mirror of
https://github.com/kovidgoyal/calibre.git
synced 2026-01-08 21:20:21 -05:00
430 lines
19 KiB
Python
430 lines
19 KiB
Python
#!/usr/bin/env python
|
||
# 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.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)
|
||
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
|
||
|
||
|
||
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:
|
||
data = json.loads(raw)['props']['pageProps']['content']
|
||
except KeyError as e:
|
||
raise JSONHasNoContent(e)
|
||
if isinstance(data, list):
|
||
data = data[0]
|
||
body = root.xpath('//body')[0]
|
||
for child in tuple(body):
|
||
body.remove(child)
|
||
article = E(body, 'article')
|
||
E(article, 'div', replace_entities(data['subheadline']) , style='color: red; font-size:small; font-weight:bold;')
|
||
E(article, 'h1', replace_entities(data['headline']))
|
||
E(article, 'div', replace_entities(data['description']), style='font-style: italic; color:#202020;')
|
||
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')
|
||
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=()):
|
||
impl = getattr(soup, 'new_tag', None)
|
||
if impl is not None:
|
||
return impl(name, attrs=dict(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 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 = (
|
||
'The World Ahead is The Economist’s future-gazing publication. It prepares audiences for what is to '
|
||
'come with mind-stretching insights and expert analysis—all in The Economist’s clear, elegant style.'
|
||
' Best downloaded in late November.'
|
||
)
|
||
extra_css = '''
|
||
em { color:#202020; }
|
||
img {display:block; margin:0 auto;}
|
||
'''
|
||
oldest_article = 7.0
|
||
resolve_internal_links = True
|
||
remove_tags = [
|
||
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'
|
||
]
|
||
}
|
||
),
|
||
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']
|
||
# economist.com has started throttling after about 60% of the total has
|
||
# downloaded with connection reset by peer (104) errors.
|
||
delay = 1
|
||
|
||
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'):
|
||
# Reduce image sizes to get file size below amazon's email
|
||
# sending threshold
|
||
self.web2disk_options.compress_news_images = True
|
||
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 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'
|
||
# }])]
|
||
url = 'https://www.economist.com/the-world-ahead'
|
||
# raw = open('/t/raw.html').read()
|
||
raw = self.index_to_soup(url, raw=True)
|
||
# with open('/t/raw.html', 'wb') as f:
|
||
# f.write(raw)
|
||
soup = self.index_to_soup(raw)
|
||
# nav = soup.find(attrs={'class':'navigation__wrapper'})
|
||
# if nav is not None:
|
||
# a = nav.find('a', href=lambda x: x and '/printedition/' in x)
|
||
# if a is not None:
|
||
# self.log('Following nav link to current edition', a['href'])
|
||
# soup = self.index_to_soup(process_url(a['href']))
|
||
ans = self.economist_parse_index(soup)
|
||
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 economist_parse_index(self, soup):
|
||
script_tag = soup.find("script", id="__NEXT_DATA__")
|
||
if script_tag is not None:
|
||
data = json.loads(script_tag.string)
|
||
# open('/t/raw.json', 'w').write(json.dumps(data, indent=2, sort_keys=True))
|
||
self.title = safe_dict(data, "props", "pageProps", "content", "headline")
|
||
# self.cover_url = 'https://mma.prnewswire.com/media/2275620/The_Economist_The_World_Ahead_2024.jpg?w=600'
|
||
|
||
feeds = []
|
||
|
||
for coll in safe_dict(data, "props", "pageProps", "content", "collections"):
|
||
section = safe_dict(coll, "headline") or ''
|
||
self.log(section)
|
||
articles = []
|
||
for part in safe_dict(coll, "hasPart", "parts"):
|
||
title = safe_dict(part, "headline") or ''
|
||
url = safe_dict(part, "url", "canonical") or ''
|
||
if not title or not url:
|
||
continue
|
||
desc = safe_dict(part, "description") or ''
|
||
sub = safe_dict(part, "subheadline") or ''
|
||
if sub:
|
||
desc = sub + ' :: ' + desc
|
||
self.log('\t', title, '\n\t', desc, '\n\t\t', url)
|
||
articles.append({'title': title, 'description':desc, 'url': url})
|
||
if articles:
|
||
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:
|
||
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])
|