[Bug]4.23.6版本对话模型获取失败

[SARFF6]

What happened / 发生了什么

在win11系统下载的astrbot桌面版更新到4.23.6版本后，本地部署的对话模型和DeepSeek模型都无法获取和使用

Reproduce / 如何复现？

点开模型提供商，点击获取模型列表

AstrBot version, deployment method (e.g., Windows Docker Desktop deployment), provider used, and messaging platform used. / AstrBot 版本、部署方式（如 Windows Docker Desktop 部署）、使用的提供商、使用的消息平台适配器

桌面客户端部署4.23.6版本，deep seek模型，QQ开放平台

OS

Windows

Logs / 报错日志

[2026-04-29 12:45:16.343] [Core] [INFO] [provider.manager:572]: 载入 openai_chat_completion(deepseek-v4-flash) 服务提供商 ...
[2026-04-29 12:45:16.344] [Core] [ERRO] [v4.23.6] [provider.manager:709]: 实例化 openai_chat_completion(deepseek-v4-flash) 提供商适配器失败：Invalid http_client argument; Expected an instance of httpx.AsyncClient but got <class 'httpx.AsyncClient'>
[2026-04-29 12:47:48.720] [Core] [INFO] [provider.manager:726]: providers in user's config: ['deepseek-v4-flash', 'openai/deepseek-v4-pro', 'ollama/moondream:latest', 'mxbai-embed-large:latest', 'openai/deepseek-chat', 'openai/deepseek-reasoner']
[2026-04-29 12:47:48.933] [Core] [INFO] [provider.manager:572]: 载入 openai_chat_completion(deepseek-v4-flash) 服务提供商 ...
[2026-04-29 12:47:48.934] [Core] [ERRO] [v4.23.6] [provider.manager:709]: 实例化 openai_chat_completion(deepseek-v4-flash) 提供商适配器失败：Invalid http_client argument; Expected an instance of httpx.AsyncClient but got <class 'httpx.AsyncClient'>
[2026-04-29 12:47:49.530] [Core] [INFO] [provider.manager:726]: providers in user's config: ['deepseek-v4-flash', 'openai/deepseek-v4-pro', 'ollama/moondream:latest', 'mxbai-embed-large:latest', 'openai/deepseek-chat', 'openai/deepseek-reasoner']
[2026-04-29 12:47:50.159] [Core] [INFO] [provider.manager:572]: 载入 openai_chat_completion(openai/deepseek-v4-pro) 服务提供商 ...
[2026-04-29 12:47:50.160] [Core] [ERRO] [v4.23.6] [provider.manager:709]: 实例化 openai_chat_completion(openai/deepseek-v4-pro) 提供商适配器失败：Invalid http_client argument; Expected an instance of httpx.AsyncClient but got <class 'httpx.AsyncClient'>
[2026-04-29 12:48:00.080] [Core] [INFO] [provider.manager:726]: providers in user's config: ['deepseek-v4-flash', 'openai/deepseek-v4-pro', 'ollama/moondream:latest', 'mxbai-embed-large:latest', 'openai/deepseek-chat', 'openai/deepseek-reasoner']
[2026-04-29 12:48:00.499] [Core] [INFO] [provider.manager:572]: 载入 openai_chat_completion(ollama/moondream:latest) 服务提供商 ...
[2026-04-29 12:48:00.500] [Core] [ERRO] [v4.23.6] [provider.manager:709]: 实例化 openai_chat_completion(ollama/moondream:latest) 提供商适配器失败：Invalid http_client argument; Expected an instance of httpx.AsyncClient but got <class 'httpx.AsyncClient'>
[2026-04-29 12:48:00.978] [Core] [INFO] [routes.config:743]: API call: /config/provider/check_one id=ollama/moondream:latest
[2026-04-29 12:48:00.979] [Core] [WARN] [v4.23.6] [routes.config:749]: Provider with id 'ollama/moondream:latest' not found in provider_manager.
[2026-04-29 12:48:05.108] [Core] [INFO] [routes.config:743]: API call: /config/provider/check_one id=mxbai-embed-large:latest
[2026-04-29 12:48:06.684] [Core] [INFO] [routes.config:706]: Provider mxbai-embed-large:latest (ID: mxbai-embed-large:latest) is available.

Are you willing to submit a PR? / 你愿意提交 PR 吗？

• Yes!

Code of Conduct

• I have read and agree to abide by the project's Code of Conduct。

[dosubot]

Hi @SARFF6! I'm Dosu and I’m helping the AstrBot team.

这个问题与 #7879 报告的完全一致——是 httpx 库版本/路径冲突 导致 OpenAI SDK 内部的 isinstance(http_client, httpx.AsyncClient) 检查失败 [1]。

根因： 环境中存在多个 httpx 安装路径（通常由插件安装引起），导致传入 AsyncOpenAI 的 httpx.AsyncClient 实例与 SDK 期望的类来自不同模块路径，isinstance 检查就会失败。

当前修复状态：

• PR #7685 已合并，集中化了 HTTP 客户端创建 [2]
• PR #7773 针对 OpenAI 模块路径冲突的专项修复，尚未合并 [3]

临时解决方法： 由于你使用的是桌面客户端，可以尝试重新安装/更新 AstrBot 桌面版，或手动在 AstrBot 的 Python 环境中强制重装 httpx：

pip install --force-reinstall httpx

然后重启 AstrBot。

To reply, just mention @dosu.

---

Docs are dead. Just use Dosu.

[NayukiChiba]

pr #7773 和pr #7587 都是修复的多个httpx路径问题，等待下个版本修复即可

[Kisaragi911]

这个bug不光个别插件的问题，很多插件都会出现这个问题，基本已经不可用了

[ltk415]

我也出现这个问题了，希望尽快修复，这是严重bug。请问如何在手动在 AstrBot 的 Python 环境中强制重装 httpx：？我是mac系统，用终端无法运行pip命令，感谢。

[NayukiChiba]

已经在修复了😭😭😭