mirror of
https://github.com/kovidgoyal/calibre.git
synced 2025-09-29 15:31:08 -04:00
321 lines
12 KiB
Python
321 lines
12 KiB
Python
#!/usr/bin/env python
|
|
# License: GPLv3 Copyright: 2008, Kovid Goyal <kovid at kovidgoyal.net>
|
|
|
|
import json
|
|
import time
|
|
from datetime import datetime, timedelta
|
|
|
|
from html5_parser import parse
|
|
from lxml import etree
|
|
|
|
from calibre.ebooks.BeautifulSoup import NavigableString, Tag
|
|
from calibre.web.feeds.news import BasicNewsRecipe
|
|
|
|
|
|
def process_list(li_node):
|
|
li_html = ''
|
|
for li in li_node['items']:
|
|
if li.get('textHtml'):
|
|
li_html += f'<li>{li.get("textHtml")}</li>'
|
|
else:
|
|
li_html += f'<li>{li.get("text", "")}</li>'
|
|
return li_html
|
|
|
|
|
|
def process_info_box(bx):
|
|
info = ''
|
|
for x in safe_dict(bx, 'components'):
|
|
info += f'<blockquote>{process_node(x)}</blockquote>'
|
|
return info
|
|
|
|
|
|
def process_node(node):
|
|
ntype = node.get('type', '')
|
|
if ntype == 'CROSSHEAD':
|
|
if node.get('textHtml'):
|
|
return f'<h4>{node.get("textHtml")}</h4>'
|
|
return f'<h4>{node.get("text", "")}</h4>'
|
|
elif ntype in ['PARAGRAPH', 'BOOK_INFO']:
|
|
if node.get('textHtml'):
|
|
return f'<p>{node.get("textHtml")}</p>'
|
|
return f'<p>{node.get("text", "")}</p>'
|
|
elif ntype == 'IMAGE':
|
|
alt = '' if node.get('altText') is None else node.get('altText')
|
|
cap = ''
|
|
if node.get('caption'):
|
|
if node['caption'].get('textHtml') is not None:
|
|
cap = node['caption']['textHtml']
|
|
return f'<div><img src="{node["url"]}" title="{alt}"></div><div style="text-align:center; font-size:small;">{cap}</div>'
|
|
elif ntype == 'PULL_QUOTE':
|
|
if node.get('textHtml'):
|
|
return f'<blockquote>{node.get("textHtml")}</blockquote>'
|
|
return f'<blockquote>{node.get("text", "")}</blockquote>'
|
|
elif ntype == 'DIVIDER':
|
|
return '<hr>'
|
|
elif ntype == 'INFOGRAPHIC':
|
|
if node.get('fallback'):
|
|
return process_node(node['fallback'])
|
|
elif ntype == 'INFOBOX':
|
|
return process_info_box(node)
|
|
elif ntype == 'UNORDERED_LIST':
|
|
if node.get('items'):
|
|
return process_list(node)
|
|
elif ntype:
|
|
print('** ', ntype)
|
|
return ''
|
|
|
|
|
|
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):
|
|
# open('/t/raw.json', 'w').write(raw)
|
|
body = ''
|
|
try:
|
|
data = json.loads(raw)['props']['pageProps']['cp2Content']
|
|
except Exception:
|
|
data = json.loads(raw)['props']['pageProps']['content']
|
|
body += f'<div style="color: red; font-size:small; font-weight:bold;">{data.get("flyTitle", "")}</div>'
|
|
body += f'<h1>{data["headline"]}</h1>'
|
|
body += f'<div style="font-style: italic; color:#202020;">{data.get("rubric", "")}</div>'
|
|
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.get('dateline') is None:
|
|
body += f'<p style="color: gray; font-size: small;">{dt}</p>'
|
|
else:
|
|
body += f'<p style="color: gray; font-size: small;">{dt + " | " + (data["dateline"])}</p>'
|
|
main_image_url = safe_dict(data, 'leadComponent') or ''
|
|
if main_image_url:
|
|
body += process_node(data['leadComponent'])
|
|
for node in data.get('body'):
|
|
body += process_node(node)
|
|
return '<html><body><article>' + body + '</article></body></html>'
|
|
|
|
|
|
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)
|
|
|
|
|
|
def process_url(url):
|
|
if url.startswith('/'):
|
|
url = 'https://www.economist.com' + url
|
|
return url
|
|
|
|
|
|
class econ_search(BasicNewsRecipe):
|
|
title = 'The Economist - Search'
|
|
language = 'en_GB'
|
|
encoding = 'utf-8'
|
|
__author__ = 'unkn0wn'
|
|
description = (
|
|
'Use the Advanced section of the recipe to search.'
|
|
)
|
|
|
|
remove_attributes = ['data-reactid', 'style', 'height', 'width']
|
|
no_stylesheets = True
|
|
ignore_duplicate_articles = {'url'}
|
|
extra_css = '''
|
|
em { color:#202020; }
|
|
img {display:block; margin:0 auto;}
|
|
'''
|
|
|
|
browser_type = 'webengine'
|
|
|
|
resolve_internal_links = True
|
|
remove_tags = [
|
|
dict(name=['script', 'noscript', 'title', 'iframe', 'cf_floatingcontent', 'aside', 'footer', 'svg']),
|
|
dict(attrs={'aria-label': 'Article Teaser'}),
|
|
dict(attrs={'id':'player'}),
|
|
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()}),
|
|
dict(attrs={'id':lambda x: x and 'gpt-ad-slot' in x}),
|
|
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)]
|
|
# economist.com has started throttling after about 60% of the total has
|
|
# downloaded with connection reset by peer (104) errors.
|
|
delay = 3
|
|
|
|
def get_browser(self, *args, **kwargs):
|
|
kwargs['user_agent'] = (
|
|
'Mozilla/5.0 (Linux; Android 14) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/127.0.6533.103 Mobile Safari/537.36 Liskov'
|
|
)
|
|
br = BasicNewsRecipe.get_browser(self, *args, **kwargs)
|
|
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>' + 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__"]')
|
|
|
|
html = load_article_from_json(script[0].text)
|
|
|
|
root = parse(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 preprocess_html(self, soup):
|
|
width = '600'
|
|
w = self.recipe_specific_options.get('res')
|
|
if w and isinstance(w, str):
|
|
width = w
|
|
for img in soup.findAll('img', src=True):
|
|
qua = 'economist.com/cdn-cgi/image/width=' + width + ',quality=80,format=auto/'
|
|
img['src'] = img['src'].replace('economist.com/', qua)
|
|
return soup
|
|
|
|
recipe_specific_options = {
|
|
'q': {
|
|
'short': 'Text Search',
|
|
'default': 'schools brief'
|
|
},
|
|
's': {
|
|
'short': 'Sort by (date/relevance)',
|
|
'long': 'you can sort by date or relevance',
|
|
'default': 'relevance'
|
|
},
|
|
'p': {
|
|
'short': 'number of pages',
|
|
'long': 'number of pages of search results you want',
|
|
'default': '2'
|
|
},
|
|
'res': {
|
|
'short': 'For hi-res images, select a resolution from the\nfollowing options: 834, 960, 1096, 1280, 1424',
|
|
'long': 'This is useful for non e-ink devices, and for a lower file size\nthan the default, use from 480, 384, 360, 256.',
|
|
'default': '600',
|
|
},
|
|
}
|
|
|
|
def parse_index(self):
|
|
url = 'https://www.economist.com/search?q={query}&sort={sort}&page={page}'
|
|
search = self.recipe_specific_options.get('q')
|
|
sort_type = self.recipe_specific_options.get('s')
|
|
page = self.recipe_specific_options.get('p')
|
|
self.title = 'The Economist - ' + search
|
|
ans = []
|
|
for num in range(1, int(page) + 1):
|
|
ans.extend(self.economist_parse_index(url.format(query=search.replace(' ', '+'), sort=sort_type, page=num)))
|
|
return [('Articles', ans)]
|
|
|
|
def economist_parse_index(self, url):
|
|
self.log('Page ', url.rsplit('=', 1)[-1])
|
|
soup = self.index_to_soup(url)
|
|
results = soup.find('ol', id='search-results')
|
|
if not results:
|
|
self.log('\tPage ', url.rsplit('=', 1)[-1], ' not found')
|
|
return
|
|
for a in results.findAll('a', attrs={'class':'_search-result'}):
|
|
url = a['href']
|
|
title = self.tag_to_string(a.find(attrs={'class':'_headline'}))
|
|
desc = self.tag_to_string(a.find(attrs={'class':'_snippet'}))
|
|
self.log('\t', title, '\n\t', desc, '\n\t\t', url)
|
|
yield {'title': title, 'url': url, 'description': desc}
|
|
|
|
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)
|