INNER CODE UNIT · Python

retrieve_from_index

kaito-project/kaito · presets/ragengine/main.py:839

async def retrieve_from_index(request: RetrieveRequest):
    """
    Retrieve relevant documents from the specified index.

    This performs pure document retrieve based on semantic similarity,
    without involving LLM generation.
    """
    start_time = time.perf_counter()
    status = STATUS_FAILURE  # Default status

    try:
        result = await rag_ops.retrieve(
            index_name=request.index_name,
            query=request.query,
            max_node_count=request.max_node_count,
            metadata_filter=request.metadata_filter,
        )
        status = STATUS_SUCCESS

View source record →