locallm

fix in api

fix in types

api: Add configurable thinking tags and progress info support

types: Add new inference parameters and interfaces

Fixes in the openai api

Update @locallm/types package to version 0.7.1 and add onThink callback to InferenceOptions