Refactor LLM processor and improve async handling

Move contextual information handling from noise filtering to extraction
and centralize LLM call logic. Wrap blocking transcription and state
update calls in asyncio.to_thread to prevent event loop blocking.
Update transcriber model size to base.

src/llm/models.py

@@ -55,10 +55,6 @@ class ContextUpdate(BaseModel):
 
 
 class FilterResult(BaseModel):
-    contextual_info: str = Field(
-        ...,
-        description="Information interesting to the user but not useful for structured extraction",
-    )
     filtered_text: str = Field(
         ..., description="Cleaned transcript used for structured data extraction"
     )