feat: add extra_body support for custom API parameters on OpenAIModelConfig

[MuFengMuXue]

feat: add extra_body support to OpenAIModelConfig for custom API params,Allows passing arbitrary parameters (e.g., DeepSeek V4 thinking, reasoning_effort), Uses serde(flatten) to merge extra_body into request root,Enables disabling thinking mode without modifying struct

[Onelevenvy]

Thanks for the PR and for adding extra_body support! 🙏

Since this branch reformatted ~70 files (just cargo fmt differences), which makes reviewing the actual feature change pretty hard, I'll take the actual feature change and apply it on top of main myself, then re-format before merging — that keeps the history clean and reviewable. No need for you to redo anything. Appreciate the contribution!

[MuFengMuXue]

Thank you so much for taking over and applying the changes yourself! 🙏

This is the first pr I made in this repo.I'm really sorry for the review hassle caused by mixing the cargo fmt --all with the feature logic. I totally get that the 70-file diff made it hard to review.

Lesson learned for next time — I'll either format only the files I touched or split the formatting into a separate commit. Really appreciate your patience and guidance on this.🙏🙏🙏

[Onelevenvy]

Thank you so much for taking over and applying the changes yourself! 🙏

This is the first pr I made in this repo.I'm really sorry for the review hassle caused by mixing the cargo fmt --all with the feature logic. I totally get that the 70-file diff made it hard to review.

Lesson learned for next time — I'll either format only the files I touched or split the formatting into a separate commit. Really appreciate your patience and guidance on this.🙏🙏🙏

No worries at all! ❤️ Honestly, part of the confusion was on my end too — I hadn't set up CI for cargo fmt and clippy yet in this repo!

I've just added CI workflows with fmt and clippy checks, as well as updated the repo formatting. Now local dev and PR checks will be much smoother for everyone.

Thanks again for the great PR and for helping make the project better!