feat: add support for Bedrock/Anthropic ToolChoice to structured_output

[liushang1997]

Description

Add a new parameter tool_choice to the Model.stream method to allow providers to configure the tool choice. We do not expose this anywhere in the Agent/event_loop, so for now this is just an internal implementation detail, but it does allow/enable providers that use tools for structured_output - specifically Bedrock & Anthropic - to force tool calling to provide structured_output; this should provide better consistency for customers as the LLM is now guaranteed to call that api.

Changes made

1. Add a new parameter to Model.stream() that takes in the toolChoice, defaulting to auto;
2. Plumb through various providers:
   • anthropic (needed)
   • bedrock (needed)
   • openai (implemented)
   • litellm (implemented via openai)
3. Emit warnings for providers that it's not plumbed through:
   • llamaapi
   • mistral
   • ollama
   • sagemaker
   • writer
4. For providers that use tool calling for structured_output (Bedrock and Anthropic), pass toolChoice=any when calling stream()
5. Add unit tests
6. Add conformance test to verify structured_output through relevant providers is forced

Related Issues

#256

Documentation PR

N/A

Type of Change

New feature

Testing

How have you tested the change? Verify that the changes do not break functionality or introduce warnings in consuming repositories: agents-docs, agents-tools, agents-cli

• I ran hatch run prepare

Checklist

• I have read the CONTRIBUTING document
• I have added any necessary tests that prove my fix is effective or my feature works
• I have updated the documentation accordingly
• I have added an appropriate example to the documentation to outline the feature, or no new docs are needed
• My changes generate no new warnings
• Any dependent changes have been merged and published

---

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

I made changes