[Bug] 通义千问视觉模型无法使用 Missing Content-Length of multimodal url

[Liangpi000]

📦 部署环境

Docker

📌 软件版本

v1.36.31

💻 系统环境

Windows

🌐 浏览器

Edge

🐛 问题描述

使用通义千问视觉模型时，上传文件会提示错误：

{
"error": {
"code": "invalid_parameter_error",
"type": "invalid_request_error",
"param": null,
"message": "<400> InternalError.Algo.InvalidParameter: Missing Content-Length of multimodal url"
},
"endpoint": "https://dashscope.aliyuncs.com/compatible-mode/v1",
"provider": "qwen"
}

📷 复现步骤

No response

🚦 期望结果

修复视觉模型的使用问题，增加Content-Length字段？

📝 补充信息

No response

[lobehubbot]

👀 @Liangpi000

Thank you for raising an issue. We will investigate into the matter and get back to you as soon as possible.
Please make sure you have given us as much context as possible.
非常感谢您提交 issue。我们会尽快调查此事，并尽快回复您。 请确保您已经提供了尽可能多的背景信息。

[lobehubbot]

Bot detected the issue body's language is not English, translate it automatically. 👯👭🏻🧑‍🤝‍🧑👫🧑🏿‍🤝‍🧑🏻👩🏾‍🤝‍👨🏿👬🏿

---

📦 Deployment environment

Docker

📌 Software version

v1.36.31

💻 System environment

Windows

🌐 Browser

Edge

🐛 Problem description

When using Tongyi Qianwen visual model, an error will be prompted when uploading files:

{
"error": {
"code": "invalid_parameter_error",
"type": "invalid_request_error",
"param": null,
"message": "<400> InternalError.Algo.InvalidParameter: Missing Content-Length of multimodal url"
},
"endpoint": "https://dashscope.aliyuncs.com/compatible-mode/v1",
"provider": "qwen"
}

📷 Steps to reproduce

No response

🚦 Expected results

Fix the usage problem of visual model and add Content-Length field?

📝 Supplementary information

No response

[dosubot[bot]]

@Liangpi000 你好！我是Dosu，一个可以帮助解决 bug、回答问题和成为贡献者的机器人。我会在等待人类维护者的过程中帮助你解决问题。

The error you're encountering indicates that the Content-Length header is missing in the request for the multimodal URL when using the 通义千问视觉模型. This header is crucial for specifying the size of the request body, especially when uploading files.

To address this issue, you should ensure that the Content-Length header is correctly set in the HTTP request. This typically involves calculating the size of the file being uploaded and including it in the request headers. If you're using a library or framework to handle HTTP requests, it might automatically set this header for you when you provide the file data correctly.

If you're modifying the code directly, look for the part of the code where the HTTP request is constructed for file uploads and ensure that the Content-Length is included in the headers. If this is part of the lobe-chat framework, you might need to check the relevant part of the codebase where file uploads are handled and adjust it to include this header.

Unfortunately, I couldn't find specific details in the lobe-chat codebase about where the Content-Length header is set for file uploads, so you might need to explore the code related to HTTP request construction for multimodal requests.

To continue talking to Dosu, mention @dosu.

---

^{Help Dosu learn! Give it feedback: Great Response | Irrelevant Answer | Incorrect Sources | Too Verbose | Hallucination | Other}

[202252197]

• 上传的图片需要使用阿里云的OSS的url图片才可以，你可以使用我这个图片url进行测试：https://dashscope.oss-cn-beijing.aliyuncs.com/images/dog_and_girl.jpeg

[lobehubbot]

Bot detected the issue body's language is not English, translate it automatically. 👯👭🏻🧑‍🤝‍🧑👫🧑🏿‍🤝‍🧑🏻👩🏾‍🤝‍👨🏿👬🏿

---

• The uploaded images need to use the url image of Alibaba Cloud's OSS. You can use my image url for testing: https://dashscope.oss-cn-beijing.aliyuncs.com/images/dog_and_girl.jpeg

[Mr-Time-01]

阿里云的OSS的url图片要怎么弄，怎么上传

[liuyingbin1922]

阿里云oss 平台注册使用， 有对应的文档