פרוקסי אמריקאי: רוטציה וסשנים קבועים על כתובות ביתיות

# pip install "requests[socks]"
import time, statistics, requests
from collections import Counter
from concurrent.futures import ThreadPoolExecutor

LINE = "host:port:login:password"
PROTOCOL = "http"
TARGET_URL = "https://www.example.com/"
REQUESTS = 50
THREADS = 10
HEADERS = {
    "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 "
                  "(KHTML, like Gecko) Chrome/130.0.0.0 Safari/537.36",
    "Accept-Language": "en-US,en;q=0.9",
}
CAPTCHA_MARKERS = ("captcha", "/sorry/", "robot check", "are you a human", "px-block", "cf-chl")
BROWSER_MARKERS = ("/httpservice/retry/enablejs", "enable javascript", "please enable js", "challenge-platform")

host, port, login, password = LINE.split(":")
scheme = "socks5h" if PROTOCOL == "socks5" else "http"
proxy_url = f"{scheme}://{login}:{password}@{host}:{port}"
PROXIES = {"http": proxy_url, "https": proxy_url}

def visit(_):
    start = time.monotonic()
    try:
        r = requests.get(TARGET_URL, headers=HEADERS, proxies=PROXIES, timeout=30)
    except requests.RequestException as e:
        return f"error {type(e).__name__}", time.monotonic() - start
    page = (r.url + r.text[:300000]).lower()
    if any(marker in page for marker in CAPTCHA_MARKERS):
        return "captcha", time.monotonic() - start
    if any(marker in page for marker in BROWSER_MARKERS):
        return "needs browser", time.monotonic() - start
    if r.status_code in (403, 429):
        return f"blocked {r.status_code}", time.monotonic() - start
    return ("ok" if r.status_code == 200 else f"http {r.status_code}"), time.monotonic() - start

with ThreadPoolExecutor(THREADS) as executor:
    results = list(executor.map(visit, range(REQUESTS)))

verdicts = Counter(verdict for verdict, _ in results)
ok_times = [t for verdict, t in results if verdict == "ok"]
print(f"{TARGET_URL} through {REQUESTS} requests")
for verdict, n in verdicts.most_common():
    print(f"{n:>5}  {n / REQUESTS:6.1%}  {verdict}")
if ok_times:
    print(f"clean pages: median {statistics.median(ok_times):.2f} s")