zhiyinya/AIStudioBuildWS
0
1import os2from playwright.sync_api import TimeoutError, Error as PlaywrightError3from utils.logger import setup_logging4from utils.cookie_manager import CookieManager5from browser.navigation import handle_successful_navigation6from camoufox.sync_api import Camoufox7from utils.paths import logs_dir8from utils.common import parse_headless_mode, ensure_dir9from utils.url_helper import extract_url_path10 11 12def run_browser_instance(config):13 """14 根据最终合并的配置，启动并管理一个单独的 Camoufox 浏览器实例。15 使用CookieManager统一管理cookie加载，避免重复的扫描逻辑。16 """17 cookie_source = config.get('cookie_source')18 if not cookie_source:19 # 使用默认logger进行错误报告20 logger = setup_logging(os.path.join(logs_dir(), 'app.log'))21 logger.error("错误: 配置中缺少cookie_source对象")22 return23 24 instance_label = cookie_source.display_name25 logger = setup_logging(26 os.path.join(logs_dir(), 'app.log'), prefix=instance_label27 )28 diagnostic_tag = instance_label.replace(os.sep, "_")29 30 expected_url = config.get('url')31 proxy = config.get('proxy')32 headless_setting = config.get('headless', 'virtual')33 34 # 使用CookieManager加载cookie35 cookie_manager = CookieManager(logger)36 all_cookies = []37 38 try:39 # 直接使用CookieSource对象加载cookie40 cookies = cookie_manager.load_cookies(cookie_source)41 all_cookies.extend(cookies)42 43 except Exception as e:44 logger.error(f"从cookie来源加载时出错: {e}")45 return46 47 # 3. 检查是否有任何cookie可用48 if not all_cookies:49 logger.error("错误: 没有可用的cookie（既没有有效的JSON文件，也没有环境变量）")50 return51 52 cookies = all_cookies53 54 headless_mode = parse_headless_mode(headless_setting)55 launch_options = {"headless": headless_mode}56 if proxy:57 logger.info(f"使用代理: {proxy} 访问")58 launch_options["proxy"] = {"server": proxy, "bypass": "localhost, 127.0.0.1"}59 # 无需禁用图片加载, 因为图片很少, 禁用还可能导致风控增加60 # launch_options["block_images"] = True61 62 screenshot_dir = logs_dir()63 ensure_dir(screenshot_dir)64 65 try:66 with Camoufox(**launch_options) as browser:67 context = browser.new_context()68 context.add_cookies(cookies)69 page = context.new_page()70 71 # ####################################################################72 # ############ 增强的 page.goto() 错误处理和日志记录 ###############73 # ####################################################################74 75 response = None76 try:77 logger.info(f"正在导航到: {expected_url} (超时设置为 120 秒)")78 # page.goto() 会返回一个 response 对象，我们可以用它来获取状态码等信息79 response = page.goto(expected_url, wait_until='domcontentloaded', timeout=120000)80 81 # 检查HTTP响应状态码82 if response:83 logger.info(f"导航初步成功，服务器响应状态码: {response.status} {response.status_text}")84 if not response.ok: # response.ok 检查状态码是否在 200-299 范围内85 logger.warning(f"警告：页面加载成功，但HTTP状态码表示错误: {response.status}")86 # 即使状态码错误，也保存快照以供分析87 page.screenshot(path=os.path.join(screenshot_dir, f"WARN_http_status_{response.status}_{diagnostic_tag}.png"))88 else:89 # 对于非http/https的导航（如 about:blank），response可能为None90 logger.warning("page.goto 未返回响应对象，可能是一个非HTTP导航。")91 92 except TimeoutError:93 # 这是最常见的错误：超时94 logger.error(f"导航到 {expected_url} 超时 (超过120秒)。")95 logger.error("可能原因：网络连接缓慢、目标网站服务器无响应、代理问题、或页面资源被阻塞。")96 # 尝试保存诊断信息97 try:98 # 截图对于看到页面卡在什么状态非常有帮助（例如，空白页、加载中、Chrome错误页）99 screenshot_path = os.path.join(screenshot_dir, f"FAIL_timeout_{diagnostic_tag}.png")100 page.screenshot(path=screenshot_path, full_page=True)101 logger.info(f"已截取超时时的屏幕快照: {screenshot_path}")102 103 # 保存HTML可以帮助分析DOM结构，即使在无头模式下也很有用104 html_path = os.path.join(screenshot_dir, f"FAIL_timeout_{diagnostic_tag}.html")105 with open(html_path, 'w', encoding='utf-8') as f:106 f.write(page.content())107 logger.info(f"已保存超时时的页面HTML: {html_path}")108 except Exception as diag_e:109 logger.error(f"在尝试进行超时诊断（截图/保存HTML）时发生额外错误: {diag_e}")110 return # 超时后，后续操作无意义，直接终止111 112 except PlaywrightError as e:113 # 捕获其他Playwright相关的网络错误，例如DNS解析失败、连接被拒绝等114 error_message = str(e)115 logger.error(f"导航到 {expected_url} 时发生 Playwright 网络错误。")116 logger.error(f"错误详情: {error_message}")117 118 # Playwright的错误信息通常很具体，例如 "net::ERR_CONNECTION_REFUSED"119 if "net::ERR_NAME_NOT_RESOLVED" in error_message:120 logger.error("排查建议：检查DNS设置或域名是否正确。")121 elif "net::ERR_CONNECTION_REFUSED" in error_message:122 logger.error("排查建议：目标服务器可能已关闭，或代理/防火墙阻止了连接。")123 elif "net::ERR_INTERNET_DISCONNECTED" in error_message:124 logger.error("排查建议：检查本机的网络连接。")125 126 # 同样，尝试截图，尽管此时页面可能完全无法访问127 try:128 screenshot_path = os.path.join(screenshot_dir, f"FAIL_network_error_{diagnostic_tag}.png")129 page.screenshot(path=screenshot_path)130 logger.info(f"已截取网络错误时的屏幕快照: {screenshot_path}")131 except Exception as diag_e:132 logger.error(f"在尝试进行网络错误诊断（截图）时发生额外错误: {diag_e}")133 return # 网络错误，终止134 135 # --- 如果导航没有抛出异常，继续执行后续逻辑 ---136 137 logger.info("页面初步加载完成，正在检查并处理初始弹窗...")138 page.wait_for_timeout(2000)139 140 final_url = page.url141 logger.info(f"导航完成。最终URL为: {final_url}")142 143 # ... 你原有的URL检查逻辑保持不变 ...144 if "accounts.google.com/v3/signin/identifier" in final_url:145 logger.error("检测到Google登录页面（需要输入邮箱）。Cookie已完全失效。")146 page.screenshot(path=os.path.join(screenshot_dir, f"FAIL_identifier_page_{diagnostic_tag}.png"))147 return148 149 # 提取路径部分进行匹配（允许域名重定向）150 expected_path = extract_url_path(expected_url).split('?')[0]151 final_path = extract_url_path(final_url)152 153 if expected_path and expected_path in final_path:154 logger.info(f"URL验证通过。预期路径: {expected_path}, 最终URL: {final_url}")155 156 # --- NEW ROBUST STRATEGY: Wait for the loading spinner to disappear ---157 # This is the key to solving the race condition. The error message or158 # content will only appear AFTER the initial loading is done.159 spinner_locator = page.locator('mat-spinner')160 try:161 logger.info("正在等待加载指示器 (spinner) 消失... (最长等待30秒)")162 # We wait for the spinner to be 'hidden' or not present in the DOM.163 spinner_locator.wait_for(state='hidden', timeout=30000)164 logger.info("加载指示器已消失。页面已完成异步加载。")165 except TimeoutError:166 logger.error("页面加载指示器在30秒内未消失。页面可能已卡住。")167 page.screenshot(path=os.path.join(screenshot_dir, f"FAIL_spinner_stuck_{diagnostic_tag}.png"))168 return # Exit if the page is stuck loading169 170 # --- NOW, we can safely check for the error message ---171 # We use the most specific text possible to avoid false positives.172 auth_error_text = "authentication error"173 auth_error_locator = page.get_by_text(auth_error_text, exact=False)174 175 # We only need a very short timeout here because the page should be stable.176 if auth_error_locator.is_visible(timeout=2000):177 logger.error(f"检测到认证失败的错误横幅: '{auth_error_text}'. Cookie已过期或无效。")178 screenshot_path = os.path.join(screenshot_dir, f"FAIL_auth_error_banner_{diagnostic_tag}.png")179 page.screenshot(path=screenshot_path)180 181 # html_path = os.path.join(screenshot_dir, f"FAIL_auth_error_banner_{diagnostic_tag}.html")182 # with open(html_path, 'w', encoding='utf-8') as f:183 # f.write(page.content())184 # logger.info(f"已保存包含错误信息的页面HTML: {html_path}")185 return # Definitive failure, so we exit.186 187 # --- If no error, proceed to final confirmation (as a fallback) ---188 logger.info("未检测到认证错误横幅。进行最终确认。")189 login_button_cn = page.get_by_role('button', name='登录')190 login_button_en = page.get_by_role('button', name='Login')191 192 if login_button_cn.is_visible(timeout=1000) or login_button_en.is_visible(timeout=1000):193 logger.error("页面上仍显示'登录'按钮。Cookie无效。")194 page.screenshot(path=os.path.join(screenshot_dir, f"FAIL_login_button_visible_{diagnostic_tag}.png"))195 return196 197 # --- If all checks pass, we assume success ---198 logger.info("所有验证通过，确认已成功登录。")199 handle_successful_navigation(page, logger, diagnostic_tag)200 elif "accounts.google.com/v3/signin/accountchooser" in final_url:201 logger.warning("检测到Google账户选择页面。登录失败或Cookie已过期。")202 page.screenshot(path=os.path.join(screenshot_dir, f"FAIL_chooser_click_failed_{diagnostic_tag}.png"))203 return204 else:205 logger.error(f"导航到了意外的URL。")206 logger.error(f" 预期路径: {expected_path}")207 logger.error(f" 最终URL: {final_url}")208 logger.error(f" 最终路径: {final_path}")209 page.screenshot(path=os.path.join(screenshot_dir, f"FAIL_unexpected_url_{diagnostic_tag}.png"))210 return211 212 except KeyboardInterrupt:213 logger.info(f"用户中断，正在关闭...")214 except Exception as e:215 # 这是一个最终的捕获，用于捕获所有未预料到的错误216 logger.exception(f"运行 Camoufox 实例时发生未预料的严重错误: {e}")217