Frontline by unkn0wn

recipes/frontline.recipe

@@ -0,0 +1,67 @@
+#!/usr/bin/env python
+# vim:fileencoding=utf-8
+from calibre.web.feeds.news import BasicNewsRecipe, classes
+
+
+class Frontline(BasicNewsRecipe):
+    title = u'Frontline'
+    __author__ = 'unkn0wn'
+    description = 'Frontline, the fortnightly English magazine from the stable of The Hindu, has been a distinguished presence in the media world since 1984.'
+    language = 'en_IN'
+    no_stylesheets = True
+    remove_javascript = True
+    use_embedded_content = False
+    encoding = 'utf-8'
+    oldest_article = 14
+    max_articles_per_feed = 50
+    ignore_duplicate_articles = {'url'}
+    # masthead_url = 'https://fl.thgim.com/static/theme/default/base/img/fllogo.png'
+    remove_attributes = ['height', 'width']
+
+    def get_cover_url(self):
+        soup = self.index_to_soup(
+            'https://frontline.thehindu.com/current-issue/')
+        tag = soup.find(attrs={'class': 'imgWrapper'})
+        if tag:
+            self.cover_url = tag.find('img')['data-proxy-image'].replace(
+                "FREE_100", "FREE_810")
+        return super().get_cover_url()
+
+    # https://fl.thgim.com/incoming/b5zy2g/article38454943.ece/alternates/FREE_100/coverpng
+
+    keep_only_tags = [
+        classes(
+            'overline mainart-title marginBottom10px articleBottomLine swiper-slide slide-caption artlead-text body-main'
+        )
+    ]
+    remove_tags = [classes('dispatche-middle bigtitle')]
+
+    remove_tags_after = [
+        classes('body-main'),
+    ]
+
+    feeds = [
+        ('Cover Story',
+         'https://frontline.thehindu.com/cover-story/feeder/default.rss'),
+        ('The Nation',
+         'https://frontline.thehindu.com/the-nation/feeder/default.rss'),
+        ('World Affairs',
+         'https://frontline.thehindu.com/world-affairs/feeder/default.rss'),
+        ('Politics',
+         'https://frontline.thehindu.com/politics/feeder/default.rss'),
+        ('Arts & Culture',
+         'https://frontline.thehindu.com/arts-and-culture/feeder/default.rss'),
+        ('Social Issues',
+         'https://frontline.thehindu.com/social-issues/feeder/default.rss'),
+        ('Books', 'https://frontline.thehindu.com/books/feeder/default.rss'),
+        ('Columns',
+         'https://frontline.thehindu.com/columns/feeder/default.rss'),
+        ('Others', 'https://frontline.thehindu.com/other/feeder/default.rss'),
+    ]
+
+    def preprocess_html(self, soup):
+        for img in soup.findAll('img', attrs={'data-proxy-image': True}):
+            # replace FREE_810 with FREE_615 FREE_1200.. FREE_300.
+            img['src'] = img['data-proxy-image'].replace(
+                "FREE_100", "FREE_810")
+        return soup