feat: xhs sign playwright version

NanmiCoder · NanmiCoder · commit f989ce07880f · 2025-11-27T10:53:08.000+08:00
diff --git a/media_platform/xhs/client.py b/media_platform/xhs/client.py
@@ -19,15 +19,12 @@
 
 import asyncio
 import json
-import time
 from typing import TYPE_CHECKING, Any, Callable, Dict, List, Optional, Union
-from urllib.parse import urlencode, urlparse, parse_qs
-
+from urllib.parse import urlencode
 
 import httpx
 from playwright.async_api import BrowserContext, Page
 from tenacity import retry, stop_after_attempt, wait_fixed
-from xhshow import Xhshow
 
 import config
 from base.base_crawler import AbstractApiClient
@@ -39,8 +36,9 @@
 
 from .exception import DataFetchError, IPBlockError
 from .field import SearchNoteType, SearchSortType
-from .help import get_search_id, sign
+from .help import get_search_id
 from .extractor import XiaoHongShuExtractor
+from .playwright_sign import sign_with_playwright
 
 
 class XiaoHongShuClient(AbstractApiClient, ProxyRefreshMixin):
@@ -67,54 +65,36 @@ def __init__(
         self.playwright_page = playwright_page
         self.cookie_dict = cookie_dict
         self._extractor = XiaoHongShuExtractor()
-        # 初始化 xhshow 客户端用于签名生成
-        self._xhshow_client = Xhshow()
         # 初始化代理池（来自 ProxyRefreshMixin）
         self.init_proxy_pool(proxy_ip_pool)
 
     async def _pre_headers(self, url: str, params: Optional[Dict] = None, payload: Optional[Dict] = None) -> Dict:
-        """请求头参数签名
+        """请求头参数签名（使用 playwright 注入方式）
 
         Args:
-            url: 请求的URL(GET请求是包含请求的参数)
+            url: 请求的URL
             params: GET请求的参数
             payload: POST请求的参数
 
         Returns:
             Dict: 请求头参数签名
         """
         a1_value = self.cookie_dict.get("a1", "")
-        parsed = urlparse(url)
-        uri = parsed.path
+
+        # 确定请求数据和 URI
         if params is not None:
-            x_s = self._xhshow_client.sign_xs_get(
-                uri=uri, a1_value=a1_value, params=params
-            )
+            data = params
         elif payload is not None:
-            x_s = self._xhshow_client.sign_xs_post(
-                uri=uri, a1_value=a1_value, payload=payload
-            )
+            data = payload
         else:
             raise ValueError("params or payload is required")
 
-        # 获取 b1 值
-        b1_value = ""
-        try:
-            if self.playwright_page:
-                local_storage = await self.playwright_page.evaluate(
-                    "() => window.localStorage"
-                )
-                b1_value = local_storage.get("b1", "")
-        except Exception as e:
-            utils.logger.warning(
-                f"[XiaoHongShuClient._pre_headers] Failed to get b1 from localStorage: {e}"
-            )
-
-        signs = sign(
+        # 使用 playwright 注入方式生成签名
+        signs = await sign_with_playwright(
+            page=self.playwright_page,
+            uri=url,
+            data=data,
             a1=a1_value,
-            b1=b1_value,
-            x_s=x_s,
-            x_t=str(int(time.time() * 1000)),
         )
 
         headers = {
@@ -177,11 +157,9 @@ async def get(self, uri: str, params: Optional[Dict] = None) -> Dict:
         """
         headers = await self._pre_headers(uri, params)
         if isinstance(params, dict):
-            # 使用 xhsshow build_url 构建完整的 URL
-            full_url = self._xhshow_client.build_url(
-                base_url=f"{self._host}{uri}",
-                params=params
-            )
+            # 构建带参数的完整 URL
+            query_string = urlencode(params)
+            full_url = f"{self._host}{uri}?{query_string}"
         else:
             full_url = f"{self._host}{uri}"
 
@@ -200,7 +178,7 @@ async def post(self, uri: str, data: dict, **kwargs) -> Dict:
 
         """
         headers = await self._pre_headers(uri, payload=data)
-        json_str = self._xhshow_client.build_json_body(payload=data)
+        json_str = json.dumps(data, separators=(",", ":"), ensure_ascii=False)
         return await self.request(
             method="POST",
             url=f"{self._host}{uri}",
diff --git a/media_platform/xhs/playwright_sign.py b/media_platform/xhs/playwright_sign.py
@@ -0,0 +1,203 @@
+# -*- coding: utf-8 -*-
+# Copyright (c) 2025 relakkes@gmail.com
+#
+# This file is part of MediaCrawler project.
+# Repository: https://github.com/NanmiCoder/MediaCrawler/blob/main/media_platform/xhs/playwright_sign.py
+# GitHub: https://github.com/NanmiCoder
+# Licensed under NON-COMMERCIAL LEARNING LICENSE 1.1
+#
+# 声明：本代码仅供学习和研究目的使用。使用者应遵守以下原则：
+# 1. 不得用于任何商业用途。
+# 2. 使用时应遵守目标平台的使用条款和robots.txt规则。
+# 3. 不得进行大规模爬取或对平台造成运营干扰。
+# 4. 应合理控制请求频率，避免给目标平台带来不必要的负担。
+# 5. 不得用于任何非法或不当的用途。
+#
+# 详细许可条款请参阅项目根目录下的LICENSE文件。
+# 使用本代码即表示您同意遵守上述原则和LICENSE中的所有条款。
+
+# 通过 Playwright 注入调用 window.mnsv2 生成小红书签名
+
+import hashlib
+import json
+import time
+from typing import Any, Dict, Optional, Union
+from urllib.parse import urlparse
+
+from playwright.async_api import Page
+
+from .xhs_sign import b64_encode, encode_utf8, get_trace_id, mrc
+
+
+def _build_sign_string(uri: str, data: Optional[Union[Dict, str]] = None) -> str:
+    """构建待签名字符串"""
+    c = uri
+    if data is not None:
+        if isinstance(data, dict):
+            c += json.dumps(data, separators=(",", ":"), ensure_ascii=False)
+        elif isinstance(data, str):
+            c += data
+    return c
+
+
+def _md5_hex(s: str) -> str:
+    """计算 MD5 哈希值"""
+    return hashlib.md5(s.encode("utf-8")).hexdigest()
+
+
+def _build_xs_payload(x3_value: str, data_type: str = "object") -> str:
+    """构建 x-s 签名"""
+    s = {
+        "x0": "4.2.1",
+        "x1": "xhs-pc-web",
+        "x2": "Mac OS",
+        "x3": x3_value,
+        "x4": data_type,
+    }
+    return "XYS_" + b64_encode(encode_utf8(json.dumps(s, separators=(",", ":"))))
+
+
+def _build_xs_common(a1: str, b1: str, x_s: str, x_t: str) -> str:
+    """构建 x-s-common 请求头"""
+    payload = {
+        "s0": 3,
+        "s1": "",
+        "x0": "1",
+        "x1": "4.2.2",
+        "x2": "Mac OS",
+        "x3": "xhs-pc-web",
+        "x4": "4.74.0",
+        "x5": a1,
+        "x6": x_t,
+        "x7": x_s,
+        "x8": b1,
+        "x9": mrc(x_t + x_s + b1),
+        "x10": 154,
+        "x11": "normal",
+    }
+    return b64_encode(encode_utf8(json.dumps(payload, separators=(",", ":"))))
+
+
+async def get_b1_from_localstorage(page: Page) -> str:
+    """从 localStorage 获取 b1 值"""
+    try:
+        local_storage = await page.evaluate("() => window.localStorage")
+        return local_storage.get("b1", "")
+    except Exception:
+        return ""
+
+
+async def call_mnsv2(page: Page, sign_str: str, md5_str: str) -> str:
+    """
+    通过 playwright 调用 window.mnsv2 函数
+
+    Args:
+        page: playwright Page 对象
+        sign_str: 待签名字符串 (uri + JSON.stringify(data))
+        md5_str: sign_str 的 MD5 哈希值
+
+    Returns:
+        mnsv2 返回的签名字符串
+    """
+    sign_str_escaped = sign_str.replace("\\", "\\\\").replace("'", "\\'").replace("\n", "\\n")
+    md5_str_escaped = md5_str.replace("\\", "\\\\").replace("'", "\\'")
+
+    try:
+        result = await page.evaluate(f"window.mnsv2('{sign_str_escaped}', '{md5_str_escaped}')")
+        return result if result else ""
+    except Exception:
+        return ""
+
+
+async def sign_xs_with_playwright(
+    page: Page,
+    uri: str,
+    data: Optional[Union[Dict, str]] = None,
+) -> str:
+    """
+    通过 playwright 注入生成 x-s 签名
+
+    Args:
+        page: playwright Page 对象（必须已打开小红书页面）
+        uri: API 路径，如 "/api/sns/web/v1/search/notes"
+        data: 请求数据（GET 的 params 或 POST 的 payload）
+
+    Returns:
+        x-s 签名字符串
+    """
+    sign_str = _build_sign_string(uri, data)
+    md5_str = _md5_hex(sign_str)
+    x3_value = await call_mnsv2(page, sign_str, md5_str)
+    data_type = "object" if isinstance(data, (dict, list)) else "string"
+    return _build_xs_payload(x3_value, data_type)
+
+
+async def sign_with_playwright(
+    page: Page,
+    uri: str,
+    data: Optional[Union[Dict, str]] = None,
+    a1: str = "",
+) -> Dict[str, Any]:
+    """
+    通过 playwright 生成完整的签名请求头
+
+    Args:
+        page: playwright Page 对象（必须已打开小红书页面）
+        uri: API 路径
+        data: 请求数据
+        a1: cookie 中的 a1 值
+
+    Returns:
+        包含 x-s, x-t, x-s-common, x-b3-traceid 的字典
+    """
+    b1 = await get_b1_from_localstorage(page)
+    x_s = await sign_xs_with_playwright(page, uri, data)
+    x_t = str(int(time.time() * 1000))
+
+    return {
+        "x-s": x_s,
+        "x-t": x_t,
+        "x-s-common": _build_xs_common(a1, b1, x_s, x_t),
+        "x-b3-traceid": get_trace_id(),
+    }
+
+
+async def pre_headers_with_playwright(
+    page: Page,
+    url: str,
+    cookie_dict: Dict[str, str],
+    params: Optional[Dict] = None,
+    payload: Optional[Dict] = None,
+) -> Dict[str, str]:
+    """
+    使用 playwright 注入方式生成请求头签名
+    可直接替换 client.py 中的 _pre_headers 方法
+
+    Args:
+        page: playwright Page 对象
+        url: 请求 URL
+        cookie_dict: cookie 字典
+        params: GET 请求参数
+        payload: POST 请求参数
+
+    Returns:
+        签名后的请求头字典
+    """
+    a1_value = cookie_dict.get("a1", "")
+    uri = urlparse(url).path
+
+    if params is not None:
+        data = params
+    elif payload is not None:
+        data = payload
+    else:
+        raise ValueError("params or payload is required")
+
+    signs = await sign_with_playwright(page, uri, data, a1_value)
+
+    return {
+        "X-S": signs["x-s"],
+        "X-T": signs["x-t"],
+        "x-S-Common": signs["x-s-common"],
+        "X-B3-Traceid": signs["x-b3-traceid"],
+    }
diff --git a/media_platform/xhs/xhs_sign.py b/media_platform/xhs/xhs_sign.py
