Switch portkey.org to non-stone soup and brute-force remove <HTML> tags.

This commit is contained in:
Jim Miller 2013-09-12 22:40:41 -05:00
parent 9f56952950
commit aacf6dc6a2
2 changed files with 5 additions and 3 deletions

View file

@ -251,8 +251,10 @@ class PortkeyOrgAdapter(BaseSiteAdapter): # XXX
logger.debug('Getting chapter text from: %s' % url)
data = self._fetchUrl(url)
soup = bs.BeautifulStoneSoup(data,
selfClosingTags=('br','hr')) # otherwise soup eats the br/hr tags.
# data = data.replace("HTML>","div>")
soup = bs.BeautifulSoup(data)
#print("soup:%s"%soup)
tag = soup.find('td', {'class' : 'story'})

View file

@ -78,7 +78,7 @@
If you have any problems with this application, please
report them in
the <a href="http://groups.google.com/group/fanfic-downloader">FanFictionDownLoader Google Group</a>. The
<a href="http://4-4-67.fanfictiondownloader.appspot.com">Previous Version</a> is also available for you to use if necessary.
<a href="http://4-4-72.fanfictiondownloader.appspot.com">Previous Version</a> is also available for you to use if necessary.
</p>
<div id='error'>
{{ error_message }}