* Refactor ChatClientBase: drop Pydantic inheritance and add typed generate() overloads
* Align all LLM chat clients with refactored base and unified response models
* Unify LLM utils across providers and delegate streaming/response to provider‑specific handlers
* Refactor LLM pipeline: add HuggingFace tool calls, unify chat client/response types, and switch DurableAgent to loop‑based workflow
* Refactor orchestrators with loops and unify LLM response handling using LLMChatResponse
* test remaining quickstarts after all changes
* run pytest after all changes
* Run linting and formatting checks to ensure code quality
* Update logging, Orchestrator Name and OTel module name
---------
Signed-off-by: Roberto Rodriguez <9653181+Cyb3rWard0g@users.noreply.github.com>
Co-authored-by: Roberto Rodriguez <9653181+Cyb3rWard0g@users.noreply.github.com>
Yaron Schneider