calibre/recipes/new_yorker.recipe
2017-06-17 01:07:30 +05:30

162 lines
5.6 KiB
Python

#!/usr/bin/env python2
# -*- coding: utf-8 -*-
__license__ = 'GPL v3'
import json
from urllib import unquote
from collections import defaultdict
from calibre.web.feeds.news import BasicNewsRecipe
from calibre import browser
def absurl(x):
if x.startswith('/') and not x.startswith('//'):
x = 'https://www.newyorker.com' + x
return x
class Tag(list):
def __init__(self, name, **attrs):
self.name = name
self.attrs = attrs
def __str__(self):
ans = ['<' + self.name]
for k, v in self.attrs.iteritems():
ans.append(' {}="{}"'.format(k, v))
ans.append('>')
for child in self:
ans.append(unicode(child))
ans.append('</{}>'.format(self.name))
return ''.join(ans)
def deserialize(node):
name = node.pop(0)
if name == 'inline-embed':
meta = node.pop(0)
t = meta['type']
if t in ('image', 'cartoon'):
meta = json.loads(unquote(meta['meta']))
ans = Tag('img', src=absurl(meta['url']))
elif t == 'section':
ans = Tag('div')
else:
ans = Tag('span')
else:
ans = Tag(name)
for child in node:
if isinstance(child, list):
ans.append(deserialize(child))
elif isinstance(child, basestring):
ans.append(child)
return ans
class NewYorker(BasicNewsRecipe):
title = u'New Yorker Magazine'
description = u'Content from the New Yorker website'
url_list = []
language = 'en'
__author__ = 'Kovid Goyal'
no_stylesheets = True
timefmt = ' [%b %d]'
encoding = 'utf-8'
extra_css = '''
.byline { font-size:xx-small; font-weight: bold;}
h3 { margin-bottom: 6px; }
.caption { font-size: xx-small; font-style: italic; font-weight: normal; }
'''
needs_subscription = 'optional'
keep_only_tags = [
dict(attrs={'class':lambda x: x and 'ArticleHeader__hed___' in x}),
dict(attrs={'class':lambda x: x and 'ArticleHeader__dek___' in x}),
dict(attrs={'class':lambda x: x and 'Byline__articleHeader___' in x}),
dict(attrs={'class':lambda x: x and 'ArticleLedeImage__container___' in x}),
dict(itemprop=['headline', 'alternativeHeadline']),
dict(name='h1'),
dict(attrs={'class':lambda x: x and 'byline-and-date' in x}),
dict(attrs={'class':lambda x: x and 'inset-mobile-crop-image' in x}),
dict(attrs={'class':lambda x: x and 'hero-image-caption' in x}),
dict(id='articleBody'),
dict(attrs={'class':lambda x: x and 'ArticleDisclaimer__articleDisclaimer___' in x}),
dict(attrs={'class':lambda x: x and 'ArticleContributors__bio___' in x}),
]
remove_tags = [
dict(attrs={'class': lambda x: x and set(x.split()).intersection(
{'content-ad-wrapper', 'social-hover', 'background-image'})}),
dict(id=['newsletter-signup']),
dict(name='meta links source'.split()),
]
# def preprocess_raw_html(self, raw, url):
# import re
# try:
# raw = re.search(r'window.__TNY__.INITIAL_STATE = ({.+?)</script', raw).group(1)
# except AttributeError:
# return raw
# data = json.loads(raw.strip().rstrip(';'))
# return '<html><body><div id="articleBody">' + unicode(deserialize(data['primary']['body']))
#
def parse_index(self):
soup = self.index_to_soup(
'https://www.newyorker.com/magazine?intcid=magazine')
# soup = self.index_to_soup('file:///t/raw.html')
cover_img = soup.find(attrs={'class': lambda x: x and 'MagazineCover__cover___' in x})
if cover_img is not None:
cover_img = cover_img.find('img')
if cover_img is not None:
self.cover_url = cover_img.get('src', cover_img.get('data-src', cover_img.get('srcset').split()[0]))
self.log('Found cover:', self.cover_url)
stories = defaultdict(list)
last_section = 'Unknown'
for story in soup.findAll(attrs={'class': lambda x: x and 'River__riverItemContent___' in x}):
try:
section = self.tag_to_string(story.find('a')['title']) or last_section
except KeyError:
section = last_section
last_section = section
a = story.find('h4').find('a')
title = a.contents[1]
url = absurl(a['href'])
desc = ''
body = story.find(attrs={'class': 'River__dek___CayIg'})
if body is not None:
desc = body.contents[0]
self.log('Found article:', title)
self.log('\t' + url)
self.log('\t' + desc)
self.log('')
stories[section].append({'title':title, 'url':url, 'description':desc})
return [(k, stories[k]) for k in sorted(stories)]
def preprocess_html(self, soup):
for attr in 'srcset data-src-mobile'.split():
for img in soup.findAll('img'):
try:
ds = img[attr].split()[0]
del img[attr]
except KeyError:
continue
if ds:
img['src'] = ds
return soup
# The New Yorker changes the content it delivers based on cookies, so the
# following ensures that we send no cookies
def get_browser(self, *args, **kwargs):
return self
def clone_browser(self, *args, **kwargs):
return self.get_browser()
def open_novisit(self, *args, **kwargs):
br = browser()
return br.open_novisit(*args, **kwargs)
open = open_novisit