Tweak XKCD recipe: Add a masthead image, put the strip title in an h1 tag

This commit is contained in:
NiLuJe 2012-09-14 17:47:38 +02:00
parent 87fc87978b
commit a7cab66b6f

View File

@ -2,6 +2,8 @@ __license__ = 'GPL v3'
__copyright__ = '2008, Kovid Goyal <kovid at kovidgoyal.net>'
'''
Changelog:
2012-04-06
Fixed empty articles, added masthead img (NiLuJe)
2011-09-24
Changed cover (drMerry)
'''
@ -13,7 +15,8 @@ import time, re
from calibre.web.feeds.news import BasicNewsRecipe
class XkcdCom(BasicNewsRecipe):
cover_url = 'http://imgs.xkcd.com/s/9be30a7.png'
cover_url = 'http://imgs.xkcd.com/static/terrible_small_logo.png'
masthead_url = 'http://imgs.xkcd.com/static/terrible_small_logo.png'
title = 'xkcd'
description = 'A webcomic of romance and math humor.'
__author__ = 'Martin Pitt updated by DrMerry.'
@ -21,13 +24,14 @@ class XkcdCom(BasicNewsRecipe):
use_embedded_content = False
oldest_article = 60
keep_only_tags = [dict(id='middleContainer')]
remove_tags = [dict(name='ul'), dict(name='h3'), dict(name='br')]
#keep_only_tags = [dict(id='middleContainer')]
#remove_tags = [dict(name='ul'), dict(name='h3'), dict(name='br')]
keep_only_tags = [dict(id='comic')]
no_stylesheets = True
# turn image bubblehelp into a paragraph
# turn image bubblehelp into a paragraph, and put alt in a heading
preprocess_regexps = [
(re.compile(r'(<img.*title=")([^"]+)(".*>)'),
lambda m: '%s%s<p>%s</p>' % (m.group(1), m.group(3), m.group(2)))
(re.compile(r'(<img.*title=")([^"]+)(".alt=")([^"]+)(".*>)'),
lambda m: '<h1>%s</h1>%s%s%s<p>%s</p>' % (m.group(4), m.group(1), m.group(3), m.group(5), m.group(2)))
]
def parse_index(self):