nojhan/weboob-devel
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

[wordreference] dirty fix for a dirty website

Browse source
This commit is contained in:
Julien Veyssier 2013-03-17 14:42:28 +01:00
commit c56f8b2eb4
1 changed files with 5 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

6
modules/wordreference/pages.py
View file

@ -41,7 +41,11 @@ class TranslatePage(BasePage):
# if only one signification is found # if only one signification is found
for div in self.document.getiterator('div'): for div in self.document.getiterator('div'):
if div.attrib.get('class','') == "trans clickable": if div.attrib.get('class','') == "trans clickable":
names = u''+" ".join(div.text_content().split(']')[1].split()[1:]).split(';')[0] if ']' in div.text_content():
tnames = div.text_content().split(']')[1].split()[1:]
else:
tnames = div.text_content().split()[1:]
names = u''+" ".join(tnames).split(';')[0]
names = LAST_THING_IN_PARENTHESIS.sub("",names) names = LAST_THING_IN_PARENTHESIS.sub("",names)
return names.strip() return names.strip()
# another numerotation possibility... # another numerotation possibility...