Spaces:
Build error
Build error
| from flask import Flask, render_template, request, jsonify | |
| from time import sleep | |
| from bs4 import BeautifulSoup | |
| from requests import get | |
| import urllib | |
| app = Flask(__name__) | |
| def _req(term, results, lang, start, proxies, timeout, safe, ssl_verify): | |
| resp = get( | |
| url="https://www.google.com/search", | |
| headers={ | |
| "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/111.0.0.0 Safari/537.36 Edg/111.0.1661.62" | |
| }, | |
| params={ | |
| "q": term, | |
| "num": results + 2, # Prevents multiple requests | |
| "hl": lang, | |
| "start": start, | |
| "safe": safe, | |
| }, | |
| proxies=proxies, | |
| timeout=timeout, | |
| verify=ssl_verify, | |
| ) | |
| resp.raise_for_status() | |
| return resp | |
| class SearchResult: | |
| def __init__(self, url, title, description): | |
| self.url = url | |
| self.title = title | |
| self.description = description | |
| def __repr__(self): | |
| return f"SearchResult(url={self.url}, title={self.title}, description={self.description})" | |
| def to_dict(self): # Add a method to convert to a dictionary | |
| return { | |
| "url": self.url, | |
| "title": self.title, | |
| "description": self.description, | |
| } | |
| def search(term, num_results=10, lang="en", proxy=None, advanced=False, sleep_interval=0, timeout=5, safe="active", ssl_verify=None): | |
| """Search the Google search engine""" | |
| escaped_term = urllib.parse.quote_plus(term) # make 'site:xxx.xxx.xxx ' works. | |
| # Proxy | |
| proxies = None | |
| if proxy: | |
| if proxy.startswith("https"): | |
| proxies = {"https": proxy} | |
| else: | |
| proxies = {"http": proxy} | |
| # Fetch | |
| start = 0 | |
| results = [] | |
| while start < num_results: | |
| # Send request | |
| resp = _req(escaped_term, num_results - start, | |
| lang, start, proxies, timeout, safe, ssl_verify) | |
| # Parse | |
| soup = BeautifulSoup(resp.text, "html.parser") | |
| result_block = soup.find_all("div", attrs={"class": "g"}) | |
| if len(result_block) == 0: | |
| start += 1 | |
| for result in result_block: | |
| # Find link, title, description | |
| link = result.find("a", href=True) | |
| title = result.find("h3") | |
| description_box = result.find( | |
| "div", {"style": "-webkit-line-clamp:2"}) | |
| if description_box: | |
| description = description_box.text | |
| if link and title and description: | |
| start += 1 | |
| if advanced: | |
| results.append(SearchResult(link["href"], title.text, description)) | |
| else: | |
| results.append(link["href"]) | |
| sleep(sleep_interval) | |
| return results | |
| def index(): | |
| return render_template("index.html") | |
| def perform_search(): | |
| search_query = request.args.get("q") | |
| max_results = int(request.args.get("max_results", 10)) # Default to 10 results | |
| search_results = search(term=search_query, num_results=max_results, advanced=True) | |
| # Convert search results to a list of dictionaries | |
| results_json = [result.to_dict() for result in search_results] | |
| return jsonify(results_json) | |
| if __name__ == "__main__": | |
| app.run(debug=True) |