Update Ars Technica

Merge branch 'patch-1' of https://github.com/Hainish/calibre
This commit is contained in:
Kovid Goyal 2014-05-02 08:38:47 +05:30
commit fbbdd931bb

View File

@ -48,16 +48,16 @@ class ArsTechnica(BasicNewsRecipe):
]
remove_attributes = ['lang']
feeds = [
(u'Infinite Loop (Apple content)' , u'http://feeds.arstechnica.com/arstechnica/apple/' )
,(u'Opposable Thumbs (Gaming content)' , u'http://feeds.arstechnica.com/arstechnica/gaming/' )
,(u'Gear and Gadgets' , u'http://feeds.arstechnica.com/arstechnica/gadgets/' )
,(u'Uptime (IT content)' , u'http://feeds.arstechnica.com/arstechnica/business/' )
(u'Infinite Loop (Apple content)' , u'http://feeds.arstechnica.com/arstechnica/apple/')
,(u'Opposable Thumbs (Gaming content)' , u'http://feeds.arstechnica.com/arstechnica/gaming/')
,(u'Gear and Gadgets' , u'http://feeds.arstechnica.com/arstechnica/gadgets/')
,(u'Uptime (IT content)' , u'http://feeds.arstechnica.com/arstechnica/business/')
,(u'Open Ended (Open Source content)' , u'http://feeds.arstechnica.com/arstechnica/open-source/')
,(u'One Microsoft Way' , u'http://feeds.arstechnica.com/arstechnica/microsoft/' )
,(u'Scientific method (Science content)' , u'http://feeds.arstechnica.com/arstechnica/science/' )
,(u'One Microsoft Way' , u'http://feeds.arstechnica.com/arstechnica/microsoft/')
,(u'Scientific method (Science content)' , u'http://feeds.arstechnica.com/arstechnica/science/')
,(u'Law & Disorder (Tech policy content)' , u'http://feeds.arstechnica.com/arstechnica/tech-policy/')
,(u'Risk Assessment (Security content)' , u'http://feeds.arstechnica.com/arstechnica/security/')
]
def append_page(self, soup, appendtag, position):
@ -75,7 +75,6 @@ class ArsTechnica(BasicNewsRecipe):
pager.extract()
appendtag.insert(position,texttag)
def preprocess_html(self, soup):
self.append_page(soup, soup.body, 3)
for item in soup.findAll('a'):
@ -91,10 +90,9 @@ class ArsTechnica(BasicNewsRecipe):
str = self.tag_to_string(item)
item.replaceWith(str)
for item in soup.findAll('img'):
if not item.has_key('alt'):
if 'alt' not in item:
item['alt'] = 'image'
return soup
def preprocess_raw_html(self, raw, url):
return '<html><head>'+raw[raw.find('</head>'):]