research/pdf_downloader/save-page-with-selenium.py

@@ -0,0 +1,322 @@
+import logging
+import os
+import queue
+import threading
+import time
+from datetime import datetime
+
+import pymysql
+from tqdm import tqdm
+
+from save_page_as_pdf import PDFSaver
+from save_remote_as_mhtml import RemoteMHTMLSaver
+from save_page_as_mhtml import MHTMLSaver
+import tldextract
+
+# 配置日志
+from save_remote_as_pdf import RemotePDFSaver
+
+logging.basicConfig(
+    level=logging.INFO,
+    format='%(asctime)s - %(levelname)s - %(message)s',
+    handlers=[
+        logging.StreamHandler(),
+        logging.FileHandler('pdf_downloader.log')
+    ]
+)
+logger = logging.getLogger(__name__)
+
+# =============== MySQL 配置 ===============
+MYSQL_CONFIG = {
+    'host': '47.113.231.200',
+    'port': 28089,
+    'user': 'root',
+    'password': 'passok123A',
+    'database': 'dsp',
+    'charset': 'utf8mb4',
+    'autocommit': False  # 手动控制事务
+}
+# =========================================
+
+# 配置参数
+BATCH_SIZE = 500
+MAX_WORKERS = 1
+TIMEOUT = 10
+PDF_OUTPUT_DIR = 'D:/data/output/pdf'
+MIN_PDF_SIZE = 80 * 1024  # 80KB
+
+MHTML_OUTPUT_DIR = 'D:/data/output/mhtml'
+os.makedirs(PDF_OUTPUT_DIR, exist_ok=True)
+
+running = True
+running_interval_seconds = 15
+
+remote_host_name = [
+    'epochtimes.com',
+    # 'secretchina.com'
+]
+
+
+class PDFDownloader:
+    def __init__(self):
+        self.db_lock = threading.Lock()
+        self.task_queue = queue.Queue(maxsize=MAX_WORKERS * 3)
+        self.processed_count = 0
+        self.success_count = 0
+        self.fail_count = 0
+        self.small_file_count = 0  # 新增：统计小文件数量
+        self.last_loadtime = self.get_last_loadtime()
+        self.total_rows = self.get_total_rows()
+        self.start_time = time.time()
+        self.skip_hosts = []
+        self.local_handler = None
+        self.remote_handler = None
+
+    # 替换 MYSQL_CONFIG 中的连接方式
+    def get_db_connection(self):
+        return pymysql.connect(
+            host=MYSQL_CONFIG['host'],
+            port=MYSQL_CONFIG['port'],
+            user=MYSQL_CONFIG['user'],
+            password=MYSQL_CONFIG['password'],
+            database=MYSQL_CONFIG['database'],
+            charset='utf8mb4',
+            autocommit=False
+        )
+
+    def get_total_rows(self):
+        """获取总记录数"""
+        with self.get_db_connection() as conn:
+            cursor = conn.cursor()
+            cursor.execute(
+                "SELECT COUNT(*) FROM indeximos "
+                "WHERE (es_video IS NULL OR es_video IN ('-2', '-1')) "
+                "AND es_loadtime > %s", self.last_loadtime
+            )
+            return cursor.fetchone()[0]
+
+    def get_last_loadtime(self):
+        """获取上次导出数据的时间"""
+        with self.get_db_connection() as conn:
+            cursor = conn.cursor()
+            cursor.execute(
+                "SELECT config_value FROM config "
+                "WHERE config_name = 'last_loadtime' "
+            )
+            return cursor.fetchone()[0]
+
+    def use_remote_selenium(self, url):
+        for host in remote_host_name:
+            if host in url:
+                return True
+        return False
+
+    def format_pdf_filename(self, row):
+        """格式化PDF文件名"""
+        es_urltitle = row[2] or 'untitled'
+        es_urltime = str(row[3]) or '19700101_000000'
+        es_sitename = row[4] or 'anonymous'
+
+        def clean_filename(text):
+            if not text:
+                return ''
+            invalid_chars = ['<', '>', ':', '"', '/', '\\', '|', '?', '*']
+            for char in invalid_chars:
+                text = text.replace(char, '_')
+            return text.strip()[:100]
+
+        try:
+            dt = datetime.strptime(es_urltime, '%Y-%m-%d %H:%M:%S')
+            es_urltime_fix = dt.strftime('%Y%m%d_%H%M%S')
+        except:
+            es_urltime_fix = '19700101_000000'
+
+        filename = f"{clean_filename(es_urltitle)}_{es_urltime_fix}_{es_sitename}.pdf"
+        return os.path.join(PDF_OUTPUT_DIR, filename)
+
+    def format_mhtml_filename(self, row):
+        """格式化PDF文件名"""
+        es_urltitle = row[2] or 'untitled'
+        es_urltime = str(row[3]) or '19700101_000000'
+        es_sitename = row[4] or 'anonymous'
+
+        def clean_filename(text):
+            if not text:
+                return ''
+            invalid_chars = ['<', '>', ':', '"', '/', '\\', '|', '?', '*']
+            for char in invalid_chars:
+                text = text.replace(char, '_')
+            return text.strip()[:100]
+
+        try:
+            dt = datetime.strptime(es_urltime, '%Y-%m-%d %H:%M:%S')
+            es_urltime_fix = dt.strftime('%Y%m%d_%H%M%S')
+        except:
+            es_urltime_fix = '19700101_000000'
+
+        filename = f"{clean_filename(es_urltitle)}_{es_urltime_fix}_{es_sitename}.mhtml"
+        return os.path.join(PDF_OUTPUT_DIR, filename)
+
+    def fetch_data_batch(self, offset):
+        """分页获取数据"""
+        with self.get_db_connection() as conn:
+            cursor = conn.cursor()
+            cursor.execute(
+                "SELECT es_sid, es_urlname, es_urltitle, es_urltime, es_sitename, es_authors FROM indeximos "
+                "WHERE (es_video IS NULL OR es_video IN ('-2', '-1')) "
+                "AND es_loadtime > %s "
+                "ORDER BY es_urltime LIMIT %s OFFSET %s",
+                (self.last_loadtime, BATCH_SIZE, offset)
+            )
+            return cursor.fetchall()
+
+    def update_file_status(self, es_sid, status, retry=3):
+        """更新数据库状态"""
+        for attempt in range(retry):
+            try:
+                with self.db_lock, self.get_db_connection() as conn:
+                    cursor = conn.cursor()
+                    cursor.execute(
+                        "UPDATE indeximos SET es_video = %s WHERE es_sid = %s",
+                        (status, es_sid))
+                    conn.commit()
+                return True
+            except Exception as e:
+                if attempt == retry - 1:
+                    logger.error(f"更新数据库失败(es_sid={es_sid}): {e}")
+                    return False
+                time.sleep(1)
+
+    def extract_main_domain(self, url):
+        extracted = tldextract.extract(url)
+        # 组合注册域名（主域名）
+        main_domain = f"{extracted.domain}.{extracted.suffix}"
+        return main_domain
+
+    def download_worker(self):
+        """工作线程函数"""
+        while True:
+            try:
+                task = self.task_queue.get(timeout=1)
+                if task is None:
+                    break
+
+                row = task
+                url = row[1]
+                if self.extract_main_domain(url) in self.skip_hosts:
+                    self.small_file_count += 1
+                    self.processed_count += 1
+                    self.task_queue.task_done()
+                    print(f"小文件规避，暂时跳过URL：{url}")
+                    continue
+                output_file = self.format_pdf_filename(row)  # 获取格式化后的文件名
+
+                try:
+                    os.makedirs(os.path.dirname(output_file), exist_ok=True)
+
+                    # 调用下载函数
+                    if self.use_remote_selenium(url):
+                        if self.remote_handler is None:
+                            self.remote_handler = RemotePDFSaver()
+                        success = self.remote_handler.save_as_pdf(
+                            url=url,
+                            output_path=output_file,
+                            timeout=TIMEOUT
+                        )
+                    else:
+                        if self.local_handler is None:
+                            self.local_handler = PDFSaver()
+                        success = self.local_handler.save_as_pdf(
+                            url=url,
+                            output_path=output_file,
+                            timeout=TIMEOUT
+                        )
+
+                    # 验证下载结果
+                    if success and os.path.exists(output_file):
+                        file_size = os.path.getsize(output_file)
+
+                        if file_size >= MIN_PDF_SIZE:  # 文件大小合格
+                            self.update_file_status(row[0], output_file)
+                            self.success_count += 1
+                        else:  # 文件太小
+                            self.update_file_status(row[0], '-2')
+                            self.small_file_count += 1
+                            logger.warning(f"文件过小({file_size}字节): {output_file}")
+                            try:
+                                os.remove(output_file)
+                                self.skip_hosts.append(self.extract_main_domain(url))
+                            except:
+                                pass
+                    else:  # 下载失败
+                        self.update_file_status(row[0], '0')
+                        self.fail_count += 1
+                        if os.path.exists(output_file):
+                            try:
+                                os.remove(output_file)
+                            except:
+                                pass
+
+                except Exception as e:
+                    logger.error(f"下载出现异常(es_sid={row[0]}, url={url}): {str(e)}")
+                    self.update_file_status(row[0], '-1')
+                    self.fail_count += 1
+
+                self.processed_count += 1
+                self.task_queue.task_done()
+
+            except queue.Empty:
+                continue
+
+    def run(self):
+        """启动下载任务"""
+        threads = []
+
+        # 创建工作线程
+        for _ in range(MAX_WORKERS):
+            t = threading.Thread(target=self.download_worker)
+            t.start()
+            threads.append(t)
+
+        # 使用进度条显示进度
+        with tqdm(total=self.total_rows, desc="处理进度", unit="条") as pbar:
+            offset = 0
+            while True:
+                batch = self.fetch_data_batch(offset)
+                if not batch:
+                    break
+
+                for row in batch:
+                    self.task_queue.put(row)
+
+                pbar.update(len(batch))
+                pbar.set_postfix({
+                    '成功': self.success_count,
+                    '失败': self.fail_count,
+                    '小文件': self.small_file_count,
+                    '速度': f"{self.processed_count / (time.time() - self.start_time):.1f}条/秒"
+                })
+
+                offset += BATCH_SIZE
+
+        self.task_queue.join()
+
+        for _ in range(MAX_WORKERS):
+            self.task_queue.put(None)
+
+        for t in threads:
+            t.join()
+
+        total_time = time.time() - self.start_time
+        print(f"\n处理完成! 总计: {self.total_rows}条")
+        print(f"成功: {self.success_count}条, 失败: {self.fail_count}条, 小文件: {self.small_file_count}条")
+        print(f"总耗时: {total_time:.2f}秒, 平均速度: {self.total_rows / total_time:.2f}条/秒")
+
+
+if __name__ == "__main__":
+    while running:
+        print(f"开始处理，总记录数: {PDFDownloader().get_total_rows()}")
+        downloader = PDFDownloader()
+        downloader.run()
+        print(f"运行完成，暂停{running_interval_seconds}秒后开始下一次运行...")
+        time.sleep(running_interval_seconds)

research/pdf_downloader/save_page_as_mhtml.py

@@ -0,0 +1,141 @@
+import logging
+import os
+import time
+from urllib.parse import urlparse
+
+from selenium import webdriver
+from selenium.webdriver.chrome.options import Options
+from selenium.webdriver.chrome.service import Service as ChromeService
+
+# 配置日志
+logging.basicConfig(
+    level=logging.INFO,
+    format='%(asctime)s - %(levelname)s - %(message)s',
+    handlers=[
+        logging.StreamHandler(),
+        logging.FileHandler('mhtml_saver.log', encoding='utf-8')
+    ]
+)
+logger = logging.getLogger(__name__)
+
+
+class MHTMLSaver:
+    def __init__(self, headless=True):
+        logger.info("正在初始化 Chrome WebDriver（自动匹配版本）...")
+        service = ChromeService(executable_path="C:/Program Files/Python38/chromedriver.exe")
+        user_agent = "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/143.0.3650.75 Safari/537.36 Edg/143.0.3650.75"
+
+        # Chrome 选项
+        chrome_options = Options()
+        chrome_options.add_argument('--headless=new')
+        chrome_options.add_argument('--disable-gpu')
+        chrome_options.add_argument('--no-sandbox')
+        chrome_options.add_argument('--disable-dev-shm-usage')
+        chrome_options.add_argument(f'--user-agent={user_agent}')
+        chrome_options.add_argument('--save-page-as-mhtml')  # 启用 MHTML 支持
+        chrome_options.add_experimental_option('excludeSwitches', ['enable-automation'])
+        chrome_options.add_experimental_option('useAutomationExtension', False)
+        chrome_options.add_argument('--lang=zh-CN')
+        chrome_options.add_experimental_option('prefs', {
+            'intl.accept_languages': 'zh-CN,zh,en'
+        })
+        # 或启动时指定（部分版本支持）
+        chrome_options.add_argument('--window-size=1920,1080')
+
+        # 隐藏 webdriver 特征
+        chrome_options.add_argument("--disable-blink-features=AutomationControlled")
+        # 隐藏 "navigator.webdriver"
+        chrome_options.add_experimental_option('excludeSwitches', ['enable-automation'])
+        chrome_options.add_experimental_option('useAutomationExtension', False)
+
+        self.driver = webdriver.Chrome(service=service, options=chrome_options)
+
+    def save_as_mhtml(self, url, output_path=None, timeout=30, wait_time=5):
+        """
+        将网页保存为 MHTML 文件
+        :param url: 目标网页 URL
+        :param output_path: 输出路径（.mhtml）
+        :param timeout: 页面加载超时（秒）
+        :param wait_time: 页面加载后等待时间（秒），用于动态内容渲染
+        :return: 保存的文件绝对路径
+        """
+        if output_path is None:
+            parsed = urlparse(url)
+            domain = parsed.netloc.replace('www.', '').split('.')[0] or 'page'
+            output_path = f"{domain}.mhtml"
+
+        if not output_path.lower().endswith('.mhtml'):
+            output_path += '.mhtml'
+
+        try:
+            # 设置超时
+            self.driver.set_page_load_timeout(timeout)
+
+            # 启动后注入脚本（双重保险）
+            self.driver.execute_cdp_cmd('Page.addScriptToEvaluateOnNewDocument', {
+                'source': '''
+                    delete navigator.__proto__.webdriver;
+                    window.navigator.permissions.query = (parameters) => {
+                        return parameters.name === 'notifications' ?
+                            Promise.resolve({ state: Notification.permission }) :
+                            originalQuery(parameters);
+                    };
+                '''
+            })
+            # 在 driver.get() 之后设置窗口大小
+
+            logger.info(f"正在加载页面: {url}")
+            self.driver.get(url)
+            self.driver.set_window_size(1920, 1080)
+
+            # 等待页面动态内容加载（可调整）
+            logger.info(f"等待 {wait_time} 秒以确保页面完全渲染...")
+            time.sleep(wait_time)
+
+            # ✅ 关键：调用 CDP 命令捕获 MHTML
+            logger.info("正在生成 MHTML 快照...")
+            result = self.driver.execute_cdp_cmd('Page.captureSnapshot', {'format': 'mhtml'})
+
+            # ✅ result['data'] 是 Base64 编码的 MHTML 文本（实际是纯字符串）
+            mhtml_content = result['data']
+
+            # ✅ 以文本模式写入（UTF-8）
+            with open(output_path, 'w', encoding='utf-8', newline='') as f:
+                f.write(mhtml_content)
+
+            # 验证文件
+            file_size = os.path.getsize(output_path)
+            if file_size == 0:
+                raise RuntimeError("生成了空文件")
+
+            logger.info(f"✅ MHTML 保存成功: {os.path.abspath(output_path)} (大小: {file_size} 字节)")
+            return os.path.abspath(output_path)
+
+        except Exception as e:
+            logger.error(f"❌ 保存失败: {e}")
+            raise
+
+    def quit(self):
+        if self.driver:
+            self.driver.quit()
+            logger.info("浏览器已关闭")
+
+
+# ===== 测试入口 =====
+if __name__ == "__main__":
+    # 示例 URL（可替换为你自己的）
+    test_url = "https://cn.ultraiso.net/jiaocheng/ke-lu-guang-pan.html"
+
+    saver = MHTMLSaver(headless=True)
+    try:
+        output_file = saver.save_as_mhtml(
+            url=test_url,
+            output_path="example.mhtml",
+            timeout=30,
+            wait_time=5
+        )
+        print(f"\n🎉 成功保存 MHTML 文件: {output_file}")
+    except Exception as e:
+        print(f"\n💥 保存失败: {e}")
+    finally:
+        saver.quit()

research/pdf_downloader/save_page_as_pdf.py

@@ -0,0 +1,144 @@
+import base64
+import logging
+import os
+import time
+from urllib.parse import urlparse
+
+from selenium import webdriver
+from selenium.webdriver.chrome.options import Options
+from selenium.webdriver.chrome.service import Service as ChromeService
+
+# 配置日志
+logging.basicConfig(
+    level=logging.INFO,
+    format='%(asctime)s - %(levelname)s - %(message)s',
+    handlers=[
+        logging.StreamHandler(),
+        logging.FileHandler('pdf_saver.log', encoding='utf-8')
+    ]
+)
+logger = logging.getLogger(__name__)
+
+
+class PDFSaver:
+    def __init__(self, headless=True):
+        logger.info("正在初始化 Chrome WebDriver（自动匹配版本）...")
+        service = ChromeService(executable_path="C:/Program Files/Python38/chromedriver.exe")
+        user_agent = "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/143.0.3650.75 Safari/537.36 Edg/143.0.3650.75"
+
+        # Chrome 选项
+        chrome_options = Options()
+        if headless:
+            chrome_options.add_argument('--headless=new')
+        chrome_options.add_argument('--disable-gpu')
+        chrome_options.add_argument('--no-sandbox')
+        chrome_options.add_argument('--disable-dev-shm-usage')
+        chrome_options.add_argument(f'--user-agent={user_agent}')
+        chrome_options.add_experimental_option('excludeSwitches', ['enable-automation'])
+        chrome_options.add_experimental_option('useAutomationExtension', False)
+        chrome_options.add_argument('--lang=zh-CN')
+        chrome_options.add_experimental_option('prefs', {
+            'intl.accept_languages': 'zh-CN,zh,en'
+        })
+        chrome_options.add_argument('--window-size=1920,1080')
+        chrome_options.add_argument("--disable-blink-features=AutomationControlled")
+
+        # 注意：PDF 打印不需要 --save-page-as-mhtml
+        self.driver = webdriver.Chrome(service=service, options=chrome_options)
+
+    def save_as_pdf(self, url, output_path=None, timeout=30, wait_time=5, print_options=None):
+        """
+        将网页保存为 PDF 文件
+        :param url: 目标网页 URL
+        :param output_path: 输出路径（.pdf）
+        :param timeout: 页面加载超时（秒）
+        :param wait_time: 页面加载后等待时间（秒），用于动态内容渲染
+        :param print_options: PDF 打印选项（可选），参考 https://chromedevtools.github.io/devtools-protocol/tot/Page/#method-printToPDF
+        :return: 保存的文件绝对路径
+        """
+        if output_path is None:
+            parsed = urlparse(url)
+            domain = parsed.netloc.replace('www.', '').split('.')[0] or 'page'
+            output_path = f"{domain}.pdf"
+
+        if not output_path.lower().endswith('.pdf'):
+            output_path += '.pdf'
+
+        # 默认打印选项（可按需调整）
+        default_print_options = {
+            'landscape': False,
+            'displayHeaderFooter': False,
+            'printBackground': True,
+            'preferCSSPageSize': True,
+            'paperWidth': 8.27,  # A4 宽（英寸）
+            'paperHeight': 11.69,  # A4 高（英寸）
+        }
+        if print_options:
+            default_print_options.update(print_options)
+
+        try:
+            self.driver.set_page_load_timeout(timeout)
+
+            # 隐藏自动化特征
+            self.driver.execute_cdp_cmd('Page.addScriptToEvaluateOnNewDocument', {
+                'source': '''
+                    delete navigator.__proto__.webdriver;
+                    window.navigator.permissions.query = (parameters) => {
+                        return parameters.name === 'notifications' ?
+                            Promise.resolve({ state: Notification.permission }) :
+                            originalQuery(parameters);
+                    };
+                    const originalQuery = window.navigator.permissions.query;
+                '''
+            })
+
+            logger.info(f"正在加载页面: {url}")
+            self.driver.get(url)
+            self.driver.set_window_size(1920, 1080)
+
+            logger.info(f"等待 {wait_time} 秒以确保页面完全渲染...")
+            time.sleep(wait_time)
+
+            logger.info("正在生成 PDF...")
+            result = self.driver.execute_cdp_cmd('Page.printToPDF', default_print_options)
+
+            # result['data'] 是 Base64 编码的 PDF
+            pdf_data = base64.b64decode(result['data'])
+
+            with open(output_path, 'wb') as f:
+                f.write(pdf_data)
+
+            file_size = os.path.getsize(output_path)
+            if file_size == 0:
+                raise RuntimeError("生成了空文件")
+
+            logger.info(f"✅ PDF 保存成功: {os.path.abspath(output_path)} (大小: {file_size} 字节)")
+            return os.path.abspath(output_path)
+
+        except Exception as e:
+            logger.error(f"❌ 保存失败: {e}")
+            raise
+
+    def quit(self):
+        if self.driver:
+            self.driver.quit()
+            logger.info("浏览器已关闭")
+
+
+# ===== 测试入口 =====
+if __name__ == "__main__":
+    test_url = "https://cn.ultraiso.net/jiaocheng/ke-lu-guang-pan.html"
+
+    saver = PDFSaver(headless=True)
+    try:
+        output_file = saver.save_as_pdf(
+            url=test_url,
+            output_path="example.pdf",
+            timeout=30,
+            wait_time=5
+        )
+        print(f"\n🎉 成功保存 PDF 文件: {output_file}")
+    except Exception as e:
+        print(f"\n💥 保存失败: {e}")
+    finally:
+        saver.quit()

research/pdf_downloader/save_remote_as_mhtml.py

@@ -0,0 +1,190 @@
+import logging
+import os
+import time
+from urllib.parse import urlparse
+
+from selenium import webdriver
+from selenium.common.exceptions import (
+    WebDriverException,
+    TimeoutException,
+    SessionNotCreatedException,
+    InvalidSessionIdException
+)
+from selenium.webdriver.chrome.options import Options
+
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(__name__)
+
+
+class RemoteMHTMLSaver:
+    def __init__(
+            self,
+            remote_url="http://144.34.185.108:28098/wd/hub",
+            headless=True,
+            max_retries=3,
+            retry_delay=2
+    ):
+        """
+        初始化远程 MHTML 保存器（支持自动重建 session）
+        :param remote_url: 远程 Selenium 地址
+        :param headless: 是否无头
+        :param max_retries: 单次操作最大重试次数
+        :param retry_delay: 重试前等待时间（秒）
+        """
+        self.remote_url = remote_url
+        self.headless = headless
+        self.max_retries = max_retries
+        self.retry_delay = retry_delay
+        self.driver = None
+        self._init_driver()
+
+    def _build_chrome_options(self):
+        """构建 Chrome 选项（可复用）"""
+        chrome_options = Options()
+        if self.headless:
+            chrome_options.add_argument('--headless=new')
+        chrome_options.add_argument('--no-sandbox')
+        chrome_options.add_argument('--disable-dev-shm-usage')
+        chrome_options.add_argument('--disable-gpu')
+        chrome_options.add_argument('--window-size=1920,1080')
+        chrome_options.add_argument(
+            "--user-agent=Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/143.0.3650.75 Safari/537.36"
+        )
+        chrome_options.add_experimental_option('excludeSwitches', ['enable-automation'])
+        chrome_options.add_experimental_option('useAutomationExtension', False)
+        return chrome_options
+
+    def _init_driver(self):
+        """初始化或重新初始化 WebDriver"""
+        if self.driver:
+            try:
+                self.driver.quit()
+            except Exception:
+                pass  # 忽略关闭失败
+
+        logger.info(f"正在创建新的远程 WebDriver 会话: {self.remote_url}")
+        for attempt in range(3):
+            try:
+                self.driver = webdriver.Remote(
+                    command_executor=self.remote_url,
+                    options=self._build_chrome_options()
+                )
+                # 注入反检测脚本
+                self.driver.execute_cdp_cmd('Page.addScriptToEvaluateOnNewDocument', {
+                    'source': '''
+                        delete navigator.__proto__.webdriver;
+                        window.chrome = { runtime: {} };
+                        Object.defineProperty(navigator, 'languages', {
+                            get: () => ['zh-CN', 'zh', 'en']
+                        });
+                    '''
+                })
+                logger.info("✅ 远程 WebDriver 会话创建成功")
+                return
+            except Exception as e:
+                logger.warning(f"创建 WebDriver 失败 (尝试 {attempt + 1}/3): {e}")
+                if attempt < 2:
+                    time.sleep(2)
+                else:
+                    raise RuntimeError(f"无法连接到远程 Selenium 服务: {e}")
+
+    def save_as_mhtml(self, url, output_path=None, timeout=30, wait_time=5):
+        """
+        保存网页为 MHTML，支持自动重试和 session 重建
+        """
+        if output_path is None:
+            domain = urlparse(url).netloc.replace('www.', '').split('.')[0] or 'page'
+            output_path = f"{domain}.mhtml"
+        if not output_path.lower().endswith('.mhtml'):
+            output_path += '.mhtml'
+
+        last_exception = None
+
+        for retry in range(self.max_retries + 1):
+            try:
+                # 检查 driver 是否有效
+                if not self.driver:
+                    self._init_driver()
+
+                self.driver.set_page_load_timeout(timeout)
+                logger.info(f"[{retry + 1}/{self.max_retries + 1}] 加载页面: {url}")
+                self.driver.get(url)
+                time.sleep(wait_time)
+
+                logger.info("生成 MHTML 快照...")
+                result = self.driver.execute_cdp_cmd('Page.captureSnapshot', {'format': 'mhtml'})
+                mhtml_content = result['data']
+
+                # 写入本地文件
+                with open(output_path, 'w', encoding='utf-8', newline='') as f:
+                    f.write(mhtml_content)
+
+                file_size = os.path.getsize(output_path)
+                if file_size == 0:
+                    raise RuntimeError("生成了空文件")
+
+                logger.info(f"✅ 保存成功: {os.path.abspath(output_path)} ({file_size} 字节)")
+                return os.path.abspath(output_path)
+
+            except (WebDriverException, InvalidSessionIdException, SessionNotCreatedException) as e:
+                last_exception = e
+                logger.warning(f"WebDriver 异常 (retry {retry + 1}): {e}")
+                if retry < self.max_retries:
+                    logger.info("正在重建 WebDriver 会话...")
+                    self._init_driver()
+                    time.sleep(self.retry_delay)
+                else:
+                    logger.error("达到最大重试次数，放弃")
+                    break
+
+            except TimeoutException as e:
+                last_exception = e
+                logger.warning(f"页面加载超时 (retry {retry + 1}): {e}")
+                if retry < self.max_retries:
+                    time.sleep(self.retry_delay)
+                else:
+                    break
+
+            except Exception as e:
+                last_exception = e
+                logger.error(f"未知错误 (retry {retry + 1}): {e}")
+                break  # 非 WebDriver 错误，不重试
+
+        # 如果所有重试失败
+        if os.path.exists(output_path):
+            try:
+                os.remove(output_path)
+            except OSError:
+                pass
+
+        raise RuntimeError(f"保存失败（{type(last_exception).__name__}）: {last_exception}")
+
+    def quit(self):
+        """显式关闭浏览器"""
+        if self.driver:
+            try:
+                self.driver.quit()
+                logger.info("WebDriver 会话已关闭")
+            except Exception:
+                pass
+            self.driver = None
+
+    def __del__(self):
+        self.quit()
+
+
+# ===== 测试 =====
+if __name__ == "__main__":
+    saver = RemoteMHTMLSaver(
+        remote_url="http://144.34.185.108:28098/wd/hub",  # ← 替换为你的云服务器公网 IP
+        headless=True
+    )
+    try:
+        saver.save_as_mhtml(
+            url="https://www.epochtimes.com/gb/25/12/22/n14660274.htm",
+            output_path="remote_example2.mhtml"
+        )
+    except Exception as e:
+        print(f"❌ 失败: {e}")
+
+    saver.quit()

research/pdf_downloader/save_remote_as_pdf.py

@@ -0,0 +1,201 @@
+import base64
+import logging
+import os
+import time
+from urllib.parse import urlparse
+
+from selenium import webdriver
+from selenium.common.exceptions import (
+    WebDriverException,
+    TimeoutException,
+    SessionNotCreatedException,
+    InvalidSessionIdException
+)
+from selenium.webdriver.chrome.options import Options
+
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(__name__)
+
+
+class RemotePDFSaver:
+    def __init__(
+            self,
+            remote_url="http://144.34.185.108:28098/wd/hub",
+            headless=True,
+            max_retries=3,
+            retry_delay=2,
+            print_options=None
+    ):
+        """
+        初始化远程 PDF 保存器（支持自动重建 session）
+        :param remote_url: 远程 Selenium 地址
+        :param headless: 是否无头模式
+        :param max_retries: 单次操作最大重试次数
+        :param retry_delay: 重试前等待时间（秒）
+        :param print_options: PDF 打印选项（参考 DevTools Protocol）
+        """
+        self.remote_url = remote_url
+        self.headless = headless
+        self.max_retries = max_retries
+        self.retry_delay = retry_delay
+        self.print_options = print_options or {
+            'landscape': False,
+            'displayHeaderFooter': False,
+            'printBackground': True,
+            'preferCSSPageSize': True,
+            'paperWidth': 8.27,  # A4 宽（英寸）
+            'paperHeight': 11.69,  # A4 高（英寸）
+        }
+        self.driver = None
+        self._init_driver()
+
+    def _build_chrome_options(self):
+        """构建 Chrome 选项（可复用）"""
+        chrome_options = Options()
+        if self.headless:
+            chrome_options.add_argument('--headless=new')
+        chrome_options.add_argument('--no-sandbox')
+        chrome_options.add_argument('--disable-dev-shm-usage')
+        chrome_options.add_argument('--disable-gpu')
+        chrome_options.add_argument('--window-size=1920,1080')
+        chrome_options.add_argument(
+            "--user-agent=Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/143.0.3650.75 Safari/537.36"
+        )
+        chrome_options.add_experimental_option('excludeSwitches', ['enable-automation'])
+        chrome_options.add_experimental_option('useAutomationExtension', False)
+        return chrome_options
+
+    def _init_driver(self):
+        """初始化或重新初始化 WebDriver"""
+        if self.driver:
+            try:
+                self.driver.quit()
+            except Exception:
+                pass  # 忽略关闭失败
+
+        logger.info(f"正在创建新的远程 WebDriver 会话: {self.remote_url}")
+        for attempt in range(3):
+            try:
+                self.driver = webdriver.Remote(
+                    command_executor=self.remote_url,
+                    options=self._build_chrome_options()
+                )
+                # 注入反检测脚本
+                self.driver.execute_cdp_cmd('Page.addScriptToEvaluateOnNewDocument', {
+                    'source': '''
+                        delete navigator.__proto__.webdriver;
+                        window.chrome = { runtime: {} };
+                        Object.defineProperty(navigator, 'languages', {
+                            get: () => ['zh-CN', 'zh', 'en']
+                        });
+                    '''
+                })
+                logger.info("✅ 远程 WebDriver 会话创建成功")
+                return
+            except Exception as e:
+                logger.warning(f"创建 WebDriver 失败 (尝试 {attempt + 1}/3): {e}")
+                if attempt < 2:
+                    time.sleep(2)
+                else:
+                    raise RuntimeError(f"无法连接到远程 Selenium 服务: {e}")
+
+    def save_as_pdf(self, url, output_path=None, timeout=30, wait_time=5):
+        """
+        保存网页为 PDF，支持自动重试和 session 重建
+        """
+        if output_path is None:
+            domain = urlparse(url).netloc.replace('www.', '').split('.')[0] or 'page'
+            output_path = f"{domain}.pdf"
+        if not output_path.lower().endswith('.pdf'):
+            output_path += '.pdf'
+
+        last_exception = None
+
+        for retry in range(self.max_retries + 1):
+            try:
+                # 检查 driver 是否有效
+                if not self.driver:
+                    self._init_driver()
+
+                self.driver.set_page_load_timeout(timeout)
+                logger.info(f"[{retry + 1}/{self.max_retries + 1}] 加载页面: {url}")
+                self.driver.get(url)
+                time.sleep(wait_time)
+
+                logger.info("生成 PDF...")
+                result = self.driver.execute_cdp_cmd('Page.printToPDF', self.print_options)
+                pdf_data = base64.b64decode(result['data'])
+
+                # 写入本地 PDF 文件（二进制）
+                with open(output_path, 'wb') as f:
+                    f.write(pdf_data)
+
+                file_size = os.path.getsize(output_path)
+                if file_size == 0:
+                    raise RuntimeError("生成了空文件")
+
+                logger.info(f"✅ 保存成功: {os.path.abspath(output_path)} ({file_size} 字节)")
+                return os.path.abspath(output_path)
+
+            except (WebDriverException, InvalidSessionIdException, SessionNotCreatedException) as e:
+                last_exception = e
+                logger.warning(f"WebDriver 异常 (retry {retry + 1}): {e}")
+                if retry < self.max_retries:
+                    logger.info("正在重建 WebDriver 会话...")
+                    self._init_driver()
+                    time.sleep(self.retry_delay)
+                else:
+                    logger.error("达到最大重试次数，放弃")
+                    break
+
+            except TimeoutException as e:
+                last_exception = e
+                logger.warning(f"页面加载超时 (retry {retry + 1}): {e}")
+                if retry < self.max_retries:
+                    time.sleep(self.retry_delay)
+                else:
+                    break
+
+            except Exception as e:
+                last_exception = e
+                logger.error(f"未知错误 (retry {retry + 1}): {e}")
+                break  # 非 WebDriver 错误，不重试
+
+        # 清理失败生成的空文件
+        if os.path.exists(output_path):
+            try:
+                os.remove(output_path)
+            except OSError:
+                pass
+
+        raise RuntimeError(f"保存失败（{type(last_exception).__name__}）: {last_exception}")
+
+    def quit(self):
+        """显式关闭浏览器"""
+        if self.driver:
+            try:
+                self.driver.quit()
+                logger.info("WebDriver 会话已关闭")
+            except Exception:
+                pass
+            self.driver = None
+
+    def __del__(self):
+        self.quit()
+
+
+# ===== 测试 =====
+if __name__ == "__main__":
+    saver = RemotePDFSaver(
+        remote_url="http://144.34.185.108:28098/wd/hub",  # ← 替换为你的云服务器公网 IP
+        headless=True
+    )
+    try:
+        saver.save_as_pdf(
+            url="https://www.epochtimes.com/gb/25/12/22/n14660274.htm",
+            output_path="remote_example2.pdf"
+        )
+    except Exception as e:
+        print(f"❌ 失败: {e}")
+
+    saver.quit()