Chewing_Bever/didier
0
0
Fork 0
mirror of https://github.com/stijndcl/didier.git synced 2026-04-07 15:48:29 +02:00
Code Issues Projects Releases Wiki Activity

Add comments

Browse source
This commit is contained in:
Stijn De Clercq 2021-01-24 13:14:44 +01:00
parent 83f28d9e21
commit 652238ce55
2 changed files with 10 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

5
functions/scraping.py
View file

@ -24,8 +24,12 @@ def google_search(query):
bs = BeautifulSoup(resp.text, "html.parser")
def getContent(element):
"""
Function to find links & titles in the HTML of a <div> element
"""
link = element.find("a", href=True)
title = element.find("h3")
if link is None or title is None:
return None
@ -35,6 +39,7 @@ def google_search(query):
return None
return link["href"], sp.text
divs = bs.find_all("div", attrs={"class": "g"})
return list(getContent(d) for d in divs), 200