MonitoredLLM

class council.llm.MonitoredLLM(name: str, llm: LLMBase)

Bases: Monitored[LLMBase]

A convenience class that wraps an LLM into a Monitor

post_chat_request(context: ContextBase, messages: Sequence[LLMMessage], budget: Budget | None = None, **kwargs: Any) → LLMResult

make a call to the wrapped llm, managing the creation of the context. See LLMBase.post_chat_request()

Parameters:

• context (ContextBase) – the context of the caller

• messages (Sequence[LLMMessage]) – see LLMBase.post_chat_request()

• budget (Optional[Budget]) – an optional budget. If none, the budget from the given context is used

• **kwargs – see LLMBase.post_chat_request()

Returns:

see LLMBase.post_chat_request()

Return type:

LLMResult