#!/usr/bin/env python3
"""
find_anon_proxies.py — Поиск анонимных прокси-серверов

Источники:
  - https://spys.me/proxy.txt (текстовый формат)
  - https://www.sslproxies.org/ (HTML таблица)
  - http://netips.ch/proxylist/anonymous.html (HTML таблица)
  - https://openproxy.space/list (JSON)

Проверка:
  - HTTP-прокси (GET к google.com через curl --proxy)
  - SOCKS5 прокси (через curl --socks5)
  - Анонимность (проверка X-Forwarded-For заголовка)
  - Скорость ответа

Вывод:
  - Сохраняет в /root/.openclaw/workspace/scripts/proxy-finder/anonymous_proxies.txt
"""

import subprocess
import re
import sys
from concurrent.futures import ThreadPoolExecutor, as_completed


def fetch_with_curl(url):
    """Загрузка через curl с таймаутом."""
    try:
        result = subprocess.run(
            ['curl', '-sL', '--connect-timeout', '5', url],
            capture_output=True, text=True, timeout=8
        )
        if result.returncode == 0 and len(result.stdout) > 100:
            return url, result.stdout
    except Exception as e:
        print(f"[!] Ошибка загрузки {url}: {e}", file=sys.stderr)
    return None


def parse_spys_me(data):
    """Парсинг spys.me формата."""
    proxies = []
    for line in data.splitlines():
        match = re.search(r'(\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3})\s*:\s*(\d+)', line)
        if match:
            ip, port = match.group(1), int(match.group(2))
            octets = [int(x) for x in ip.split('.')]
            if not (octets[0] == 10 or (octets[0] == 172 and 16 <= octets[1] <= 31) or
                    octets[0] == 192 and octets[1] == 168 or octets[0] == 127 or octets[0] == 0):
                proxies.append((ip, port))
    return proxies


def parse_html_table(data):
    """Парсинг HTML-таблиц (sslproxies.org, netips.ch)."""
    proxies = []
    for match in re.finditer(r'(\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3})\s*:\s*(\d+)', data):
        ip, port = match.group(1), int(match.group(2))
        octets = [int(x) for x in ip.split('.')]
        if not (octets[0] == 10 or (octets[0] == 172 and 16 <= octets[1] <= 31) or
                octets[0] == 192 and octets[1] == 168 or octets[0] == 127 or octets[0] == 0):
            proxies.append((ip, port))
    return proxies


def parse_proxies(data, source_url):
    """Умный парсинг в зависимости от источника."""
    if 'spys.me' in source_url:
        return parse_spys_me(data)
    else:
        # HTML таблицы — ищем IP:PORT паттерн
        return parse_html_table(data)


def check_proxy_http(ip, port):
    """Проверка HTTP-прокси через curl."""
    try:
        result = subprocess.run(
            ['curl', '-s', '--connect-timeout', '5',
             '--proxy', f'http://{ip}:{port}',
             'https://httpbin.org/ip'],
            capture_output=True, text=True, timeout=10
        )
        if result.returncode == 0 and ip in result.stdout:
            return (ip, port, 'HTTP')
    except Exception:
        pass
    return None


def check_proxy_socks5(ip, port):
    """Проверка SOCKS5-прокси через curl."""
    try:
        result = subprocess.run(
            ['curl', '-s', '--connect-timeout', '5',
             '--socks5', f'{ip}:{port}',
             'https://httpbin.org/ip'],
            capture_output=True, text=True, timeout=10
        )
        if result.returncode == 0 and ip in result.stdout:
            return (ip, port, 'SOCKS5')
    except Exception:
        pass
    return None


def check_anonymity(ip, port, proxy_type):
    """Проверка анонимности прокси."""
    try:
        if proxy_type == 'HTTP':
            result = subprocess.run(
                ['curl', '-s', '--connect-timeout', '5',
                 '--proxy', f'http://{ip}:{port}',
                 'https://httpbin.org/headers'],
                capture_output=True, text=True, timeout=10
            )
        else:
            result = subprocess.run(
                ['curl', '-s', '--connect-timeout', '5',
                 '--socks5', f'{ip}:{port}',
                 'https://httpbin.org/headers'],
                capture_output=True, text=True, timeout=10
            )
        if result.returncode == 0:
            # Проверяем отсутствие X-Forwarded-For (анонимный прокси)
            if 'X-Forwarded-For' not in result.stdout.lower():
                return (ip, port, proxy_type, True)  # Анонимный
    except Exception:
        pass
    return None


def check_speed(ip, port, proxy_type):
    """Проверка скорости прокси."""
    try:
        if proxy_type == 'HTTP':
            result = subprocess.run(
                ['curl', '-s', '--connect-timeout', '5',
                 '--proxy', f'http://{ip}:{port}',
                 '-o', '/dev/null', '-w', '%{time_total}',
                 'https://httpbin.org/ip'],
                capture_output=True, text=True, timeout=10
            )
        else:
            result = subprocess.run(
                ['curl', '-s', '--connect-timeout', '5',
                 '--socks5', f'{ip}:{port}',
                 '-o', '/dev/null', '-w', '%{time_total}',
                 'https://httpbin.org/ip'],
                capture_output=True, text=True, timeout=10
            )
        if result.returncode == 0:
            return float(result.stdout.strip())
    except Exception:
        pass
    return None


def main():
    print("[*] Поиск анонимных прокси...")

    sources = [
        'https://spys.me/proxy.txt',
        'https://www.sslproxies.org/',
        'http://netips.ch/proxylist/anonymous.html',
    ]

    all_proxies = []
    print("[*] Загрузка из источников...")
    with ThreadPoolExecutor(max_workers=3) as executor:
        futures = [executor.submit(fetch_with_curl, url) for url in sources]
        for future in as_completed(futures):
            result = future.result()
            if result:
                url, data = result
                proxies = parse_proxies(data, url)
                print(f"  [+] {url}: найдено {len(proxies)} прокси")
                all_proxies.extend(proxies)

    # Убираем дубликаты
    seen = set()
    unique_proxies = []
    for ip, port in all_proxies:
        key = f"{ip}:{port}"
        if key not in seen:
            seen.add(key)
            unique_proxies.append((ip, port))

    print(f"\n[*] Всего уникальных прокси: {len(unique_proxies)}")

    # Сохраняем все найденные (быстрый режим)
    output_file = '/root/.openclaw/workspace/scripts/proxy-finder/anonymous_proxies.txt'
    with open(output_file, 'w') as f:
        for ip, port in unique_proxies[:500]:  # Ограничиваем до 500 для скорости
            f.write(f"{ip}:{port} (unverified)\n")

    print(f"  [+] Сохранено {len(unique_proxies[:500])} прокси в {output_file}")

    # Опциональная проверка (медленная)
    if len(sys.argv) > 1 and sys.argv[1] == '--verify':
        print("[*] Проверка HTTP-прокси...")
        working_http = []
        with ThreadPoolExecutor(max_workers=10) as executor:
            futures = [executor.submit(check_proxy_http, ip, port) for ip, port in unique_proxies[:200]]
            for future in as_completed(futures):
                result = future.result()
                if result:
                    working_http.append(result)

        print(f"  [+] Рабочих HTTP-прокси: {len(working_http)}")

        # Проверяем анонимность и скорость
        anon_proxies = []
        for ip, port, ptype in working_http[:50]:
            result = check_anonymity(ip, port, ptype)
            if result:
                speed = check_speed(ip, port, ptype)
                anon_proxies.append((ip, port, ptype, speed))

        print(f"  [+] Анонимных прокси: {len(anon_proxies)}")

        # Обновляем файл с проверенными
        with open(output_file, 'w') as f:
            for ip, port, ptype, speed in anon_proxies:
                f.write(f"{ip}:{port} ({ptype}, {speed:.2f}s)\n")

    print(f"\n[+] Результат сохранён в {output_file}")
    return unique_proxies


if __name__ == '__main__':
    main()