From this code, you'll be able to know that the query is searched against nodes inside the vector index and retrieves the highest k similarity nodes. The retrieved Nodes are provided to LLM with Prompt and Query to generate the reaction. When a question here is specified, the method begins by randomly deciding upon a single chunk vector, also know