feat : support deepseek llm

[pengjunfeng11]

here is specific description #218

[devwdave]

Hi @pengjunfeng11

Thanks a lot for opening this PR and adding DeepSeek support!

I’ve looked through the changes and the overall approach, using the OpenAI SDK with a custom base URL is totally reasonable. That’s actually the same pattern we use for Nebius support.

One thing I’d like to ask is if you could trim back some of the changes and align this a bit more closely with how we’ve implemented Nebius in the SDK. In particular, it would be great if DeepSeek could:

Reuse the existing OpenAI client setup as much as possible

Follow a similar structure / minimal set of files as the Nebius integration

Could you take a look at how we currently support Nebius in the SDK and refactor this PR to mirror that approach where it makes sense?

Thanks again for the contribution, we're excited to get DeepSeek support in!

[pengjunfeng11]

No problem, I will do another PR to support this issue in the near future

[pengjunfeng11]

I've updated part of the code to see if there's a problem?

[devwdave]

Hi @pengjunfeng11 thanks again for the update.

One small request before we move forward: can we keep the example output / print statements in English so the repo is consistent and accessible for the broader OSS audience? In examples/deepseek/main.py there are a few print(...) lines (and some example prompts) not in English. Could you translate those to English (and ideally make the sample content more generic, not personal), then push an update?

Thank you!

[pengjunfeng11]

Hi @pengjunfeng11 thanks again for the update.

One small request before we move forward: can we keep the example output / print statements in English so the repo is consistent and accessible for the broader OSS audience? In examples/deepseek/main.py there are a few print(...) lines (and some example prompts) not in English. Could you translate those to English (and ideally make the sample content more generic, not personal), then push an update?

Thank you!

你好，首先，感谢你们的贡献，让我们能了解一个优秀的项目！其次，我不认为在print方法中修改语言有任何区别。exmaple的目的是为了让更多人了解这些方法的使用方式，从这个角度出发，我的代码完全实现了它的目的。如果有必要的话，我甚至可以支持更多语言而不是一种。但从开发者的角度而言，中文是我的母语，你所看到的这段话是由LLM翻译的（感谢AI！），所以使用英文对我反而是不太方便的方式
但你说的对，我确实不应该在代码中加入太多的个人信息，我将会把它修改的更通用

Hello, first of all, thank you for your contributions that allow us to understand an excellent project! Secondly, I don't think there is any difference in modifying the language in the print method. The purpose of the example is to let more people understand how to use these methods, and from this perspective, my code fully achieves its goal. If necessary, I can even support more languages than just one. But from a developer's perspective, Chinese is my mother tongue, and the text you see here was translated by an LLM (thank you for the AI!), so using English is actually not very convenient for me.
But you are right, I definitely shouldn't include too much personal information in the code. I will modify it to be more general.

[pengjunfeng11]

Hi @pengjunfeng11 thanks again for the update.

One small request before we move forward: can we keep the example output / print statements in English so the repo is consistent and accessible for the broader OSS audience? In examples/deepseek/main.py there are a few print(...) lines (and some example prompts) not in English. Could you translate those to English (and ideally make the sample content more generic, not personal), then push an update?

Thank you!

Hello, first of all, thank you for your contributions that allow us to understand an excellent project! Secondly, I don't think there is any difference in modifying the language in the print method. The purpose of the example is to help more people understand how to use these methods. From this perspective, my code fully achieves its goal (all my comments and README are in English). If necessary, I can even support multiple languages instead of just one. But from a developer's perspective, Chinese is my native language, and the text you see here was translated by an LLM (thank you for the AI!), so using English is actually less convenient for me.
But you are right, I definitely shouldn't include too much personal information in the code. I will modify it to be more general.

[devwdave]

Hi @pengjunfeng11. Thank you for following up and for the contribution here.

To clarify, this isn’t about the adequacy or correctness of the implementation, and I agree the language in print(...) doesn’t change functionality. This is purely for repository consistency and long term maintainability. We keep examples, prompts, and console output in English across the codebase so the project remains consistent and maintainable for reviewers and future contributors.

Could you please update this PR so the example strings/output are in English? Once that’s done, we’re happy to continue with the review process.