Principle:Infiniflow Ragflow Search Query Execution
| Knowledge Sources | |
|---|---|
| Domains | RAG, Information_Retrieval, Search |
| Last Updated | 2026-02-12 06:00 GMT |
Overview
A search execution pattern that processes queries with hybrid retrieval, streams AI-synthesized answers via SSE, and generates related questions and mind maps.
Description
Search Query Execution handles the runtime processing of search queries. The async_ask function performs hybrid retrieval, builds context from retrieved chunks, and streams an AI-synthesized answer via SSE. Additionally, the search application supports generating related questions (LLM-powered suggestions) and mind maps (structured knowledge visualization).
Usage
Triggered when a user submits a search query in the search application UI or via the API.
Theoretical Basis
Search query execution extends the basic RAG pattern with search-specific features:
- Answer synthesis: LLM generates a comprehensive answer from retrieved chunks
- Related questions: LLM suggests follow-up queries based on the current question
- Mind maps: Structured visualization of topic relationships
Related Pages
Implemented By
Page Connections
Double-click a node to navigate. Hold to expand connections.
Principle
Implementation
Heuristic
Environment