return report
report = build_report(raw_url) print(json.dumps(report, indent=2, ensure_ascii=False)) xnexx hot
import requests from bs4 import BeautifulSoup return report report = build_report(raw_url) print(json
title = soup.title.string.strip() if soup.title and soup.title.string else "" meta_desc = extract_meta(soup, "description") og_title = extract_meta(soup, "og:title") og_desc = extract_meta(soup, "og:description") keywords = extract_meta(soup, "keywords") "description") og_title = extract_meta(soup
def extract_meta(soup: BeautifulSoup, name: str) -> str: """Return the content of a meta tag (name or property).""" tag = soup.find("meta", attrs="name": name) or soup.find( "meta", attrs="property": name ) return (tag["content"] if tag and tag.has_attr("content") else "")
# Gather all visible text for a quick adult‑content heuristic visible_text = " ".join( s.get_text(separator=" ", strip=True) for s in soup.find_all(string=True) if s.parent.name not in "script", "style", "noscript" ) adult_flag = is_adult_content(visible_text)
report = "url": url, "final_url": resp.url, "status_code": resp.status_code, "title": title, "meta_description": meta_desc, "og_title": og_title, "og_description": og_desc, "keywords": keywords, "is_adult_content": adult_flag, "content_length_bytes": len(resp.content),