Vector Store: FAISS Index

Overview

The vector store handles the “Search and Rescue” operation. It stores the mathematical embeddings of the Child Chunks to optimize search speed using the FAISS library.

Structure / Schema

• File: vector_index.faiss
• Content: The embeddings (vectors) of the Child Chunks.

Flow / Behavior

graph TD
    User["👤 User Query"] --> Embed["🔢 Convert to Vector"]
    Embed --> Search["🔎 FAISS Search in vector_index.faiss"]
    Search --> Match["📍 Found Child Chunk"]

Usage Examples

# FAISS loads the vector map into RAM
import faiss
index = faiss.read_index('vector_index.faiss')

Related Components

• document-store
• chat-engine

---

Last Updated: 2026-05-01