Update economist_espresso.recipe

recipes/economist_espresso.recipe

@@ -2,8 +2,92 @@
 https://www.economist.com/the-world-in-brief
 '''
 
-from calibre.ebooks.BeautifulSoup import Tag
-from calibre.web.feeds.news import BasicNewsRecipe, classes
+import json
+import time
+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.web.feeds.news import BasicNewsRecipe
+from html5_parser import parse
+from lxml import etree
+
+
+def E(parent, name, text='', **attrs):
+    ans = parent.makeelement(name, **attrs)
+    ans.text = text
+    parent.append(ans)
+    return ans
+
+
+def process_node(node, html_parent):
+    ntype = node.get('type')
+    if ntype == 'tag':
+        c = html_parent.makeelement(node['name'])
+        c.attrib.update({k: v or '' for k, v in node.get('attribs', {}).items()})
+        html_parent.append(c)
+        for nc in node.get('children', ()):
+            process_node(nc, c)
+    elif ntype == 'text':
+        text = node.get('data')
+        if text:
+            text = replace_entities(text)
+            if len(html_parent):
+                t = html_parent[-1]
+                t.tail = (t.tail or '') + text
+            else:
+                html_parent.text = (html_parent.text or '') + text
+
+
+def safe_dict(data, *names):
+    ans = data
+    for x in names:
+        ans = ans.get(x) or {}
+    return ans
+
+
+class JSONHasNoContent(ValueError):
+    pass
+
+
+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;')
+    E(article, 'div', data['byline'], style='font-style: italic; color:#202020;')
+    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)
+
+def cleanup_html_article(root):
+    main = root.xpath('//main')[0]
+    body = root.xpath('//body')[0]
+    for child in tuple(body):
+        body.remove(child)
+    body.append(main)
+    main.set('id', '')
+    main.tag = 'article'
+    for x in root.xpath('//*[@style]'):
+        x.set('style', '')
+    for x in root.xpath('//button'):
+        x.getparent().remove(x)
+
+
+def classes(classes):
+    q = frozenset(classes.split(' '))
+    return dict(attrs={
+        'class': lambda x: x and frozenset(x.split()).intersection(q)})
 
 
 def new_tag(soup, name, attrs=()):
@@ -13,64 +97,179 @@ def new_tag(soup, name, attrs=()):
     return Tag(soup, name, attrs=attrs or None)
 
 
+class NoArticles(Exception):
+    pass
+
+
+def process_url(url):
+    if url.startswith('/'):
+        url = 'https://www.economist.com' + url
+    return url
+
+
 class Espresso(BasicNewsRecipe):
     title = 'The Economist Espresso'
     language = 'en'
     __author__ = 'unkn0wn'
+    encoding = 'utf-8'
+    masthead_url = 'https://www.livemint.com/lm-img/dev/economist-logo-oneline.png'
+    cover_url = 'https://downloadr2.apkmirror.com/wp-content/uploads/2021/10/75/615777cc6611b.png'
     description = (
         'Espresso is a rich, full-flavoured shot of daily global analysis'
         ' from the editors of The Economist to get you up to speed, fast.'
          'Maximise your understanding of the most significant business, '
          'economic, political and cultural developments globally.'
     )
-    cover_url = 'https://downloadr2.apkmirror.com/wp-content/uploads/2021/10/75/615777cc6611b.png'
-    no_stylesheets = True
-    remove_attributes = ['height', 'width', 'style']
-    use_embedded_content = False
-    masthead_url = 'https://www.livemint.com/lm-img/dev/economist-logo-oneline.png'
 
     extra_css = '''
-        h1 { text-align:center; }
-        ._main-image, ._description, .sub { text-align:center; font-size:small; }
-        ._quote-container { font-size:x-large; font-style:italic; color:#202020; }
+        em { color:#202020; }
+        img {display:block; margin:0 auto;}
     '''
 
-    keep_only_tags = [
-        dict(name='main', attrs={'id':'content'})
-    ]
-
     remove_tags = [
-        classes('_podcast-promo _newsletter-promo-container _time-last-updated')
-    ]
-
-    def parse_index(self):
-        return [
-            ('Espresso',
-                [
-                    {
-                        'title': 'The World in Brief',
-                        'url': 'https://www.economist.com/the-world-in-brief',
-                        'description': 'Catch up quickly on the global stories that matter'
-                    },
+        dict(name=['script', 'noscript', 'title', 'iframe', 'cf_floatingcontent', 'aside', 'footer']),
+        dict(attrs={'aria-label': "Article Teaser"}),
+        dict(attrs={
+                'class': [
+                    'dblClkTrk', 'ec-article-info', 'share_inline_header',
+                    'related-items', 'main-content-container', 'ec-topic-widget',
+                    'teaser', 'blog-post__bottom-panel-bottom', 'blog-post__comments-label',
+                    'blog-post__foot-note', 'blog-post__sharebar', 'blog-post__bottom-panel',
+                    'newsletter-form','share-links-header','teaser--wrapped', 'latest-updates-panel__container',
+                    'latest-updates-panel__article-link','blog-post__section'
                 ]
-            ),
-        ]
-
-    def print_version(self, url):
-        return 'https://webcache.googleusercontent.com/search?q=cache:' + url
-
-    def preprocess_html(self, soup):
-        if h1 := soup.find('h1'):
-            if p := h1.find_next_sibling('p'):
-                p['class'] = 'sub'
-        for hr in soup.findAll(attrs={'class':['_gobbet', '_article']}):
-            nt = new_tag(soup, 'hr')
-            hr.append(nt)
-        return soup
+            }
+        ),
+        dict(attrs={
+                'class': lambda x: x and 'blog-post__siblings-list-aside' in x.split()}),
+        classes(
+            'share-links-header teaser--wrapped latest-updates-panel__container'
+            ' latest-updates-panel__article-link blog-post__section newsletter-form blog-post__bottom-panel'
+        )
+    ]
+    keep_only_tags = [dict(name='article', id=lambda x: not x)]
+    no_stylesheets = True
+    remove_attributes = ['data-reactid', 'width', 'height']
 
     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_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
+
+    def parse_index(self):
+		# return self.economist_test_article()
+		# url = 'https://www.economist.com/weeklyedition/archive'
+        query = {
+            'query': 'query EspressoQuery($ref:String!){espresso:canonical(ref:$ref){...EspressoFragment __typename}}fragment EspressoFragment on Content{id type hasPart(size:1 sort:"datePublished:desc"){parts{id type rubric:description hasPart(sort:"publication.context.position:asc,datePublished:desc"){parts{...ArticleFragment __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': 'EspressoQuery',
+            'variables': '{"ref":"/content/jakj5ed3rml75i8j0d5i74p8adf6eem4"}',
+        }
+        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 after some time.')
+        ans = self.economist_parse_index(raw)
+        return self.economist_return_index(ans)
+
+    def economist_parse_index(self, raw):
+        data = json.loads(raw)['data']['espresso']['hasPart']['parts'][0]
+        self.description = data['rubric']
+
+        ans = []
+        for part in safe_dict(data, "hasPart", "parts"):
+            title = safe_dict(part, "title")
+            pt = PersistentTemporaryFile('.html')
+            pt.write(json.dumps(part).encode('utf-8'))
+            pt.close()
+            url = 'file:///' + pt.name
+            ans.append({"title": title, "url": url})
+        return [('The world in brief', ans)]
+
+    def populate_article_metadata(self, article, soup, first):
+        article.url = soup.find('h1')['title']
+
+    def preprocess_raw_html(self, raw, url):
+        # open('/t/raw.html', 'wb').write(raw.encode('utf-8'))
+        body = '<html><body><article></article></body></html>'
+        root = parse(body)
+        load_article_from_json(raw, 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 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:
+                yield x
+
+    def postprocess_html(self, soup, first):
+        for img in soup.findAll('img', srcset=True):
+            del img['srcset']
+        for table in list(self.eco_find_image_tables(soup)):
+            caption = table.find('font')
+            img = table.find('img')
+            div = new_tag(soup, 'div')
+            div['style'] = 'text-align:left;font-size:70%'
+            ns = NavigableString(self.tag_to_string(caption))
+            div.insert(0, ns)
+            div.insert(1, new_tag(soup, 'br'))
+            del img['width']
+            del img['height']
+            img.extract()
+            div.insert(2, img)
+            table.replaceWith(div)
+        return soup
+
+    def canonicalize_internal_url(self, url, is_link=True):
+        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])