martyn/ghostgle
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Remove sone google classes.

Browse source
This commit is contained in:
Martyn 2026-07-29 00:14:27 +02:00
parent 43dae28a36
commit 845fba5f32
1 changed files with 5 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

6
main.py
View file

@ -97,7 +97,9 @@ def googlepage_to_cleanhtml(googlehtml,searchstring="",currentpage=1):
postamble = "</body></html>" postamble = "</body></html>"
results = "" results = ""
for h3 in searchresults.select("h3"): for h3 in searchresults.select("h3"):
preview = str(h3.parent.parent.parent.parent.parent.next_sibling.div) preview = h3.parent.parent.parent.parent.parent.next_sibling.div
del preview["class"]
del preview["style"]
heading = h3.string heading = h3.string
cite = h3.parent.cite cite = h3.parent.cite
del cite["class"] del cite["class"]
@ -115,7 +117,7 @@ def googlepage_to_cleanhtml(googlehtml,searchstring="",currentpage=1):
url = geturlfromdata(h3.parent['data-sb']) url = geturlfromdata(h3.parent['data-sb'])
except: except:
url = h3.parent['href'] url = h3.parent['href']
results = results+imagetag+"<div class=\"source\">"+str(source)+"<br/>"+str(cite)+"</div><h3><a href=\""+url+"\">"+heading+"</a></h3><p>"+preview+"</p>" results = results+imagetag+"<div class=\"source\">"+str(source)+"<br/>"+str(cite)+"</div><h3><a href=\""+url+"\">"+heading+"</a></h3><p>"+str(preview)+"</p>"
navigation = soup.find("td", attrs={"role":"heading"}).parent navigation = soup.find("td", attrs={"role":"heading"}).parent
pagetds = navigation.find_all("td") pagetds = navigation.find_all("td")
del pagetds[0] del pagetds[0]