mdbq 4.0.71__py3-none-any.whl → 4.0.73__py3-none-any.whl

mdbq/__version__.py

@@ -1 +1 @@
-VERSION = '4.0.71'
+VERSION = '4.0.73'

mdbq/selenium/get_driver.py

@@ -9,6 +9,8 @@ import socket
 import tempfile
 import shutil
 import uuid
+import subprocess
+import json
 
 dir_path = os.path.expanduser("~")
 
@@ -52,54 +54,6 @@ class GetDriver:
             self.user_agent = user_agents[random.randint(0, len(user_agents) - 1)]
         self.maximize_window = maximize_window
 
-    def __enter__(self):
-        """
-        支持with语法自动获取driver
-        :return: selenium.webdriver.Chrome实例
-        """
-        self.driver = self.getdriver()
-        return self.driver
-
-    def __exit__(self, exc_type, exc_val, exc_tb):
-        """
-        支持with语法自动清理资源
-        """
-        self.quit()
-
-    def close(self):
-        """
-        关闭浏览器窗口并清理临时目录
-        """
-        if self.driver:
-            try:
-                self.driver.close()
-            except:
-                pass
-        self._cleanup_temp_dirs()
-
-    def quit(self):
-        """
-        彻底退出浏览器并清理临时目录
-        """
-        if self.driver:
-            try:
-                self.driver.quit()
-            except:
-                pass
-        self._cleanup_temp_dirs()
-
-    def _cleanup_temp_dirs(self):
-        """
-        清理所有创建的临时目录
-        """
-        for temp_dir in self.temp_dirs:
-            try:
-                if os.path.exists(temp_dir):
-                    shutil.rmtree(temp_dir)
-            except:
-                pass
-        self.temp_dirs = []
-
     def check_proxy(self):
         """
         校验代理格式和连通性，支持http/https/socks5
@@ -119,14 +73,177 @@ class GetDriver:
         except:
             return False
 
+    def _get_chrome_version(self, chrome_path):
+        """
+        获取Chrome版本号
+        :param chrome_path: Chrome可执行文件路径
+        :return: 版本号字符串，如"120.0.6099.109"
+        """
+        try:
+            if platform.system().lower() == 'windows':
+                # Windows下尝试多种方式获取版本
+                # 方法1: 尝试--version参数
+                try:
+                    result = subprocess.run([chrome_path, '--version'], 
+                                          capture_output=True, text=True, timeout=10, shell=True)
+                    if result.returncode == 0:
+                        version_match = re.search(r'Chrome\s+(\d+\.\d+\.\d+\.\d+)', result.stdout)
+                        if version_match:
+                            return version_match.group(1)
+                except:
+                    pass
+                
+                # 方法2: 尝试从注册表获取版本
+                try:
+                    import winreg
+                    key_path = r"SOFTWARE\Google\Chrome\BLBeacon"
+                    with winreg.OpenKey(winreg.HKEY_CURRENT_USER, key_path) as key:
+                        version = winreg.QueryValueEx(key, "version")[0]
+                        return version
+                except:
+                    pass
+                
+                # 方法3: 尝试从文件属性获取版本
+                try:
+                    result = subprocess.run(['wmic', 'datafile', 'where', f'name="{chrome_path.replace("/", "\\")}"', 'get', 'version', '/value'], 
+                                          capture_output=True, text=True, timeout=10, shell=True)
+                    if result.returncode == 0:
+                        version_match = re.search(r'Version=(\d+\.\d+\.\d+\.\d+)', result.stdout)
+                        if version_match:
+                            return version_match.group(1)
+                except:
+                    pass
+                
+                # 方法4: 尝试直接启动Chrome获取版本信息
+                try:
+                    result = subprocess.run([chrome_path, '--headless', '--disable-gpu', '--dump-dom', 'about:version'], 
+                                          capture_output=True, text=True, timeout=15, shell=True)
+                    if result.returncode == 0:
+                        version_match = re.search(r'Chrome/(\d+\.\d+\.\d+\.\d+)', result.stdout)
+                        if version_match:
+                            return version_match.group(1)
+                except:
+                    pass
+                    
+            else:
+                # macOS和Linux下使用--version参数
+                result = subprocess.run([chrome_path, '--version'], 
+                                      capture_output=True, text=True, timeout=10)
+                if result.returncode == 0:
+                    # 输出格式: "Google Chrome 120.0.6099.109"
+                    version_match = re.search(r'Chrome\s+(\d+\.\d+\.\d+\.\d+)', result.stdout)
+                    if version_match:
+                        return version_match.group(1)
+        except Exception as e:
+            print(f"获取Chrome版本失败: {e}")
+        return None
+
+    def _get_chromedriver_version(self, chromedriver_path):
+        """
+        获取Chromedriver版本号
+        :param chromedriver_path: Chromedriver可执行文件路径
+        :return: 版本号字符串，如"120.0.6099.109"
+        """
+        try:
+            if platform.system().lower() == 'windows':
+                # Windows下使用shell=True确保参数正确传递
+                result = subprocess.run([chromedriver_path, '--version'], 
+                                      capture_output=True, text=True, timeout=10, shell=True)
+            else:
+                result = subprocess.run([chromedriver_path, '--version'], 
+                                      capture_output=True, text=True, timeout=10)
+            
+            if result.returncode == 0:
+                # 输出格式: "ChromeDriver 120.0.6099.109"
+                version_match = re.search(r'ChromeDriver\s+(\d+\.\d+\.\d+\.\d+)', result.stdout)
+                if version_match:
+                    return version_match.group(1)
+        except Exception as e:
+            print(f"获取Chromedriver版本失败: {e}")
+        return None
+
+    def _check_version_compatibility(self, chrome_path, chromedriver_path):
+        """
+        检查Chrome和Chromedriver版本兼容性
+        :param chrome_path: Chrome可执行文件路径
+        :param chromedriver_path: Chromedriver可执行文件路径
+        :return: (is_compatible, chrome_version, chromedriver_version)
+        """
+        chrome_version = self._get_chrome_version(chrome_path)
+        chromedriver_version = self._get_chromedriver_version(chromedriver_path)
+        
+        # 如果无法获取版本信息，返回True允许尝试启动
+        if not chrome_version or not chromedriver_version:
+            print(f"警告: 无法获取版本信息 - Chrome: {chrome_version}, Chromedriver: {chromedriver_version}")
+            return True, chrome_version, chromedriver_version
+        
+        # 提取主版本号进行比较
+        chrome_major = chrome_version.split('.')[0]
+        chromedriver_major = chromedriver_version.split('.')[0]
+        
+        is_compatible = chrome_major == chromedriver_major
+        return is_compatible, chrome_version, chromedriver_version
+
+    def _try_create_driver(self, chrome_path, chromedriver_path, option, temp_dir):
+        """
+        尝试创建Chrome WebDriver实例
+        :param chrome_path: Chrome可执行文件路径
+        :param chromedriver_path: Chromedriver可执行文件路径
+        :param option: ChromeOptions实例
+        :param temp_dir: 临时目录路径
+        :return: Chrome WebDriver实例或None
+        """
+        try:
+            option.binary_location = chrome_path
+            service = Service(chromedriver_path)
+            driver = webdriver.Chrome(service=service, options=option)
+            if self.maximize_window:
+                driver.maximize_window()
+            
+            # --- 防反爬：注入多段JS隐藏Selenium特征 ---
+            js_hide_features = [
+                # 隐藏webdriver属性
+                "Object.defineProperty(navigator, 'webdriver', {get: () => undefined, configurable: true});",
+                # 模拟真实浏览器插件
+                "Object.defineProperty(navigator, 'plugins', {get: () => [1,2,3,4,5], configurable: true});",
+                # 设置语言
+                "Object.defineProperty(navigator, 'languages', {get: () => ['zh-CN', 'zh', 'en'], configurable: true});",
+                # 模拟Chrome运行时
+                "window.chrome = {runtime: {}, loadTimes: function(){}, csi: function(){}, app: {}};",
+                # 删除原型链上的webdriver
+                "delete window.navigator.__proto__.webdriver;",
+                # 删除Selenium相关属性
+                r"for (let key in window) {if (key.match(/^[\$\_]{3,}/)) {try {delete window[key];} catch(e){}}}",
+                # 隐藏自动化相关属性
+                "Object.defineProperty(navigator, 'permissions', {get: () => ({query: () => Promise.resolve({state: 'granted'})}), configurable: true});",
+                # 模拟真实的navigator属性
+                "Object.defineProperty(navigator, 'hardwareConcurrency', {get: () => 8, configurable: true});",
+                "Object.defineProperty(navigator, 'deviceMemory', {get: () => 8, configurable: true});",
+                # 防止检测自动化工具
+                "Object.defineProperty(navigator, 'maxTouchPoints', {get: () => 0, configurable: true});",
+                # 隐藏CDP相关属性
+                "delete window.cdc_adoQpoasnfa76pfcZLmcfl_Array;",
+                "delete window.cdc_adoQpoasnfa76pfcZLmcfl_Promise;",
+                "delete window.cdc_adoQpoasnfa76pfcZLmcfl_Symbol;"
+            ]
+            for js in js_hide_features:
+                driver.execute_cdp_cmd("Page.addScriptToEvaluateOnNewDocument", {"source": js})
+            
+            return driver
+        except Exception as e:
+            print(f"创建Chrome WebDriver失败: {e}")
+            return None
+
     def getdriver(self):
         """
         创建并返回Chrome WebDriver实例，自动注入反检测JS，异常时抛出GetDriverException
+        智能版本检测：优先使用正式版，版本不匹配时自动切换到测试版
         :return: selenium.webdriver.Chrome实例
         :raises: GetDriverException
         """
         if not self.check_proxy():
             raise GetDriverException(f"代理不可用或格式错误: {self.proxy}")
+        
         option = webdriver.ChromeOptions()  # 浏览器启动选项
         if self.headless:
             option.add_argument("--headless")  # 设置无界面模式
@@ -134,16 +251,19 @@ class GetDriver:
         option.add_argument("--disable-gpu")
         option.add_argument("--no-sandbox")
         option.add_argument("--disable-dev-shm-usage")
+        # 隐藏Chrome测试版提示信息
+        option.add_argument("--disable-blink-features=AutomationControlled")
+        option.add_argument("--disable-features=VizDisplayCompositor")
+        option.add_argument("--disable-background-timer-throttling")
+        option.add_argument("--disable-backgrounding-occluded-windows")
+        option.add_argument("--disable-renderer-backgrounding")
+        option.add_argument("--disable-features=TranslateUI")
+        option.add_argument("--disable-ipc-flooding-protection")
         # 添加唯一的用户数据目录，避免Chrome实例冲突
         temp_dir = tempfile.mkdtemp(prefix=f'chrome_automation_{uuid.uuid4().hex[:8]}_')
         option.add_argument(f'--user-data-dir={temp_dir}')
         option.add_argument('--no-first-run')
         option.add_argument('--no-default-browser-check')
-        option.add_argument('--disable-background-timer-throttling')
-        option.add_argument('--disable-backgrounding-occluded-windows')
-        option.add_argument('--disable-renderer-backgrounding')
-        option.add_argument('--disable-features=TranslateUI')
-        option.add_argument('--disable-ipc-flooding-protection')
         # 关键安全浏览禁用参数
         option.add_argument('--allow-insecure-localhost')
         option.add_argument('--allow-running-insecure-content')
@@ -177,91 +297,171 @@ class GetDriver:
             "profile.password_manager_enabled": False,
             "download_restrictions": 0,
         }
+        option.add_experimental_option("prefs", prefs)
+        
         # 平台与路径自动检测
         sys_platform = platform.system().lower()
         chrome_path = self.chrome_path
         chromedriver_path = self.chromedriver_path
+        
         try:
             if sys_platform == 'windows':
                 if not chrome_path:
-                    chrome_path = os.path.join(f'C:\\Users\\{getpass.getuser()}', 'chrome\\chrome_win64\\chrome.exe')
+                    chrome_path_candidates = [
+                        'C:\\Program Files\\Google\\Chrome\\Application\\chrome.exe',  # 正式版
+                        os.path.join(f'C:\\Users\\{getpass.getuser()}', 'chrome\\chrome_win64\\chrome.exe'),  # 测试版
+                    ]
                 if not chromedriver_path:
-                    chromedriver_path = os.path.join(f'C:\\Users\\{getpass.getuser()}', 'chrome\\chromedriver.exe')
-                option.binary_location = chrome_path
-                service = Service(chromedriver_path)
+                    chromedriver_path_candidates = [
+                        os.path.join(f'C:\\Users\\{getpass.getuser()}', 'chrome\\chromedriver.exe'),
+                        os.path.join(f'C:\\Users\\{getpass.getuser()}', 'chrome\\chrome_win64\\chromedriver.exe'),
+                    ]
             elif sys_platform == 'linux':
                 if not chrome_path:
-                    chrome_path = '/usr/bin/chrome/chrome'
-                    """
-                    # sudo mv /usr/bin/google-chrome /usr/bin/google-chrome.bak  # 备份原有
-                    # sudo ln -s /usr/bin/chrome /usr/bin/google-chrome # 创建软链接
-                    """
+                    chrome_path_candidates = [
+                        '/usr/bin/google-chrome',  # 正式版
+                        '/usr/bin/chrome/chrome',  # 测试版
+                    ]
                 if not chromedriver_path:
-                    chromedriver_path = '/usr/local/bin/chromedriver'
-                option.binary_location = chrome_path
-                service = Service(chromedriver_path)
+                    chromedriver_path_candidates = [
+                        '/usr/local/bin/chromedriver',
+                        '/usr/bin/chromedriver',
+                    ]
             elif sys_platform == 'darwin':
                 if not chrome_path:
-                    # 优先使用用户指定的默认路径
                     chrome_path_candidates = [
-                        '/usr/local/chrome/Google Chrome for Testing.app/Contents/MacOS/Google Chrome',
-                        '/usr/local/chrome/Google Chrome for Testing.app',
-                        '/Applications/Google Chrome.app/Contents/MacOS/Google Chrome',
-                        '/Applications/Google Chrome for Testing.app/Contents/MacOS/Google Chrome',
+                        '/Applications/Google Chrome.app/Contents/MacOS/Google Chrome',  # 正式版
+                        '/usr/local/chrome/Google Chrome for Testing.app/Contents/MacOS/Google Chrome for Testing',  # 测试版
                     ]
-                    chrome_path = next((p for p in chrome_path_candidates if os.path.exists(p)), None)
                 if not chromedriver_path:
                     chromedriver_path_candidates = [
                         '/usr/local/chrome/chromedriver',
                         '/usr/local/bin/chromedriver',
                         '/opt/homebrew/bin/chromedriver',
                     ]
-                    chromedriver_path = next((p for p in chromedriver_path_candidates if os.path.exists(p)), None)
-                if not chrome_path or not chromedriver_path:
-                    raise GetDriverException("未找到Chrome或Chromedriver，请手动指定chrome_path和chromedriver_path")
-                # option.binary_location = chrome_path  # macOS 设置此参数报错
-                service = Service(chromedriver_path)
             else:
                 raise GetDriverException(f"不支持的平台: {sys_platform}")
+            
+            # 如果用户指定了路径，直接使用
+            if chrome_path and chromedriver_path:
+                driver = self._try_create_driver(chrome_path, chromedriver_path, option, temp_dir)
+                if driver:
+                    self.temp_dirs.append(temp_dir)
+                    self.driver = driver
+                    return driver
+                else:
+                    raise GetDriverException(f"指定的Chrome路径无法启动: {chrome_path}")
+            
+            # 智能版本检测和切换
+            chrome_paths = [p for p in chrome_path_candidates if os.path.exists(p)]
+            chromedriver_paths = [p for p in chromedriver_path_candidates if os.path.exists(p)]
+            
+            if not chrome_paths:
+                raise GetDriverException("未找到Chrome浏览器，请手动指定chrome_path")
+            if not chromedriver_paths:
+                raise GetDriverException("未找到Chromedriver，请手动指定chromedriver_path")
+            
+            # 优先尝试正式版Chrome
+            for chrome_path in chrome_paths:
+                for chromedriver_path in chromedriver_paths:
+                    # 检查版本兼容性
+                    is_compatible, chrome_version, chromedriver_version = self._check_version_compatibility(chrome_path, chromedriver_path)
+                    
+                    if is_compatible:
+                        # print(f"版本兼容: Chrome {chrome_version}, Chromedriver {chromedriver_version}")
+                        driver = self._try_create_driver(chrome_path, chromedriver_path, option, temp_dir)
+                        if driver:
+                            self.temp_dirs.append(temp_dir)
+                            self.driver = driver
+                            return driver
+                    else:
+                        print(f"版本不兼容: Chrome {chrome_version}, Chromedriver {chromedriver_version}")
+                        # 即使版本不兼容也尝试启动，有时可能仍然可以工作
+                        driver = self._try_create_driver(chrome_path, chromedriver_path, option, temp_dir)
+                        if driver:
+                            print("警告：版本不兼容但启动成功，建议更新Chromedriver")
+                            self.temp_dirs.append(temp_dir)
+                            self.driver = driver
+                            return driver
+            
+            # 如果所有组合都失败，抛出异常
+            raise GetDriverException("所有Chrome和Chromedriver组合都无法启动，请检查版本兼容性")
+            
         except Exception as e:
-            raise GetDriverException(f"浏览器路径配置异常: {e}")
-        option.add_experimental_option("prefs", prefs)
-        try:
-            driver = webdriver.Chrome(service=service, options=option)
-            if self.maximize_window:
-                driver.maximize_window()
-            # --- 防反爬：注入多段JS隐藏Selenium特征 ---
-            js_hide_features = [
-                "Object.defineProperty(navigator, 'webdriver', {get: () => false});",
-                "Object.defineProperty(navigator, 'plugins', {get: () => [1,2,3,4,5]});",
-                "Object.defineProperty(navigator, 'languages', {get: () => ['zh-CN', 'zh', 'en']});",
-                "window.chrome = {runtime: {}};",
-                "delete window.navigator.__proto__.webdriver;",
-                r"for (let key in window) {if (key.match(/^[\$\_]{3,}/)) {try {delete window[key];} catch(e){}}}"
-            ]
-            for js in js_hide_features:
+            try:
+                if os.path.exists(temp_dir):
+                    shutil.rmtree(temp_dir)
+            except Exception as cleanup_error:
                 pass
-                driver.execute_cdp_cmd("Page.addScriptToEvaluateOnNewDocument", {"source": js})
-            self.temp_dirs.append(temp_dir)
-            self.driver = driver
-            return driver
-        except Exception as e:
+            if isinstance(e, GetDriverException):
+                raise e
+            else:
+                raise GetDriverException(f"启动ChromeDriver失败: {e}")
+
+    def _cleanup_temp_dirs(self):
+        """
+        清理所有创建的临时目录
+        """
+        for temp_dir in self.temp_dirs:
             try:
                 if os.path.exists(temp_dir):
                     shutil.rmtree(temp_dir)
-            except Exception as e:
+            except:
                 pass
-            raise GetDriverException(f"启动ChromeDriver失败: {e}")
+        self.temp_dirs = []
+
+    def __enter__(self):
+        """
+        支持with语法自动获取driver
+        :return: selenium.webdriver.Chrome实例
+        """
+        self.driver = self.getdriver()
+        return self.driver
+
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        """
+        支持with语法自动清理资源
+        """
+        self.quit()
+
+    def close(self):
+        """
+        关闭浏览器窗口并清理临时目录
+        """
+        if self.driver:
+            try:
+                self.driver.close()
+            except:
+                pass
+        self._cleanup_temp_dirs()
+
+    def quit(self):
+        """
+        彻底退出浏览器并清理临时目录
+        """
+        if self.driver:
+            try:
+                self.driver.quit()
+            except:
+                pass
+        self._cleanup_temp_dirs()
 
 
 if __name__ == '__main__':
-    with GetDriver(
-        headless=True,
-        proxy=None,  # 代理（'socks5://127.0.0.1:1080'）
-        user_agent=None,
-        download_dir=None, 
-        chrome_path=None, 
-        chromedriver_path=None,
-    ) as driver:
-        driver.get('https://www.baidu.com')
-        print(driver.title)
+    # with GetDriver(
+    #     headless=True,
+    #     proxy=None,  # 代理（'socks5://127.0.0.1:1080'）
+    #     user_agent=None,
+    #     download_dir=None, 
+    #     chrome_path=None, 
+    #     chromedriver_path=None,
+    # ) as driver:
+    #     driver.get('https://www.baidu.com')
+    #     print(driver.title)
+
+    
+    driver = GetDriver(headless=False).getdriver()
+    driver.get('https://www.baidu.com')
+    print(driver.title)
+    import time
+    time.sleep(1000)

{mdbq-4.0.71.dist-info → mdbq-4.0.73.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: mdbq
-Version: 4.0.71
+Version: 4.0.73
 Home-page: https://pypi.org/project/mdbq
 Author: xigua, 
 Author-email: 2587125111@qq.com

{mdbq-4.0.71.dist-info → mdbq-4.0.73.dist-info}/RECORD

@@ -1,5 +1,5 @@
 mdbq/__init__.py,sha256=Il5Q9ATdX8yXqVxtP_nYqUhExzxPC_qk_WXQ_4h0exg,16
-mdbq/__version__.py,sha256=YSGjE28atlTDM_O3TrBB0glPwobYMIp1lwetQ7I89w4,18
+mdbq/__version__.py,sha256=v9CxLcSDpoYTepVqra_1iaKkOd7scJSGR8ITdcM5Nr4,18
 mdbq/log/__init__.py,sha256=Mpbrav0s0ifLL7lVDAuePEi1hJKiSHhxcv1byBKDl5E,15
 mdbq/log/mylogger.py,sha256=kPe3wsQNaB1slfX-Z7VMqzZoMoqPfc7ylYXZDBeFzzI,24945
 mdbq/myconf/__init__.py,sha256=jso1oHcy6cJEfa7udS_9uO5X6kZLoPBF8l3wCYmr5dM,18
@@ -22,9 +22,9 @@ mdbq/pbix/refresh_all.py,sha256=OBT9EewSZ0aRS9vL_FflVn74d4l2G00wzHiikCC4TC0,5926
 mdbq/redis/__init__.py,sha256=YtgBlVSMDphtpwYX248wGge1x-Ex_mMufz4-8W0XRmA,12
 mdbq/redis/getredis.py,sha256=vpBuNc22uj9Vr-_Dh25_wpwWM1e-072EAAIBdB_IpL0,23494
 mdbq/selenium/__init__.py,sha256=AKzeEceqZyvqn2dEDoJSzDQnbuENkJSHAlbHAD0u0ZI,10
-mdbq/selenium/get_driver.py,sha256=MGZJKfu_73CSxXqeW6DmPAh2UNyJfAmI5Zb9AGa1YXY,11694
+mdbq/selenium/get_driver.py,sha256=ppV_CJvvRu2__VIYQS-nfhW0lOlesuR11bG_d2MfmNs,21523
 mdbq/spider/__init__.py,sha256=RBMFXGy_jd1HXZhngB2T2XTvJqki8P_Fr-pBcwijnew,18
-mdbq-4.0.71.dist-info/METADATA,sha256=pr4vwkvALKhNNHKwwcwYiA4-2dmBqVL0t0agvzg4KDc,364
-mdbq-4.0.71.dist-info/WHEEL,sha256=jB7zZ3N9hIM9adW7qlTAyycLYW9npaWKLRzaoVcLKcM,91
-mdbq-4.0.71.dist-info/top_level.txt,sha256=2FQ-uLnCSB-OwFiWntzmwosW3X2Xqsg0ewh1axsaylA,5
-mdbq-4.0.71.dist-info/RECORD,,
+mdbq-4.0.73.dist-info/METADATA,sha256=i2uf1d6ZPcI60XOSOGXW6wgePbz_pHsumb0-LGJD8Ns,364
+mdbq-4.0.73.dist-info/WHEEL,sha256=jB7zZ3N9hIM9adW7qlTAyycLYW9npaWKLRzaoVcLKcM,91
+mdbq-4.0.73.dist-info/top_level.txt,sha256=2FQ-uLnCSB-OwFiWntzmwosW3X2Xqsg0ewh1axsaylA,5
+mdbq-4.0.73.dist-info/RECORD,,