""" Browser service с поддержкой SOCKS5 через локальный HTTP туннель. """ import asyncio import random import logging from pathlib import Path from datetime import datetime from typing import Optional from dataclasses import dataclass from camoufox import DefaultAddons from camoufox.async_api import AsyncCamoufox from config.settings import settings from managers.proxy_manager import ProxyManager, Proxy, ProxyType from services.socks5_to_http_proxy import Socks5ProxyPool logger = logging.getLogger(__name__) def _patch_camoufox(): """Патчит camoufox.utils.generate_fingerprint чтобы убирать screen constraint при ошибке.""" import camoufox.utils as _cu _orig = _cu.generate_fingerprint def _patched(**config): try: return _orig(**config) except ValueError: config.pop("screen", None) try: return _orig(**config) except ValueError: minimal = {k: v for k, v in config.items() if k == "locale"} return _orig(**minimal) _cu.generate_fingerprint = _patched _patch_camoufox() def _locale_for_ip(ip: str) -> str: """Возвращает BCP47 locale без script-тега по реальному IP прокси.""" try: from camoufox.utils import get_geolocation geo = get_geolocation(ip) lang = geo.locale.language region = geo.locale.region if lang and region: return f"{lang}-{region}" except Exception: pass return "en-US" @dataclass class VisitResult: """Результат посещения страницы.""" url: str initial_url: Optional[str] = None final_url: Optional[str] = None redirect_occurred: bool = False proxy_info: str = "direct" reading_time: int = 0 screenshot_path: Optional[str] = None success: bool = False error: Optional[str] = None class BrowserService: """ Сервис для посещения страниц через Camoufox. Поддерживает HTTP, SOCKS5 (через туннель) и прямое соединение. """ def __init__(self, proxy_manager: ProxyManager): self.proxy_manager = proxy_manager self.socks5_pool = Socks5ProxyPool(idle_timeout=300) self.screenshots_dir = Path(settings.SCREENSHOTS_DIR) self.screenshots_dir.mkdir(exist_ok=True) async def visit_page( self, url: str, reading_time: Optional[int] = None ) -> VisitResult: """Посещает страницу.""" if reading_time is None: reading_time = random.randint( settings.DEFAULT_MIN_READING, settings.DEFAULT_MAX_READING ) for attempt in range(3): proxy = None use_socks5 = False try: if self.proxy_manager.has_proxies(): proxy = await self.proxy_manager.acquire_proxy(timeout=30) socks5_locale = None if proxy and proxy.proxy_type in [ProxyType.SOCKS5, ProxyType.SOCKS4]: proxy_config = await self.socks5_pool.get_proxy_config(proxy) use_socks5 = True # При SOCKS5 camoufox видит 127.0.0.1 → geoip берёт IP сервера → ru-Cyrl-RU # Делаем lookup по реальному IP прокси сами и передаём чистый locale real_ip = proxy.id.split("://")[-1].split(":")[0] socks5_locale = _locale_for_ip(real_ip) logger.info(f"Using SOCKS5 tunnel for: {proxy.id} locale={socks5_locale}") elif proxy: proxy_config = proxy.proxy_config logger.info(f"Using HTTP proxy: {proxy.id}") else: proxy_config = None logger.info("Direct connection") if proxy_config: camoufox_kwargs = dict( headless=True, geoip=True, humanize=True, exclude_addons=[DefaultAddons.UBO], proxy=proxy_config, ) if socks5_locale: camoufox_kwargs["locale"] = socks5_locale async with AsyncCamoufox(**camoufox_kwargs) as browser: result = await self._browse_page( browser, url, proxy.id if proxy else "direct", reading_time ) else: async with AsyncCamoufox( headless=True, geoip=False, humanize=True, locale="ru-RU", exclude_addons=[DefaultAddons.UBO] ) as browser: result = await self._browse_page( browser, url, "direct", reading_time ) if proxy: await self.proxy_manager.release_proxy(proxy, result.success) if use_socks5: await self.socks5_pool.release(proxy) return result except ValueError as e: logger.warning(f"Visit attempt {attempt + 1} fingerprint error, retrying with new proxy: {e}") if proxy: await self.proxy_manager.release_proxy(proxy, success=False) if use_socks5: await self.socks5_pool.release(proxy) except Exception as e: logger.error(f"Visit error: {e}", exc_info=True) if proxy: await self.proxy_manager.release_proxy(proxy, success=False) if use_socks5: await self.socks5_pool.release(proxy) return VisitResult(url=url, success=False, error=str(e)) return VisitResult(url=url, success=False, error="All attempts failed (fingerprint error)") async def _browse_page( self, browser, url: str, proxy_info: str, reading_time: int ) -> VisitResult: """Выполняет просмотр страницы.""" page = await browser.new_page() initial_url = url final_url = url redirect = False reading_completed = False try: await page.set_viewport_size({ "width": settings.VIEWPORT_WIDTH, "height": settings.VIEWPORT_HEIGHT }) goto_task = asyncio.create_task( page.goto(url, wait_until="commit", timeout=30000) ) await self._move_mouse(page, goto_task) await goto_task initial_url = page.url try: final_url, redirect = await self._wait_redirect(page, initial_url) except Exception as e: logger.warning(f"Redirect wait error (ignored): {e}") final_url = page.url await self._simulate_reading(page, reading_time) reading_completed = True screenshot_path = None if settings.SCREENSHOTS_DIR: try: screenshot_path = await self._take_screenshot(page, final_url) except Exception: pass return VisitResult( url=url, initial_url=initial_url, final_url=final_url, redirect_occurred=redirect, proxy_info=proxy_info, reading_time=reading_time, screenshot_path=screenshot_path, success=True ) except asyncio.CancelledError: raise except Exception as e: return VisitResult( url=url, initial_url=initial_url, final_url=final_url, redirect_occurred=redirect, proxy_info=proxy_info, reading_time=reading_time, success=reading_completed, error=str(e) if not reading_completed else None ) finally: await page.close() async def _move_mouse(self, page, task): """Двигает мышь.""" w, h = settings.VIEWPORT_WIDTH, settings.VIEWPORT_HEIGHT try: while not task.done(): await page.mouse.move( random.randint(100, w - 100), random.randint(100, h - 100) ) await asyncio.sleep(random.uniform(0.1, 0.3)) except Exception: pass async def _wait_redirect(self, page, initial_url): """Ждет редирект.""" loop = asyncio.get_running_loop() start = loop.time() redirect = False current = initial_url while loop.time() - start < settings.MAX_REDIRECT_WAIT: await asyncio.sleep(2) new_url = page.url if new_url != current: current = new_url redirect = True task = asyncio.create_task( page.wait_for_load_state("load", timeout=30000) ) await self._move_mouse(page, task) await task break await self._random_action(page) return page.url, redirect async def _simulate_reading(self, page, duration): """Симулирует чтение.""" loop = asyncio.get_running_loop() start = loop.time() while loop.time() - start < duration: await self._random_action(page) async def _random_action(self, page): """Случайное действие.""" try: action = random.choice(['move', 'scroll', 'pause']) if action == 'move': x = random.randint(100, settings.VIEWPORT_WIDTH - 100) y = random.randint(100, settings.VIEWPORT_HEIGHT - 100) await page.mouse.move(x, y) await asyncio.sleep(random.uniform(0.3, 1.0)) elif action == 'scroll': await page.evaluate(f"window.scrollBy(0, {random.randint(100, 300)})") await asyncio.sleep(random.uniform(0.3, 0.8)) else: await asyncio.sleep(random.uniform(1, 3)) except Exception: await asyncio.sleep(0.5) async def _take_screenshot(self, page, url): """Скриншот.""" try: ts = datetime.now().strftime("%Y%m%d_%H%M%S") domain = url.replace("https://", "").replace("http://", "").split("/")[0][:30] fp = self.screenshots_dir / f"{ts}_{domain}.png" await page.screenshot(path=str(fp), full_page=True) return str(fp) except Exception as e: logger.error(f"Screenshot: {e}") return None async def cleanup(self): """Очистка с таймаутом.""" try: await asyncio.wait_for(self.socks5_pool.stop_all(), timeout=10) except asyncio.TimeoutError: logger.warning("SOCKS5 pool cleanup timeout") except Exception as e: logger.error(f"Cleanup error: {e}") async def visit_page_from_twitch(self, url: str, channel: str, reading_time: int = None) -> VisitResult: if reading_time is None: reading_time = random.randint(settings.DEFAULT_MIN_READING, settings.DEFAULT_MAX_READING) proxy = None for attempt in range(2): try: if self.proxy_manager.has_proxies(): proxy = await self.proxy_manager.acquire_proxy(timeout=90) if proxy and proxy.proxy_type in [ProxyType.SOCKS5, ProxyType.SOCKS4]: proxy_config = await self.socks5_pool.get_proxy_config(proxy) elif proxy: proxy_config = proxy.proxy_config else: proxy_config = None browser_kwargs = { "headless": True, "geoip": True, "humanize": True, "exclude_addons": [DefaultAddons.UBO], } if proxy_config: browser_kwargs["proxy"] = proxy_config async with AsyncCamoufox(**browser_kwargs) as browser: result = await self._browse_page( browser, url, proxy.id if proxy else "direct", reading_time ) if proxy: await self.proxy_manager.release_proxy(proxy, result.success) if result.success: return result except Exception as e: logger.warning(f"Attempt {attempt + 1} failed: {e}") if proxy: await self.proxy_manager.release_proxy(proxy, success=False) await self.socks5_pool.release(proxy) proxy = None if attempt < 1: await asyncio.sleep(2) return VisitResult(url=url, success=False, error="All attempts failed")