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

factorise

Browse source
This commit is contained in:
Juke 2011-03-07 09:44:02 +01:00 • committed by Romain Bignon
commit 397b5875f1
9 changed files with 59 additions and 54 deletions
Download patch file Download diff file Expand all files Collapse all files

16
weboob/backends/inrocks/pages/article.py
View file

@ -17,25 +17,25 @@
# Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
from weboob.tools.parsers.lxmlparser import select
from weboob.tools.genericArticle import GenericNewsPage, try_remove
from weboob.tools.genericArticle import GenericNewsPage, try_remove, \
try_remove_from_selector_list
class ArticlePage(GenericNewsPage):
"ArticlePage object for inrocks"
def on_loaded(self):
self.main_div = self.document.getroot()
self.element_title_selector = "h1"
self.element_author_selector = "div.name>span"
self.element_title_selector = "h1"
self.element_body_selector = "div.maincol"
def get_body(self):
element_body = self.get_element_body()
try_remove(element_body, "div.sidebar")
details = select(element_body, "div.details", 1)
try_remove(details, "div.footer")
header = select(element_body, "div.header", 1)
for selector in ["h1", "div.picture", "div.date", "div.news-single-img",
"div.metas_img", "strong"]:
try_remove(header, selector)
try_remove(select(element_body, "div.details", 1), "div.footer")
try_remove_from_selector_list(select(element_body, "div.header", 1),
["h1", "div.picture", "div.date",
"div.news-single-img",
"div.metas_img", "strong"])
return self.browser.parser.tostring(element_body)