mirror of
https://github.com/kovidgoyal/calibre.git
synced 2025-07-31 14:33:54 -04:00
Update Economist News
Eliminated functions and imports related to HTML parsing and cleanup that are no longer used, simplifying the recipe. Updated article processing to use JSON-based content directly, and removed custom tag and attribute handling for HTML output.
This commit is contained in:
unkn0w7n
parent
8acb5ddb37
commit
cebab6815a
@ -7,10 +7,6 @@ from datetime import datetime, timedelta
|
||||
from urllib.parse import quote, urlencode
|
||||
from uuid import uuid4
|
||||
|
||||
from html5_parser import parse
|
||||
from lxml import etree
|
||||
|
||||
from calibre.ebooks.BeautifulSoup import NavigableString, Tag
|
||||
from calibre.ptempfile import PersistentTemporaryFile
|
||||
from calibre.web.feeds.news import BasicNewsRecipe
|
||||
|
||||
@ -48,14 +44,25 @@ def parse_txt(ty):
|
||||
|
||||
tag_map = {
|
||||
'text': lambda: [ty.get('value', '')],
|
||||
'scaps': lambda: [f'<span style="font-variant: all-small-caps;">{"".join(parse_txt(c))}</span>' for c in children],
|
||||
'scaps': lambda: [
|
||||
f'<span style="font-variant: all-small-caps;">{"".join(parse_txt(c))}</span>'
|
||||
for c in children
|
||||
],
|
||||
'bold': lambda: [f'<b>{"".join(parse_txt(c))}</b>' for c in children],
|
||||
'drop_caps': lambda: [f'<b>{"".join(parse_txt(c))}</b>' for c in children],
|
||||
'italic': lambda: [f'<i>{"".join(parse_txt(c))}</i>' for c in children],
|
||||
'linebreak': lambda: ['<hr>'],
|
||||
'external_link': lambda: [f'<a href="{attr}">{"".join(parse_txt(children[0]))}</a>'] if children else [],
|
||||
'internal_link': lambda: [f'<a href="{attr}">{"".join(parse_txt(children[0]))}</a>'] if children else [],
|
||||
'ufinish': lambda: [text for c in children for text in parse_txt(c)]
|
||||
'external_link': lambda: [
|
||||
f'<a href="{attr}">{"".join(parse_txt(children[0]))}</a>'
|
||||
]
|
||||
if children
|
||||
else [],
|
||||
'internal_link': lambda: [
|
||||
f'<a href="{attr}">{"".join(parse_txt(children[0]))}</a>'
|
||||
]
|
||||
if children
|
||||
else [],
|
||||
'ufinish': lambda: [text for c in children for text in parse_txt(c)],
|
||||
}
|
||||
|
||||
if typ in tag_map:
|
||||
@ -76,10 +83,10 @@ def process_web_node(node):
|
||||
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'\n<p>{node.get("textHtml")}</p>'
|
||||
elif node.get('textJson'):
|
||||
return f'<p>{parse_textjson(node["textJson"])}</p>'
|
||||
return f'<p>{node.get("text", "")}</p>'
|
||||
return f'\n<p>{parse_textjson(node["textJson"])}</p>'
|
||||
return f'\n<p>{node.get("text", "")}</p>'
|
||||
elif (ntype == 'IMAGE') or (node.get('__typename', '') == 'ImageComponent'):
|
||||
alt = '' if node.get('altText') is None else node.get('altText')
|
||||
cap = ''
|
||||
@ -141,33 +148,6 @@ def load_article_from_web_json(raw):
|
||||
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)
|
||||
|
||||
|
||||
class NoArticles(Exception):
|
||||
pass
|
||||
|
||||
@ -185,7 +165,7 @@ def get_content(url_):
|
||||
'x-economist-consumer': 'TheEconomist-Liskov-android',
|
||||
'x-teg-client-name': 'Economist-Android',
|
||||
'x-teg-client-os': 'Android',
|
||||
'x-teg-client-version': '4.40.0'
|
||||
'x-teg-client-version': '4.40.0',
|
||||
}
|
||||
br = browser()
|
||||
req = Request(
|
||||
@ -221,35 +201,11 @@ class EconomistNews(BasicNewsRecipe):
|
||||
cover_url = 'https://m.media-amazon.com/images/M/MV5BNzJiZGYzNzgtNWY5Yi00NWYyLThmZGUtODQyM2ZkOWVlMDI1XkEyXkFqcGc@.jpg'
|
||||
oldest_article = 15
|
||||
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)]
|
||||
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
|
||||
|
||||
remove_empty_feeds = True
|
||||
ignore_duplicate_articles = {'title'}
|
||||
|
||||
@ -261,7 +217,7 @@ class EconomistNews(BasicNewsRecipe):
|
||||
'days': {
|
||||
'short': 'Oldest article to download from this news source. In days ',
|
||||
'long': 'For example, 0.5, gives you articles from the past 12 hours',
|
||||
'default': str(oldest_article)
|
||||
'default': str(oldest_article),
|
||||
},
|
||||
'res': {
|
||||
'short': 'For hi-res images, select a resolution from the\nfollowing options: 834, 960, 1096, 1280, 1424',
|
||||
@ -303,7 +259,9 @@ class EconomistNews(BasicNewsRecipe):
|
||||
'variables': '{"homepageType":"MOBILE"}',
|
||||
'query': 'query FindHomepage($homepageType: HomepageType!) { findHomepage(homepageType: $homepageType) { __typename ...HomepageFragment } } fragment CtaFragment on Cta { link text } fragment ContentIdentityFragment on ContentIdentity { articleType forceAppWebView leadMediaType } fragment ImageTeaserFragment on ImageComponent { altText height imageType source url width } fragment NarrationFragment on Narration { album bitrate duration filename id provider url } fragment PodcastAudioFragment on PodcastEpisode { id audio { url durationInSeconds } } fragment ArticleTeaserFragment on Article { id tegId url rubric headline flyTitle brand byline dateFirstPublished dateline dateModified datePublished dateRevised estimatedReadTime wordCount printHeadline contentIdentity { __typename ...ContentIdentityFragment } section { tegId name } teaserImage { __typename type ...ImageTeaserFragment } leadComponent { __typename type ...ImageTeaserFragment } narration { __typename ...NarrationFragment } podcast { __typename ...PodcastAudioFragment } } fragment OverridesFragment on Overrides { flyTitle headline rubric teaserImage { __typename ...ImageTeaserFragment } } fragment CollectionItemFragment on CollectionItem { __typename type ... on CollectionArticleItem { article { __typename ...ArticleTeaserFragment } overrides { __typename ...OverridesFragment } } ... on CollectionRelatedArticleItem { article { __typename ...ArticleTeaserFragment } overrides { __typename ...OverridesFragment } } ... on CollectionExternalLinkItem { url overrides { __typename ...OverridesFragment } } } fragment HomepageFragment on Homepage { components { __typename id headline type ... on StandardCollection { cta { __typename ...CtaFragment } items { __typename ...CollectionItemFragment } } ... on TakeoverCollection { cta { __typename ...CtaFragment } items { __typename ...CollectionItemFragment } } ... on DiscoverRailCollection { cta { __typename ...CtaFragment } items { __typename ...CollectionItemFragment } } ... on TopStoriesCollection { cta { __typename ...CtaFragment } items { __typename ...CollectionItemFragment } } ... on EmbedsCollection { cta { __typename ...CtaFragment } items { __typename ...CollectionItemFragment } } ... on CarouselCollection { cta { __typename ...CtaFragment } items { __typename ...CollectionItemFragment } imageLayout variant } ... on VideoCarouselCollection { cta { __typename ...CtaFragment } playlistId source fallbackStoryLink { sourceId } } ... on CoverPackageCollection { cta { __typename ...CtaFragment } items { __typename ...CollectionItemFragment } } ... on LatestEditionCollection { cta { __typename ...CtaFragment } items { __typename ...CollectionItemFragment } } ... on Newsletter { slug items { __typename ...CollectionItemFragment } } } }', # noqa: E501
|
||||
}
|
||||
url = 'https://cp2-graphql-gateway.p.aws.economist.com/graphql?' + urlencode(query, safe='()!', quote_via=quote)
|
||||
url = 'https://cp2-graphql-gateway.p.aws.economist.com/graphql?' + urlencode(
|
||||
query, safe='()!', quote_via=quote
|
||||
)
|
||||
raw = get_content(url)
|
||||
ans = self.economist_parse_index(raw)
|
||||
return self.economist_return_index(ans)
|
||||
@ -347,43 +305,15 @@ class EconomistNews(BasicNewsRecipe):
|
||||
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/'
|
||||
qua = (
|
||||
'economist.com/cdn-cgi/image/width=' + width + ',quality=80,format=auto/'
|
||||
)
|
||||
img['src'] = img['src'].replace('economist.com/', qua)
|
||||
return soup
|
||||
|
||||
def preprocess_raw_html(self, raw, url):
|
||||
# open('/t/raw.html', 'wb').write(raw.encode('utf-8'))
|
||||
html = load_article_from_web_json(raw)
|
||||
|
||||
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
|
||||
return load_article_from_web_json(raw)
|
||||
|
||||
def get_article(self, url):
|
||||
query = {
|
||||
@ -391,7 +321,9 @@ class EconomistNews(BasicNewsRecipe):
|
||||
'variables': '{{"ref":"{}"}}'.format(url),
|
||||
'query': 'query ArticleDeeplinkQuery($ref: String!, $includeRelatedArticles: Boolean = true ) { findArticleByUrl(url: $ref) { __typename ...ArticleDataFragment } } fragment ContentIdentityFragment on ContentIdentity { articleType forceAppWebView leadMediaType } fragment NarrationFragment on Narration { album bitrate duration filename id provider url isAiGenerated fileHash } fragment ImageTeaserFragment on ImageComponent { altText height imageType source url width } fragment PodcastAudioFragment on PodcastEpisode { id audio { url durationInSeconds } } fragment ArticleTeaserFragment on Article { id tegId url rubric headline flyTitle brand byline dateFirstPublished dateline dateModified datePublished dateRevised estimatedReadTime wordCount printHeadline contentIdentity { __typename ...ContentIdentityFragment } section { tegId name } teaserImage { __typename type ...ImageTeaserFragment } leadComponent { __typename type ...ImageTeaserFragment } narration(selectionMethod: PREFER_ACTOR_NARRATION) { __typename ...NarrationFragment } podcast { __typename ...PodcastAudioFragment } } fragment AnnotatedTextFragment on AnnotatedText { text textJson annotations { type length index attributes { name value } } } fragment ImageComponentFragment on ImageComponent { altText caption { __typename ...AnnotatedTextFragment } credit height imageType mode source url width } fragment BlockQuoteComponentFragment on BlockQuoteComponent { text textJson annotations { type length index attributes { name value } } } fragment BookInfoComponentFragment on BookInfoComponent { text textJson annotations { type length index attributes { name value } } } fragment ParagraphComponentFragment on ParagraphComponent { text textJson annotations { type length index attributes { name value } } } fragment PullQuoteComponentFragment on PullQuoteComponent { text textJson annotations { type length index attributes { name value } } } fragment CrossheadComponentFragment on CrossheadComponent { text } fragment OrderedListComponentFragment on OrderedListComponent { items { __typename ...AnnotatedTextFragment } } fragment UnorderedListComponentFragment on UnorderedListComponent { items { __typename ...AnnotatedTextFragment } } fragment VideoComponentFragment on VideoComponent { url title thumbnailImage } fragment InfoboxComponentFragment on InfoboxComponent { components { __typename type ...BlockQuoteComponentFragment ...BookInfoComponentFragment ...ParagraphComponentFragment ...PullQuoteComponentFragment ...CrossheadComponentFragment ...OrderedListComponentFragment ...UnorderedListComponentFragment ...VideoComponentFragment } } fragment InfographicComponentFragment on InfographicComponent { url title width fallback { __typename ...ImageComponentFragment } altText height width } fragment ArticleDataFragment on Article { id url brand byline rubric headline layout { headerStyle } contentIdentity { __typename ...ContentIdentityFragment } dateline dateFirstPublished dateModified datePublished dateRevised estimatedReadTime narration(selectionMethod: PREFER_ACTOR_NARRATION) { __typename ...NarrationFragment } printFlyTitle printHeadline printRubric flyTitle wordCount section { tegId name articles(pagingInfo: { pagingType: OFFSET pageSize: 6 pageNumber: 1 } ) @include(if: $includeRelatedArticles) { edges { node { __typename ...ArticleTeaserFragment } } } } teaserImage { __typename type ...ImageComponentFragment } tegId leadComponent { __typename type ...ImageComponentFragment } body { __typename type ...BlockQuoteComponentFragment ...BookInfoComponentFragment ...ParagraphComponentFragment ...PullQuoteComponentFragment ...CrossheadComponentFragment ...OrderedListComponentFragment ...UnorderedListComponentFragment ...InfoboxComponentFragment ...ImageComponentFragment ...VideoComponentFragment ...InfographicComponentFragment } footer { __typename type ...ParagraphComponentFragment } tags { name } ads { adData } podcast { __typename ...PodcastAudioFragment } }', # noqa: E501
|
||||
}
|
||||
deep_url = 'https://cp2-graphql-gateway.p.aws.economist.com/graphql?' + urlencode(query, safe='()!', quote_via=quote)
|
||||
deep_url = 'https://cp2-graphql-gateway.p.aws.economist.com/graphql?' + urlencode(
|
||||
query, safe='()!', quote_via=quote
|
||||
)
|
||||
raw = get_content(deep_url)
|
||||
return raw
|
||||
|
||||
@ -401,31 +333,3 @@ class EconomistNews(BasicNewsRecipe):
|
||||
pt.write(art_cont)
|
||||
pt.close()
|
||||
return 'file:///' + pt.name
|
||||
|
||||
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)
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user