drawText(r, Qt.AlignmentFlag.AlignJustify | Qt.AlignmentFlag.AlignVCenter | Qt.AlignmentFlag.AlignCenter, edition)
            p.end()
            return pixmap_to_data(img)

        try:
            today=datetime.today()
            date = today.strftime('%d %b %Y')
            img_data = create_cover_mediapart(date)
            cover_file.write(img_data)
            cover_file.flush()
        except Exception:
            self.log.exception('Failed to generate default cover')
            return False
        return True