Update internet.py

This commit is contained in:
H Lohaus 2024-11-15 18:25:54 +01:00 committed by GitHub
parent 0f81b61739
commit 9e4a63e74b
No known key found for this signature in database
GPG key ID: B5690EEEBB952194

View file

@ -46,8 +46,6 @@ class SearchResultEntry():
def scrape_text(html: str, max_words: int = None) -> str: def scrape_text(html: str, max_words: int = None) -> str:
soup = BeautifulSoup(html, "html.parser") soup = BeautifulSoup(html, "html.parser")
for exclude in soup(["script", "style"]):
exclude.extract()
for selector in [ for selector in [
"main", "main",
".main-content-wrapper", ".main-content-wrapper",