Backup_Repos/FanFicFare
1
0
Fork 0
mirror of https://github.com/JimmXinu/FanFicFare.git synced 2026-05-05 02:51:48 +02:00
Code Issues Projects Releases Wiki Activity

Remove extra html body from ancient ffnet chapters.

Browse source
This commit is contained in:
Jim Miller 2012-07-30 17:00:13 -05:00
parent 850567afde
commit 42473d4f1d
1 changed files with 9 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

9
fanficdownloader/adapters/adapter_fanfictionnet.py
View file

@ -239,6 +239,15 @@ class FanFictionNetSiteAdapter(BaseSiteAdapter):
## additional to what ever the
## slow_down_sleep_time setting is.
data = self._fetchUrl(url)
# some ancient stories have body tags inside them that cause
# soup parsing to discard the content. For story text we
# don't care about anything before "<div class='storytextp"
# (there's a space after storytextp, so no close quote(')) and
# this kills any body tags.
data = data[data.index("<div class='storytextp"):]
data.replace("<body","<notbody").replace("<BODY","<NOTBODY")
soup = bs.BeautifulSoup(data)
## Remove the 'share' button.