Query

Level-0 query entry points for GLLM RAG.

query(question, datastore, lm_invoker, top_k=5) async

Run a minimal RAG query asynchronously.

Examples:

result = await query(
    question="What is RAG?",
    datastore=my_datastore.with_vector(em_invoker=my_em_invoker),
    lm_invoker=my_lm_invoker,
    top_k=5,
)

Parameters:

Name	Type	Description	Default
question	str	User question.	required
datastore	BaseDataStore	Datastore instance from gllm-datastore. Must have vector capability registered.	required
lm_invoker	BaseLMInvoker	LM invoker from gllm-inference.	required
top_k	int	Number of chunks to retrieve. Defaults to 5.	5

Returns:

Type	Description
dict[str, Any]	dict[str, Any]: Result payload containing at least response and chunks.

Raises:

Type	Description
TypeError	If datastore is not BaseDataStore or lm_invoker is not BaseLMInvoker.
ValueError	If question/top_k are invalid, or datastore has no vector capability.