Jarvis-Brain 0.1.7.7__py3-none-any.whl

jarvis_brain-0.1.7.7.dist-info/METADATA

@@ -0,0 +1,10 @@
+Metadata-Version: 2.4
+Name: Jarvis_Brain
+Version: 0.1.7.7
+Summary: Jarvis brain mcp
+Requires-Python: >=3.10
+Requires-Dist: beautifulsoup4
+Requires-Dist: curl-cffi
+Requires-Dist: drissionpage
+Requires-Dist: fastmcp
+Requires-Dist: minify-html

jarvis_brain-0.1.7.7.dist-info/RECORD

@@ -0,0 +1,11 @@
+mcp_tools/__init__.py,sha256=nqhlHRalTYQF8gU5RBfanOn-zSONLYON1mjPIvP-f4w,113
+mcp_tools/dp_tools.py,sha256=O51cz2ectsbJJ6fN9X1YarSC7JereXeUWFtehx71E3Q,10796
+mcp_tools/main.py,sha256=Fdt2N3oKGwvruuno_ywnuWSlm1BexE9ZY669H2LTo9w,1056
+tools/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+tools/browser_manager.py,sha256=EXM7n-sDOtdQGpWkVTAZHWhepVU-7PAoUTDNgGF9_fQ,1938
+tools/browser_proxy.py,sha256=cdMRxcUYyaOqGU17lldltHOvt9rxXD5Dwh7hBXEBby4,6780
+tools/tools.py,sha256=TaWs-CNXy-py9BFmCnJrQ09ke938xXpImf-N2Qo_Rvc,4708
+jarvis_brain-0.1.7.7.dist-info/METADATA,sha256=u_pB128wsC-0P2-hKsYlgkZFaDZ45dCoM81ZdnZDRdU,241
+jarvis_brain-0.1.7.7.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
+jarvis_brain-0.1.7.7.dist-info/entry_points.txt,sha256=YFQT4xpkUqt5dM5wlKPQQOqcjMuFrT9iuRAzIpAyH7U,51
+jarvis_brain-0.1.7.7.dist-info/RECORD,,

jarvis_brain-0.1.7.7.dist-info/WHEEL

@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: hatchling 1.28.0
+Root-Is-Purelib: true
+Tag: py3-none-any

jarvis_brain-0.1.7.7.dist-info/entry_points.txt

@@ -0,0 +1,2 @@
+[console_scripts]
+jarvis-mcp = mcp_tools.main:main

mcp_tools/__init__.py

@@ -0,0 +1,4 @@
+a = ["a", "b", "c", "d", "e", "f", "g", "h"]
+print(len(a))
+for index, num in enumerate(a):
+    print(index, num)

mcp_tools/dp_tools.py

@@ -0,0 +1,197 @@
+import hashlib
+import json
+import os
+from typing import Any
+
+from fastmcp import FastMCP
+
+from tools.browser_manager import BrowserManager
+from tools.tools import compress_html, requests_html, dp_headless_html, assert_waf_cookie, dp_mcp_message_pack
+from tools.browser_proxy import DPProxyClient, DPProxyClientManager
+
+html_source_code_local_save_path = os.path.join(os.getcwd(), "html-source-code")
+waf_status_code_dict = {
+    412: "瑞数",
+    521: "加速乐"
+}
+# 一轮最大输入，以免单个html最大长度超过ai最大输入
+one_turn_max_token = 20000
+
+
+def register_visit_url(mcp: FastMCP, browser_manager: BrowserManager, client_manager: DPProxyClientManager):
+    @mcp.tool(name="visit_url",
+              description="使用Drissionpage打开url访问某个网站，并开始监听初始tab页的所有的XHR请求，当需要使用手机版浏览器Ua时use_mobile_user_agent为True")
+    async def visit_url(url: str, use_mobile_user_agent=False) -> dict[str, Any]:
+        mobile_user_agent = None
+        if use_mobile_user_agent:
+            mobile_user_agent = "Mozilla/5.0 (Linux; Android 6.0; Nexus 5 Build/MRA58N) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/143.0.0.0 Mobile Safari/537.36"
+        port, _browser = browser_manager.create_browser(mobile_user_agent)
+        tab = _browser.get_tab()
+        client_manager.create_client(tab)
+        tab.get(url)
+        tab_id = tab.tab_id
+        return dp_mcp_message_pack(
+            f"已在[{port}]端口创建浏览器对象，并已打开链接：{url}，打开的模式是：{'手机版' if use_mobile_user_agent else '电脑版'}",
+            tab_id=tab_id,
+            browser_port=port
+        )
+
+
+def register_get_html(mcp: FastMCP, browser_manager, client_manager: DPProxyClientManager):
+    @mcp.tool(name="get_html", description="使用Drissionpage获取某一个tab页的html")
+    async def get_html(browser_port: int, tab_id: str) -> dict[str, Any]:
+        _browser = browser_manager.get_browser(browser_port)
+        tab = _browser.get_tab(tab_id)
+        file_name_prefix = hashlib.md5(str(tab.title).encode('utf-8')).hexdigest()
+        if not os.path.exists(html_source_code_local_save_path):
+            os.makedirs(html_source_code_local_save_path)
+        min_html, compress_rate = compress_html(tab.html)
+        html_str_list = [min_html[i:i + one_turn_max_token] for i in range(0, len(min_html), one_turn_max_token)]
+        html_file_list = []
+        for index, html_str in enumerate(html_str_list):
+            file_name = file_name_prefix + f"_{tab_id}_segment{index}.html"
+            abs_path = os.path.join(html_source_code_local_save_path, file_name)
+            with open(abs_path, "w", encoding="utf-8") as f:
+                f.write(html_str)
+            html_file_list.append(abs_path)
+        message = f"已保存tab页：【{tab_id}】的html源码片段共{len(html_file_list)}个"
+        return dp_mcp_message_pack(message, tab_id=tab_id, htmls_local_path=html_file_list)
+
+
+def register_get_new_tab(mcp: FastMCP, browser_manager, client_manager: DPProxyClientManager):
+    @mcp.tool(name="get_new_tab",
+              description="使用Drissionpage创建一个新的tab页，在新的tab页中打开url，并开始监听新的tab页的所有XHR请求")
+    async def get_new_tab(browser_port: int, url: str) -> dict[str, Any]:
+        _browser = browser_manager.get_browser(browser_port)
+        tab = _browser.new_tab()
+        client_manager.create_client(tab)
+        tab.get(url)
+        _browser.activate_tab(tab)
+        tab_id = tab.tab_id
+        return dp_mcp_message_pack(f"已创建新的tab页，并打开链接：{url}", tab_id=tab_id)
+
+
+def register_switch_tab(mcp: FastMCP, browser_manager, client_manager: DPProxyClientManager):
+    @mcp.tool(name="switch_tab", description="根据传入的tab_id切换到对应的tab页", )
+    async def switch_tab(browser_port: int, tab_id: str) -> dict[str, Any]:
+        _browser = browser_manager.get_browser(browser_port)
+        _browser.activate_tab(tab_id)
+        return dp_mcp_message_pack(f"已将tab页:【{tab_id}】切换至最前端")
+
+
+def register_close_tab(mcp: FastMCP, browser_manager, client_manager: DPProxyClientManager):
+    @mcp.tool(name="close_tab", description="根据传入的tab_id关闭tab页", )
+    async def close_tab(browser_port, tab_id) -> dict[str, Any]:
+        _browser = browser_manager.get_browser(browser_port)
+        _browser.close_tabs(tab_id)
+        return dp_mcp_message_pack(f"已将tab页:【{tab_id}】关闭")
+
+
+def register_check_selector(mcp: FastMCP, browser_manager, client_manager: DPProxyClientManager):
+    @mcp.tool(name="check_selector", description="查找tab页中是否包含元素，并返回元素attr_name所对应的值")
+    async def check_selector(browser_port: int, tab_id: str, css_selector: str, attr_name: str = "text") -> dict[
+        str, Any]:
+        _browser = browser_manager.get_browser(browser_port)
+        target_tab = _browser.get_tab(tab_id)
+        css_selector = css_selector
+        if "css:" not in css_selector:
+            css_selector = "css:" + css_selector
+        target_eles = target_tab.eles(css_selector)
+        exist_flag = False
+        if len(target_eles) != 0:
+            exist_flag = True
+        if attr_name == "text":
+            ele_text_list = [i.text.replace("\n", "") for i in target_eles]
+            attr_output = "\n".join(ele_text_list)
+        else:
+            attr_output = json.dumps([i.attr(attr_name) for i in target_eles])
+        # 对attr_output逐个截断，截断的长度为：一轮最大token除以元素个数+6个点
+        attr_output = [attr_str[:one_turn_max_token // (len(attr_str) + 6)] + "......" for attr_str in attr_output]
+        return dp_mcp_message_pack(
+            f"已完成tab页:【{tab_id}】对：【{css_selector}】的检查",
+            tab_id=tab_id,
+            selector=css_selector,
+            selector_ele_exist=exist_flag,
+            attr_output=attr_output
+        )
+
+
+def register_quit_browser(mcp: FastMCP, browser_manager, client_manager: DPProxyClientManager):
+    @mcp.tool(name="quit_browser", description="退出浏览器会话，关闭浏览器")
+    async def quit_browser(browser_port: int) -> dict[str, Any]:
+        flag, _browser = browser_manager.remove_page(browser_port)
+        if flag:
+            _browser.quit()
+        return dp_mcp_message_pack(
+            f"浏览器[{browser_port}]，退出会话，关闭浏览器{'成功' if flag else '失败'}",
+            browser_port=browser_port,
+            quit_flag=flag
+        )
+
+
+def register_pop_first_packet(mcp: FastMCP, browser_manager, client_manager: DPProxyClientManager):
+    @mcp.tool(name="pop_first_packet",
+              description="每调用一次就会弹出传入的tab页所监听到的数据包中的第一个packet_message，当一个packet_message的response body过长时会被切分成多个包，具体一个请求是否还有下一个包，可以参考body_completed字段")
+    async def pop_first_packet(browser_port: int, tab_id: str) -> dict[str, Any]:
+        _browser = browser_manager.get_browser(browser_port)
+        client = client_manager.get_client(tab_id)
+        packet_message = client.pop_first_packet()
+        message = f"tab页:【{tab_id}】，暂时没有监听到XHR数据包"
+        if packet_message:
+            message = f"tab页:【{tab_id}】，监听到XHR数据包",
+        return dp_mcp_message_pack(
+            message,
+            browser_port=browser_port,
+            tab_id=tab_id,
+            packet_message=packet_message
+        )
+
+
+def register_assert_waf(mcp: FastMCP, browser_manager, client_manager: DPProxyClientManager):
+    @mcp.tool(name="assert_waf",
+              description="通过对比requests、有头浏览器、无头浏览器获取到的html，判断网页是否使用了waf以及是否为动态渲染的网页")
+    async def assert_waf(browser_port: int, tab_id: str) -> dict[str, Any]:
+        _browser = browser_manager.get_browser(browser_port)
+        target_tab = _browser.get_tab(tab_id)
+        recommend_team = "drissionpage_head"
+        head_cookies = target_tab.cookies()
+        # 通过cookie判断是否有waf
+        waf_flag, waf_type = assert_waf_cookie(head_cookies)
+        head_html = target_tab.html
+        min_head_html, head_rate = compress_html(head_html, only_text=True)
+        raw_html, status_code = requests_html(target_tab.url)
+        min_raw_html, raw_rate = compress_html(raw_html, only_text=True)
+        r_h_rate_diff = abs(head_rate - raw_rate)
+        # 如果有已知的防火墙，则不浪费时间使用无头获取html和压缩比了
+        if waf_flag or status_code in waf_status_code_dict.keys():
+            return dp_mcp_message_pack(
+                f"已完成tab页:【{tab_id}】的分析，该tab页存在waf",
+                tab_id=tab_id,
+                recommend_team=recommend_team,
+                raw_head_rate_difference=r_h_rate_diff,
+                raw_headless_rate_difference=0,
+                head_headless_rate_difference=0
+            )
+
+        headless_html = dp_headless_html(target_tab.url)
+        min_headless_html, headless_rate = compress_html(headless_html, only_text=True)
+        r_hless_rate_diff = abs(raw_rate - headless_rate)
+        h_hless_rate_diff = abs(head_rate - headless_rate)
+        # 最优情况：requests，dp有头和无头拿到的结果基本一致，认定为没有防护的静态网页
+        if r_h_rate_diff < 40 and r_hless_rate_diff < 40 and h_hless_rate_diff < 40:
+            recommend_team = "requests"
+        # 最差情况：requests，dp有头和无头拿到的结果差距都很大，认定为有浏览器无头检测+动态网页
+        # if r_h_rate_diff < 40 and r_hless_rate_diff < 40 and h_hless_rate_diff < 40:
+        # 较差1：dp有头和无头差距很小，但是requests拿不到正确结果，认定为有requests防护 or 动态网页
+        elif h_hless_rate_diff < 30 and r_hless_rate_diff > 40:
+            recommend_team = "drissionpage_headless"
+        # 较差2：有头和无头差距很大，但是requests和有头拿到的结果基本一致，认定为要么有别的没有防护requests的waf，或者间歇性的瑞数【此时应该拿有头的cookie去判断其中是否有瑞数特征，上面已经做了】
+        # if r_h_rate_diff < 15 and h_hless_rate_diff > 40:
+        return dp_mcp_message_pack(
+            f"已完成tab页:【{tab_id}】的分析，该tab页存在waf",
+            tab_id=tab_id,
+            recommend_team=recommend_team,
+            raw_head_rate_difference=r_h_rate_diff,
+            raw_headless_rate_difference=h_hless_rate_diff,
+            head_headless_rate_difference=h_hless_rate_diff
+        )

mcp_tools/main.py

@@ -0,0 +1,33 @@
+from fastmcp import FastMCP
+
+from mcp_tools.dp_tools import *
+from tools.browser_manager import browser_manager
+from tools.browser_proxy import client_manager
+
+mcp = FastMCP("Jarvis Brain Mcp Tools")
+
+# 根据环境变量加载模块
+enabled_modules = os.getenv("MCP_MODULES", "TeamNode-Dp").split(",")
+base_cwd = os.getenv("BASE_CWD", os.path.expanduser('~'))
+
+if "TeamNode-Dp" in enabled_modules:
+    # 页面管理
+    register_close_tab(mcp, browser_manager, client_manager)
+    register_switch_tab(mcp, browser_manager, client_manager)
+    register_get_new_tab(mcp, browser_manager, client_manager)
+    # 功能
+    register_visit_url(mcp, browser_manager, client_manager)
+    register_get_html(mcp, browser_manager, client_manager)
+    register_check_selector(mcp, browser_manager, client_manager)
+    register_pop_first_packet(mcp, browser_manager, client_manager)
+
+if "JarvisNode" in enabled_modules:
+    register_assert_waf(mcp, browser_manager, client_manager)
+
+
+def main():
+    mcp.run(transport="stdio")
+
+
+if __name__ == '__main__':
+    main()

tools/browser_manager.py

@@ -0,0 +1,52 @@
+"""浏览器池管理模块 - 单例模式确保状态共享"""
+import random
+from typing import Optional, Tuple
+import os
+from DrissionPage import ChromiumPage, ChromiumOptions
+import platform
+
+
+class BrowserManager:
+    """浏览器池管理器 - 使用单例模式"""
+    _instance = None
+
+    def __new__(cls):
+        if cls._instance is None:
+            cls._instance = super().__new__(cls)
+            cls._instance.browser_pool = {}
+        return cls._instance
+
+    def create_browser(self, user_agent: str = None) -> Tuple[int, ChromiumPage]:
+        """创建新的浏览器实例"""
+        random_port = random.randint(9223, 9934)
+        while random_port in self.browser_pool:
+            random_port = random.randint(9223, 9934)
+
+        co = ChromiumOptions().set_local_port(random_port)
+        if user_agent:
+            co.set_user_agent(user_agent)
+        if platform.system() != 'Windows':
+            co.set_argument('--no-sandbox')
+        custom_data_dir = os.path.join(os.path.expanduser('~'), 'DrissionPage', "userData", f"{random_port}")
+        co.set_user_data_path(custom_data_dir)  # 设置用户数据路径
+        # if not os.path.exists(custom_data_dir):
+        #     os.makedirs(custom_data_dir)
+        self.browser_pool[random_port] = ChromiumPage(co)
+        return random_port, self.browser_pool[random_port]
+
+    def get_browser(self, port: int) -> Optional[ChromiumPage]:
+        """根据端口获取浏览器实例"""
+        return self.browser_pool.get(port)
+
+    def remove_browser(self, port: int) -> Tuple[bool, Optional[ChromiumPage]]:
+        """根据端口移除浏览器实例"""
+        browser = self.browser_pool.pop(port, None)
+        return browser is not None, browser
+
+    def list_browsers(self) -> list[int]:
+        """列出所有活跃的浏览器端口"""
+        return list(self.browser_pool.keys())
+
+
+# 创建全局单例实例
+browser_manager = BrowserManager()

tools/browser_proxy.py

@@ -0,0 +1,171 @@
+import threading
+from collections import deque
+import time
+from DrissionPage import ChromiumPage, ChromiumOptions
+from DrissionPage._pages.chromium_tab import ChromiumTab
+from DrissionPage._units.listener import DataPacket
+from typing import Tuple, Optional
+import json
+
+one_turn_max_token = 20000
+
+
+class DPProxyClient:
+    def __init__(self, driver: ChromiumTab, self_kill=False):
+        self.tab_id = driver.tab_id
+        self.driver = ChromePageProxy(driver, self)
+        self.thread = None
+        self.self_kill = self_kill
+        # self.packet_list = []
+        self.packet_queue = deque()
+
+    def get_driver(self, start_listen, count=None, timeout=10) -> ChromiumTab:
+        """
+        获取代理后的driver、tab
+        :param start_listen: 若你自己写的代码里已经使用自动化框架监听发包的功能了，则该值应该置为False。若没监听，则必须将该值置为True
+        :param count: 需捕获的数据包总数，为None表示无限
+        :param timeout: 两个数据包之间等待的最大时长（秒），为None表示无限，默认为10秒
+        :return:
+        """
+        if start_listen:
+            self.driver.listen.set_targets(res_type="XHR")
+            self.driver.listen.start()
+            self.thread = threading.Thread(target=self.start_listen, args=(count, timeout,))
+            self.thread.start()
+        return self.driver
+
+    def start_listen(self, count=None, timeout=10):
+        for _ in self.driver.listen.steps(count=count, timeout=timeout, gap=1):
+            pass
+
+    # 每次调用函数，都从队列的左端弹出一个数据包
+    def pop_first_packet(self):
+        if self.packet_queue:
+            result = self.packet_queue.popleft()
+            return json.dumps(result, ensure_ascii=False)
+        else:
+            return None
+
+
+class DPProxyClientManager:
+    """浏览器池管理器 - 使用单例模式"""
+    _instance = None
+
+    def __new__(cls):
+        if cls._instance is None:
+            cls._instance = super().__new__(cls)
+            cls._instance.tab_pool = {}
+        return cls._instance
+
+    def create_client(self, tab: ChromiumTab, self_kill=False) -> Tuple[str, DPProxyClient, ChromiumTab]:
+        """创建新的tab页面代理实例"""
+        client = DPProxyClient(tab, self_kill=self_kill)
+        tab = client.get_driver(True)
+        tab_id = tab.tab_id
+        self.tab_pool[tab_id] = {"client": client, "driver": tab}
+        return tab_id, client, tab
+
+    def get_client(self, tab_id: str) -> Optional[DPProxyClient]:
+        """根据端口获取浏览器实例"""
+        return self.tab_pool.get(tab_id).get("client", None)
+
+    def remove_client(self, tab_id: str) -> Tuple[bool, Optional[ChromiumPage]]:
+        """根据端口移除浏览器实例"""
+        client = self.tab_pool.pop(tab_id, None)
+        return client is not None, client
+
+    def list_clients(self) -> list[int]:
+        """列出所有活跃的浏览器端口"""
+        return list(self.tab_pool.keys())
+
+
+class ChromePageProxy:
+    def __init__(self, page, client=None):
+        self.__dict__['page'] = page
+        self.__dict__['client'] = client
+
+    def __getattr__(self, item):
+        attr = getattr(self.page, item)
+        print(item, attr)
+        if item == 'listen':
+            listen_proxy = DrissionPageListenerProxy(attr, self.__dict__['client'])
+            return listen_proxy
+        return attr
+
+
+class DrissionPageListenerProxy:
+    def __init__(self, listener, client=None):
+        self.listener = listener
+        self.client = client
+
+    def __getattr__(self, item):
+        attr = getattr(self.listener, item)
+        # 当监听到wait被调用的时候
+        if item == "wait":
+            def wrapper(*args, **kwargs):
+                result = attr(*args, **kwargs)
+                check_data_packet(result, self.client)
+                return result
+
+            return wrapper
+        # 当监听到steps被调用的时候
+        if item == "steps":
+            def wrapper(*args, **kwargs):
+                if kwargs.get("gap", 1) > 1:
+                    raise Exception("暂不支持多包监控")
+                result = attr(*args, **kwargs)
+                if attr.__name__ == "steps":
+                    for step in result:
+                        check_data_packet(step, self.client)
+                        yield step
+
+            return wrapper
+        return attr
+
+
+def check_data_packet(packet: DataPacket, client: DPProxyClient):
+    """
+    封装监听到的数据包，并将其存放在client的packet_queue中
+    :param packet:
+    :param client:
+    :return:
+    """
+    url = packet.url
+    method = packet.request.method
+    data = None
+    if packet.request.hasPostData:
+        data = packet.request.postData
+    body = packet.response.body
+    body_str = json.dumps(body, ensure_ascii=False)
+    body_str_list = [body_str[i:i + one_turn_max_token] for i in range(0, len(body_str), one_turn_max_token)]
+    body_completed = True
+    for index, body_str in enumerate(body_str_list):
+        if (index + 1) != len(body_str_list):
+            body_completed = False
+        temp_dict = {
+            "url": url,
+            "body_completed": body_completed,
+            "method": method,
+            "request_data": data,
+            "request_headers": dict(packet.request.headers),
+            "response_headers": dict(packet.response.headers),
+            "response_body_segment": body_str,
+        }
+        client.packet_queue.append(temp_dict)
+
+
+client_manager = DPProxyClientManager()
+
+# if __name__ == '__main__':
+#     co = ChromiumOptions().set_user_agent(
+#         "Mozilla/5.0 (Linux; Android 6.0; Nexus 5 Build/MRA58N) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/143.0.0.0 Mobile Safari/537.36")
+#     tab = ChromiumPage(co).latest_tab
+#     client = DPProxyClient(tab, self_kill=False)
+#     # client = CaptchaClient(tab, self_kill=True)
+#     tab = client.get_driver(True)
+#     url = "https://api.toutiaoapi.com/feoffline/hotspot_and_local/html/hot_list/index.html?client_extra_params=%7B%22custom_log_pb%22%3A%22%7B%5C%22style_id%5C%22%3A%5C%2240030%5C%22%2C%5C%22entrance_hotspot%5C%22%3A%5C%22search%5C%22%2C%5C%22location%5C%22%3A%5C%22hot_board%5C%22%2C%5C%22category_name%5C%22%3A%5C%22hotboard_light%5C%22%7D%22%7D&count=50&log_pb=%7B%22style_id%22%3A%2240030%22%2C%22entrance_hotspot%22%3A%22search%22%2C%22location%22%3A%22hot_board%22%2C%22category_name%22%3A%22hotboard_light%22%7D&only_hot_list=1&tab_name=stream&enter_keyword=%23%E7%BE%8E%E5%9B%BD%E9%80%80%E5%87%BA66%E4%B8%AA%E5%9B%BD%E9%99%85%E7%BB%84%E7%BB%87%23"
+#     tab.get(url)
+#     for _ in range(5056):
+#         new_packet = client.pop_first_packet()
+#         print(new_packet, "23")
+#         time.sleep(1)

tools/tools.py

@@ -0,0 +1,108 @@
+import time
+import random
+import os
+import minify_html
+from DrissionPage import ChromiumPage, ChromiumOptions
+from bs4 import BeautifulSoup
+from curl_cffi import requests
+from lxml import html, etree
+
+
+# 使用requests获取html，用于测试是否使用了瑞数和jsl
+def requests_html(url):
+    headers = {
+        "User-Agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/142.0.0.0 Safari/537.36",
+    }
+    response = requests.get(url, headers=headers, verify=False)
+    response.encoding = "utf-8"
+    return response.text, response.status_code
+
+
+# 使用dp无头模式获取html，用于测试是否使用了其他waf，如移动waf
+def dp_headless_html(url):
+    opt = ChromiumOptions().headless(True)
+    opt.set_argument('--no-sandbox')
+    """创建新的浏览器实例"""
+    random_port = random.randint(9934, 10034)
+    custom_data_dir = os.path.join(os.path.expanduser('~'), 'DrissionPage', "userData", f"{random_port}")
+    opt.set_user_data_path(custom_data_dir)  # 设置用户数据路径
+    opt.set_local_port(random_port)
+    page = ChromiumPage(opt)
+    tab = page.latest_tab
+    tab.get(url)
+    # todo: 目前没有更好的方式，为了数据渲染完全，只能硬等【受网速波动影响比较大】
+    time.sleep(10)
+    page_html = tab.html
+    # 无头浏览器在用完之后一定要记得再page级别进行quit
+    page.quit()
+    return page_html
+
+
+# 压缩html
+def compress_html(content, only_text=False):
+    doc = html.fromstring(content)
+    # 删除 style 和 script 标签
+    for element in doc.xpath('//style | //script'):
+        element.getparent().remove(element)
+
+    # 删除 link 标签
+    for link in doc.xpath('//link[@rel="stylesheet"]'):
+        link.getparent().remove(link)
+
+    # 删除 meta 标签（新增功能）
+    for meta in doc.xpath('//meta'):
+        meta.getparent().remove(meta)
+
+    # 删除 style 属性
+    for element in doc.xpath('//*[@style]'):
+        element.attrib.pop('style')
+
+    # 删除所有 on* 事件属性
+    for element in doc.xpath('//*'):
+        for attr in list(element.attrib.keys()):
+            if attr.startswith('on'):
+                element.attrib.pop(attr)
+
+    result = etree.tostring(doc, encoding='unicode')
+    result = minify_html.minify(result)
+    compress_rate = round(len(content) / len(result) * 100)
+    print(f"html压缩比=> {compress_rate}%")
+    if not only_text:
+        return result, compress_rate
+    soup = BeautifulSoup(result, 'html.parser')
+    result = soup.get_text(strip=True)
+    return result, compress_rate
+
+
+# 通过cookie判断是否有waf，需要通过遇到的例子，不断的完善cookie判别函数
+def assert_waf_cookie(cookies: list):
+    for cookie in cookies:
+        cookie_name = cookie['name']
+        cookie_value = cookie['value']
+        if len(cookie_name) == 13 and len(cookie_value) == 88:
+            return True, "瑞数"
+        if "_jsl" in cookie_name:
+            return True, "加速乐"
+    return False, "没有waf"
+
+
+# 对dp_mcp的消息打包
+def dp_mcp_message_pack(message: str, **kwargs):
+    text_obj = {key: value for key, value in kwargs.items()}
+    text_obj.update({"message": message})
+    return {
+        "content": [{
+            "type": "text",
+            # "text": json.dumps(text_obj, ensure_ascii=False)
+            "text": text_obj
+        }]
+    }
+
+# todo: 大致盘一下各种判定的逻辑【以下的所有压缩比之间的差距均取“绝对值”】
+#  1. 如果requests、无头、有头获取到的压缩比之间从差距都在15%以内，则认定该页面是静态页面，此时优先使用requests请求
+#  2. 如果requests的status_code为特定的412，或者521，则判定是瑞数和jsl。[此时还有一个特点：requests的压缩比会与其他两种方式获取到的压缩比差距非常大(一两千的那种)]
+#  3. 如果requests、无头、有头获取到的压缩比之间差距都在40%以上，则判定该页面只可以用有头采集
+#  4. 如果无头和有头获取到的压缩比之间差距小于15%，但是requests和无头的差距大于40%，则认定该页面可以使用无头浏览器采集
+#  5. 如果requests和有头获取到的压缩比之间差距小于15%，但是无头和有头的差距大于40%，则认定该页面优先使用有头浏览器采集
+#  【此时可能是：1.使用了别的检测无头的waf。2.网站使用瑞数，但是这次请求没有拦截requests（不知道是不是瑞数那边故意设置的），
+#   此时如果想进一步判定是否是瑞数，可以使用有头浏览器取一下cookies，如果cookies里面存在瑞数的cookie，那么就可以断定是瑞数】