- Implement sanitization for `tool_choice` and removal of `disable_parallel_tool_use` in request payloads.
- Introduce logging for tool choice changes in `DoRequestHelper`.
- Update `ConvertRequest` to handle tool-call compatibility and maintain structured tool history.
- Add `ThoughtSignature` to `Part` struct for better tracking of reasoning content.
- Refactor request handling in `getRequestBody` to ensure strict compliance with OpenAI API requirements.
* feat: support gemini-2.0-flash
- Enhance model support by adding new entries and refining checks for system instruction compatibility.
- Update logging display behavior and adjust default quotas for better user experience.
- Revamp pricing structures in the billing system to reflect current model values and deprecate outdated entries.
- Streamline code by replacing hardcoded values with configurations for maintainability.
* feat: add new Gemini 2.0 flash models to adapter and billing ratio
* fix: update GetRequestURL to support gemini-1.5 model in versioning
* feat: add MILLI_USD constant and update pricing for deepseek services
* feat: add support for new Gemini model version 'gemini-2.0-flash-thinking-exp-01-21'
* Update model.go
* Support Gemini tool_calls.
* Fix gemini tool calls (also keep support functions).
* Fixed the problem of arguments not being stringified.
Fix panic: candidate.Content.Parts out of range
