def get_soup(site, cssSelector=None):
try:
f = urllib.urlopen(site)
except IOError:
print "error connecting to site " + site
exit()
html = f.read()
f.close()
soup = BeautifulSoup(html)
if not cssSelector:
return soup
return select(soup, cssSelector)
def getLinkHrefs( content, selector ): soup = BeautifulSoup(content) getHrefFromLink = lambda link : link["href"] links = select( soup, selector ) return map( getHrefFromLink, links )
def getLinkHrefs(content, selector):
soup = BeautifulSoup(content)
getHrefFromLink = lambda link: link["href"]
links = select(soup, selector)
return map(getHrefFromLink, links)