- Removed unnecessary commented-out code and simplified the logic for checking if reasoning is configured and active. This enhances code readability and maintainability.
- Introduced a new README.md file detailing the server handling module, including support for reasoning models, provider differences, effort level mappings, and usage examples.
- Cleaned up the ReasoningConfig class by removing unnecessary comments and clarifying logic related to reasoning injection and provider-specific requirements.
- Updated server classes (OpenAIServer, SGLangServer, TrlVllmServer, VLLMServer) to accept a ReasoningConfig parameter during initialization.
- Enhanced ReasoningConfig to allow flexible max_tokens without strict validation, accommodating varying provider limits.
- Implemented reasoning configuration injection in APIServer methods for chat and completion handling.
- Updated tests to reflect changes in max_tokens validation logic.
This commit integrates reasoning capabilities into the server handling architecture, improving compatibility with diverse reasoning models.
