Update The World Ahead

recipes/economist_world_ahead.recipe

@@ -2,13 +2,20 @@
 # License: GPLv3 Copyright: 2008, Kovid Goyal <kovid at kovidgoyal.net>
 
 import json
+import time
+from collections import defaultdict
+from datetime import datetime, timedelta
+from urllib.parse import quote, urlencode
 
 from calibre import replace_entities
 from calibre.ebooks.BeautifulSoup import NavigableString, Tag
+from calibre.ptempfile import PersistentTemporaryFile
+from calibre.utils.date import parse_only_date
 from calibre.web.feeds.news import BasicNewsRecipe
 from html5_parser import parse
 from lxml import etree
 
+use_archive = True
 
 def E(parent, name, text='', **attrs):
     ans = parent.makeelement(name, **attrs)
@@ -47,6 +54,35 @@ class JSONHasNoContent(ValueError):
     pass
 
 
+if use_archive:
+    def load_article_from_json(raw, root):
+        # open('/t/raw.json', 'w').write(raw)
+        data = json.loads(raw)
+        body = root.xpath('//body')[0]
+        article = E(body, 'article')
+        E(article, 'div', data['flyTitle'] , style='color: red; font-size:small; font-weight:bold;')
+        E(article, 'h1', data['title'], title=safe_dict(data, "url", "canonical") or '')
+        E(article, 'div', data['rubric'], style='font-style: italic; color:#202020;')
+        try:
+            date = data['dateModified']
+        except Exception:
+            date = data['datePublished']
+        dt = datetime.fromisoformat(date[:-1]) + timedelta(seconds=time.timezone)
+        dt = dt.strftime('%b %d, %Y, %I:%M %p')
+        if data['dateline'] is None:
+            E(article, 'p', dt, style='color: gray; font-size:small;')
+        else:
+            E(article, 'p', dt + ' | ' + (data['dateline']), style='color: gray; font-size:small;')
+        main_image_url = safe_dict(data, 'image', 'main', 'url').get('canonical')
+        if main_image_url:
+            div = E(article, 'div')
+            try:
+                E(div, 'img', src=main_image_url)
+            except Exception:
+                pass
+        for node in data.get('text') or ():
+            process_node(node, article)
+else:
     def load_article_from_json(raw, root):
         # open('/t/raw.json', 'w').write(raw)
         try:
@@ -59,10 +95,13 @@ def load_article_from_json(raw, root):
         for child in tuple(body):
             body.remove(child)
         article = E(body, 'article')
-    E(article, 'h4', data['subheadline'], style='color: red; margin: 0')
+        E(article, 'div', replace_entities(data['subheadline']) , style='color: red; font-size:small; font-weight:bold;')
-    E(article, 'h1', data['headline'], style='font-size: x-large')
+        E(article, 'h1', replace_entities(data['headline']))
-    E(article, 'div', data['description'], style='font-style: italic; color: #202020;')
+        E(article, 'div', replace_entities(data['description']), style='font-style: italic; color:#202020;')
-    E(article, 'div', (data['datePublishedString'] or '') + ' | ' + (data['dateline'] or ''), style='color: gray; margin: 1em')
+        if data['dateline'] is None:
+            E(article, 'p', (data['datePublishedString'] or ''), style='color: gray; font-size:small;')
+        else:
+            E(article, 'p', (data['datePublishedString'] or '') + ' | ' + (data['dateline']), style='color: gray; font-size:small;')
         main_image_url = safe_dict(data, 'image', 'main', 'url').get('canonical')
         if main_image_url:
             div = E(article, 'div')
@@ -116,6 +155,7 @@ class Economist(BasicNewsRecipe):
     title = 'The Economist World Ahead'
     language = 'en'
     encoding = 'utf-8'
+    masthead_url = 'https://www.livemint.com/lm-img/dev/economist-logo-oneline.png'
 
     __author__ = "Kovid Goyal"
     description = (
@@ -124,29 +164,7 @@ class Economist(BasicNewsRecipe):
         ' Best downloaded in late November.'
     )
     extra_css = '''
-        .headline {font-size: x-large;}
+        em { color:#202020; }
-        h2 { font-size: small;  }
-        h1 { font-size: medium;  }
-        em.Bold {font-weight:bold;font-style:normal;}
-        em.Italic {font-style:italic;}
-        p.xhead {font-weight:bold;}
-        .pullquote {
-            float: right;
-            font-size: larger;
-            font-weight: bold;
-            font-style: italic;
-            page-break-inside:avoid;
-            border-bottom: 3px solid black;
-            border-top: 3px solid black;
-            width: 228px;
-            margin: 0px 0px 10px 15px;
-            padding: 7px 0px 9px;
-        }
-        .flytitle-and-title__flytitle {
-            display: block;
-            font-size: smaller;
-            color: red;
-        }
         img {display:block; margin:0 auto;}
     '''
     oldest_article = 7.0
@@ -181,6 +199,83 @@ class Economist(BasicNewsRecipe):
 
     needs_subscription = False
 
+    def get_browser(self, *args, **kwargs):
+        # Needed to bypass cloudflare
+        kwargs['user_agent'] = 'common_words/based'
+        br = BasicNewsRecipe.get_browser(self, *args, **kwargs)
+        br.addheaders += [('Accept-Language', 'en-GB,en-US;q=0.9,en;q=0.8')]
+        return br
+
+    def economist_test_article(self):
+        return [('Articles', [{'title':'test',
+            'url':'https://www.economist.com/the-americas/2024/04/14/elon-musk-is-feuding-with-brazils-powerful-supreme-court'
+        }])]
+
+    def economist_return_index(self, ans):
+        if not ans:
+            raise NoArticles(
+                'Could not find any articles, either the '
+                'economist.com server is having trouble and you should '
+                'try later or the website format has changed and the '
+                'recipe needs to be updated.'
+            )
+        return ans
+
+    if use_archive:
+        def parse_index(self):
+            # return self.economist_test_article()
+            soup = self.index_to_soup('https://www.economist.com/the-world-ahead')
+            script_tag = soup.find("script", id="__NEXT_DATA__")
+            if script_tag is None:
+                raise ValueError('No script tag with JSON data found in the weeklyedition archive')
+            data = json.loads(script_tag.string)
+            content_id = data['props']['pageProps']['content']['tegID'].split('/')[-1]
+            query = {
+                'query': 'query HubsDataQuery($id:String!$size:Int!){canonical(ref:$id){id headline description url{canonical __typename}image{ident{url{canonical __typename}width height __typename}__typename}text(mode:"hub" format:"json")hasPart(size:$size){parts{id title:headline isPartOf{context{title:headline __typename}__typename}hasPart{parts{...ArticleFragment isPartOf{id context{title:headline flyTitle:subheadline rubric:description dateline image{...ImageMainFragment ...ImagePromoFragment __typename}__typename}__typename}__typename}__typename}__typename}__typename}__typename}}fragment ArticleFragment on Content{ad{grapeshot{channels{name __typename}__typename}__typename}articleSection{internal{id title:headline __typename}__typename}audio{main{id duration(format:"seconds")source:channel{id __typename}url{canonical __typename}__typename}__typename}byline dateline dateModified datePublished dateRevised flyTitle:subheadline id image{...ImageInlineFragment ...ImageMainFragment ...ImagePromoFragment __typename}print{title:headline flyTitle:subheadline rubric:description section{id title:headline __typename}__typename}publication{id tegID title:headline flyTitle:subheadline datePublished regionsAllowed url{canonical __typename}__typename}rubric:description source:channel{id __typename}tegID text(format:"json")title:headline type url{canonical __typename}topic contentIdentity{forceAppWebview mediaType articleType __typename}__typename}fragment ImageInlineFragment on Media{inline{url{canonical __typename}width height __typename}__typename}fragment ImageMainFragment on Media{main{url{canonical __typename}width height __typename}__typename}fragment ImagePromoFragment on Media{promo{url{canonical __typename}id width height __typename}__typename}',  # noqa
+                'operationName': 'HubsDataQuery',
+                'variables': '{{"id":"/content/{}","size":40}}'.format(content_id),
+            }
+            url = 'https://cp2-graphql-gateway.p.aws.economist.com/graphql?' + urlencode(query, safe='()!', quote_via=quote)
+            try:
+                raw = self.index_to_soup(url, raw=True)
+            except Exception:
+                raise ValueError('Server is not reachable, try again some other time.')
+            ans = self.economist_parse_index(raw)
+            return self.economist_return_index(ans)
+
+        def economist_parse_index(self, raw):
+            data = json.loads(raw)['data']['canonical']
+            self.description = data['description']
+
+            feeds_dict = defaultdict(list)
+            for part in safe_dict(data, "hasPart", "parts"):
+                section = part['title']
+                self.log(section)
+                for art in safe_dict(part, "hasPart", "parts"):
+                    title = safe_dict(art, "title")
+                    desc = safe_dict(art, "rubric") or ''
+                    sub = safe_dict(art, "flyTitle") or ''
+                    if sub and section != sub:
+                        desc = sub + ' :: ' + desc
+                    pt = PersistentTemporaryFile('.html')
+                    pt.write(json.dumps(art).encode('utf-8'))
+                    pt.close()
+                    url = 'file:///' + pt.name
+                    feeds_dict[section].append({"title": title, "url": url, "description": desc})
+                    self.log('\t', title, '\n\t\t', desc)
+            return [(section, articles) for section, articles in feeds_dict.items()]
+
+        def populate_article_metadata(self, article, soup, first):
+            article.url = soup.find('h1')['title']
+
+        def preprocess_html(self, soup):
+            for img in soup.findAll('img', src=True):
+                img['src'] = img['src'].replace('economist.com/', 
+                    'economist.com/cdn-cgi/image/width=600,quality=80,format=auto/')
+            return soup
+
+    else: # Load articles from individual article pages {{{
+
         def __init__(self, *args, **kwargs):
             BasicNewsRecipe.__init__(self, *args, **kwargs)
             if self.output_profile.short_name.startswith('kindle'):
@@ -190,52 +285,6 @@ class Economist(BasicNewsRecipe):
                 self.web2disk_options.compress_news_images_auto_size = 5
                 self.log.warn('Kindle Output profile being used, reducing image quality to keep file size below amazon email threshold')
 
-
-    def get_browser(self, *args, **kwargs):
-        # Needed to bypass cloudflare
-        kwargs['user_agent'] = 'common_words/based'
-        br = BasicNewsRecipe.get_browser(self, *args, **kwargs)
-        br.addheaders += [('Accept-Language', 'en-GB,en-US;q=0.9,en;q=0.8')]
-        return br
-
-    def preprocess_raw_html(self, raw, url):
-        # open('/t/raw.html', 'wb').write(raw.encode('utf-8'))
-        root = parse(raw)
-        if '/interactive/' in url:
-            return '<html><body><article><h1 class="headline">' + root.xpath('//h1')[0].text + '</h1><em>' \
-                        + 'This article is supposed to be read in a browser' \
-                            + '</em></article></body></html>'
-        script = root.xpath('//script[@id="__NEXT_DATA__"]')
-        if script:
-            try:
-                load_article_from_json(script[0].text, root)
-            except JSONHasNoContent:
-                cleanup_html_article(root)
-        for div in root.xpath('//div[@class="lazy-image"]'):
-            noscript = list(div.iter('noscript'))
-            if noscript and noscript[0].text:
-                img = list(parse(noscript[0].text).iter('img'))
-                if img:
-                    p = noscript[0].getparent()
-                    idx = p.index(noscript[0])
-                    p.insert(idx, p.makeelement('img', src=img[0].get('src')))
-                    p.remove(noscript[0])
-        for x in root.xpath('//*[name()="script" or name()="style" or name()="source" or name()="meta"]'):
-            x.getparent().remove(x)
-        # the economist uses <small> for small caps with a custom font
-        for x in root.xpath('//small'):
-            if x.text and len(x) == 0:
-                x.text = x.text.upper()
-                x.tag = 'span'
-                x.set('style', 'font-variant: small-caps')
-        for x in root.xpath('//figcaption'):
-            x.set('style', 'text-align:center; font-size:small;')
-        for x in root.xpath('//cite'):
-            x.tag = 'blockquote'
-            x.set('style', 'color:#404040;')
-        raw = etree.tostring(root, encoding='unicode')
-        return raw
-
         def parse_index(self):
             # return [('Articles', [{'title':'test',
             #     'url':'https://www.economist.com/interactive/briefing/2022/06/11/huge-foundation-models-are-turbo-charging-ai-progress'
@@ -291,6 +340,58 @@ class Economist(BasicNewsRecipe):
                         feeds.append((section, articles))
                 return feeds
 
+        # }}}
+
+
+    def preprocess_raw_html(self, raw, url):
+        # open('/t/raw.html', 'wb').write(raw.encode('utf-8'))
+        if use_archive:
+            body = '<html><body><article></article></body></html>'
+            root = parse(body)
+            load_article_from_json(raw, root)
+        else:
+            root = parse(raw)
+            script = root.xpath('//script[@id="__NEXT_DATA__"]')
+            if script:
+                try:
+                    load_article_from_json(script[0].text, root)
+                except JSONHasNoContent:
+                    cleanup_html_article(root)
+
+        if '/interactive/' in url:
+            return '<html><body><article><h1>' + root.xpath('//h1')[0].text + '</h1><em>' \
+                        + 'This article is supposed to be read in a browser' \
+                            + '</em></article></body></html>'
+
+        for div in root.xpath('//div[@class="lazy-image"]'):
+            noscript = list(div.iter('noscript'))
+            if noscript and noscript[0].text:
+                img = list(parse(noscript[0].text).iter('img'))
+                if img:
+                    p = noscript[0].getparent()
+                    idx = p.index(noscript[0])
+                    p.insert(idx, p.makeelement('img', src=img[0].get('src')))
+                    p.remove(noscript[0])
+        for x in root.xpath('//*[name()="script" or name()="style" or name()="source" or name()="meta"]'):
+            x.getparent().remove(x)
+        # the economist uses <small> for small caps with a custom font
+        for init in root.xpath('//span[@data-caps="initial"]'):
+            init.set('style', 'font-weight:bold;')
+        for x in root.xpath('//small'):
+            if x.text and len(x) == 0:
+                x.text = x.text.upper()
+                x.tag = 'span'
+                x.set('style', 'font-variant: small-caps')
+        for h2 in root.xpath('//h2'):
+            h2.tag = 'h4'
+        for x in root.xpath('//figcaption'):
+            x.set('style', 'text-align:center; font-size:small;')
+        for x in root.xpath('//cite'):
+            x.tag = 'blockquote'
+            x.set('style', 'color:#404040;')
+        raw = etree.tostring(root, encoding='unicode')
+        return raw
+
     def eco_find_image_tables(self, soup):
         for x in soup.findAll('table', align=['right', 'center']):
             if len(x.findAll('font')) in (1, 2) and len(x.findAll('img')) == 1:
@@ -318,3 +419,12 @@ class Economist(BasicNewsRecipe):
         if url.endswith('/print'):
             url = url.rpartition('/')[0]
         return BasicNewsRecipe.canonicalize_internal_url(self, url, is_link=is_link)
+
+
+def get_login_cookies(username, password):
+    print(33333333333, username, password)
+
+
+if __name__ == '__main__':
+    import sys
+    get_login_cookies(sys.argv[-2], sys.argv[-1])