feat: Support stream + tool call + structured output

[Zephyroam]

Description

Fixes #3782

Checklist

Go over all the following points, and put an x in all the boxes that apply.

• I have read the CONTRIBUTION guide (required)
• I have linked this PR to an issue using the Development section on the right sidebar or by adding Fixes #issue-number in the PR description (required)
• I have checked if any dependencies need to be added or updated in pyproject.toml and uv lock
• I have updated the tests accordingly (required for a bug fix or a new feature)
• I have updated the documentation if needed:
• I have added examples if this is a new feature

If you are unsure about any of these, don't hesitate to ask. We are here to help!

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

Run ID: aad5f641-cbba-4ecf-b298-82d9f0d2d7dc

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[fengju0213]

Nice implementation. I noticed that we also added a fallback mechanism. I think retaining the native capabilities of the OpenAI SDK is already sufficient, and introducing a fallback would bring significant maintenance overhead. I suggest removing the related code.

Since this PR was submitted via a forked repository, I’m not able to submit a PR myself, so I made the changes directly in your PR.

[Zephyroam]

Now Gemini and Claude are supported.

[waleedalzarooni]

Great job @Zephyroam!,

Couple of quick comments

[fengju0213]

I've been wondering about something. Since the OpenAI SDK doesn't natively support simultaneous structured outputs and tool calling in streaming mode, our current approach departs from the native pattern. Should we avoid building these custom workarounds for capabilities the model doesn't currently support? After all, as the model's capabilities evolve, this adaptation code will likely become obsolete.
cc @Wendong-Fan

[Wendong-Fan]

I've been wondering about something. Since the OpenAI SDK doesn't natively support simultaneous structured outputs and tool calling in streaming mode, our current approach departs from the native pattern. Should we avoid building these custom workarounds for capabilities the model doesn't currently support? After all, as the model's capabilities evolve, this adaptation code will likely become obsolete. cc @Wendong-Fan

It brings value to the agent framework by adding useful features beyond the current base model. While model platforms may support this natively in the future, the implementation cost is low. Given how important and useful it is, I think doing it ourselves makes sense

[bytecii]

In general LGTM, left some comments

[Zephyroam]

Done, thanks for your comments @bytecii

[Zephyroam]

https://github.com/camel-ai/camel/pull/3821/changes#diff-f90260a7e58e0c09b9d2e737a580635d71a7310f1a4adebdb3b41323a960698cR744
@Wendong-Fan max_tokens is a mandatory parameter for the new SDK version. I have to set a default value. I refer to pydantic/pydantic-ai#2553.