✨ 使用 playwright 绕开 cf 限制 (#52)

Co-authored-by: Akirami <[email protected]>

README.md

@@ -71,6 +71,8 @@ _✨ ChatGPT AI 对话 ✨_
 
 在 nonebot2 项目的`.env`文件中添加下表中的必填配置（在arm平台，可能必须使用CHATGPT_SESSION_TOKEN登录）
 
+windows系统需要在env文件中设置fastapi_reload=false
+
 | 配置项 | 必填 | 默认值 | 说明 |
 |:-----:|:----:|:----:|:----:|
 | CHATGPT_SESSION_TOKEN | 否 | 空字符串 | ChatGPT 的 session_token，如配置则优先使用 |

nonebot_plugin_chatgpt/__init__.py

@@ -19,10 +19,9 @@
 
 from nonebot_plugin_apscheduler import scheduler
 
-if config.chatgpt_image:
-    require("nonebot_plugin_htmlrender")
+require("nonebot_plugin_htmlrender")
 
-    from nonebot_plugin_htmlrender import md_to_pic
+from nonebot_plugin_htmlrender import md_to_pic
 
 chat_bot = Chatbot(
     token=setting.token or config.chatgpt_session_token,

nonebot_plugin_chatgpt/chatgpt.py

@@ -1,7 +1,7 @@
 import uuid
 from typing import Any, Dict, Optional
 from urllib.parse import urljoin
-
+import asyncio
 import httpx
 from nonebot.log import logger
 from nonebot.utils import escape_tag, run_sync
@@ -11,8 +11,12 @@
     import ujson as json
 except ModuleNotFoundError:
     import json
+from playwright.async_api import async_playwright
+
+js = "Object.defineProperties(navigator, {webdriver:{get:()=>undefined}});"
 
 SESSION_TOKEN_KEY = "__Secure-next-auth.session-token"
+CF_CLEARANCE_KEY = "cf_clearance"
 
 
 class Chatbot:
@@ -34,6 +38,9 @@ def __init__(
         self.timeout = timeout
         self.authorization = ""
 
+        self.cf_clearance = ""
+        self.user_agent = ""
+
         if self.session_token:
             self.auto_auth = False
         elif self.account and self.password:
@@ -59,7 +66,7 @@ def headers(self) -> Dict[str, str]:
             "Accept": "text/event-stream",
             "Authorization": f"Bearer {self.authorization}",
             "Content-Type": "application/json",
-            "User-Agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/16.1 Safari/605.1.15",
+            "User-Agent": self.user_agent,
             "X-Openai-Assistant-App-Id": "",
             "Connection": "close",
             "Accept-Language": "en-US,en;q=0.9",
@@ -82,19 +89,31 @@ def get_payload(self, prompt: str) -> Dict[str, Any]:
         }
 
     async def get_chat_response(self, prompt: str) -> str:
-        if not self.authorization:
-            await self.refresh_session()
+        for i in range(2):
+            if not self.authorization:
+                await self.refresh_session()
+            else:
+                break
+        else:
+            return "获取session失败，请检查后台报错"
+        cookies = {SESSION_TOKEN_KEY: self.session_token}
+        if self.cf_clearance:
+            cookies[CF_CLEARANCE_KEY] = self.cf_clearance
         async with httpx.AsyncClient(proxies=self.proxies) as client:
             response = await client.post(
                 urljoin(self.api_url, "backend-api/conversation"),
                 headers=self.headers,
+                cookies=cookies,
                 json=self.get_payload(prompt),
                 timeout=self.timeout,
             )
         if response.status_code == 429:
             return "请求过多，请放慢速度"
         if response.status_code == 401:
             return "token失效，请重新设置token"
+        if response.status_code == 403:
+            await self.get_cf_cookies()
+            return await self.get_chat_response(prompt)
         if response.is_error:
             logger.opt(colors=True).error(
                 f"非预期的响应内容: <r>HTTP{response.status_code}</r> {escape_tag(response.text)}"
@@ -108,10 +127,13 @@ async def get_chat_response(self, prompt: str) -> str:
         return response["message"]["content"]["parts"][0]
 
     async def refresh_session(self) -> None:
+        logger.debug("正在刷新session")
         if self.auto_auth:
             await self.login()
         else:
             cookies = {SESSION_TOKEN_KEY: self.session_token}
+            if self.cf_clearance:
+                cookies[CF_CLEARANCE_KEY] = self.cf_clearance
             async with httpx.AsyncClient(
                 cookies=cookies,
                 proxies=self.proxies,
@@ -120,14 +142,19 @@ async def refresh_session(self) -> None:
                 response = await client.get(
                     urljoin(self.api_url, "api/auth/session"),
                     headers={
-                        "User-Agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/16.1 Safari/605.1.15",
+                        "User-Agent": self.user_agent,
                     },
                 )
             try:
+                if response.status_code == 403:
+                    await self.get_cf_cookies()
+                    await self.refresh_session()
+                    return
                 self.session_token = (
                     response.cookies.get(SESSION_TOKEN_KEY) or self.session_token
                 )
                 self.authorization = response.json()["accessToken"]
+                logger.debug("刷新会话成功: " + self.session_token + self.cf_clearance)
             except Exception as e:
                 logger.opt(colors=True, exception=e).error(
                     f"刷新会话失败: <r>HTTP{response.status_code}</r> {escape_tag(response.text)}"
@@ -159,3 +186,42 @@ def login(self) -> None:
                     logger.opt(exception=e).error("ChatGPT 登陆错误!")
         else:
             logger.error("ChatGPT 登陆错误!")
+
+    async def get_cf_cookies(self) -> None:
+        logger.debug("正在获取cf cookies")
+        async with async_playwright() as p:
+            try:
+                browser = await p.firefox.launch(
+                    headless=True,
+                    proxy={"server": self.proxies}
+                    if self.proxies
+                    else None,  # your proxy
+                )
+            except Exception as e:
+                logger.opt(exception=e).error(
+                    "playwright未安装，请先在shell中运行playwright install"
+                )
+                return
+            ua = f"Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:72.0) Gecko/20100101 Firefox/{browser.version}"
+            content = await browser.new_context(user_agent=ua)
+            page = await content.new_page()
+            await page.add_init_script(js)
+            cf_clearance = None
+            await page.goto("https://chat.openai.com/chat")
+            for j in range(6):
+                if cf_clearance:
+                    break
+                await asyncio.sleep(5)
+                cookies = await content.cookies()
+                for i in cookies:
+                    if i["name"] == "cf_clearance":
+                        cf_clearance = i
+                        break
+            else:
+                logger.error("cf cookies获取失败，可能遇到了人工校验")
+            self.cf_clearance = cf_clearance["value"]
+            self.user_agent = ua
+            await page.close()
+            await content.close()
+            await browser.close()
+        logger.debug("cf cookies获取成功")