向量数据库
定义
向量数据库存储高维向量（嵌入）并支持快速相似性搜索 (例如 k-NN, approximate nearest neighbor). 它们是 the backbone of 检索 in RAG.
They 位于之间 embeddings (which produce the vectors) and the RAG retriever (which needs the top-k chunks). 与…不同 keyword search, they support semantic similarity: “customer support” can match “help desk” if the embedding model maps them close together. See RAG architecture for how the index fits into the full pipeline.
工作原理
Documents are embedded and their vectors 被写入一个 index (例如 HNSW, IVF, or flat for small datasets). At query time, the query vector is compared against the index via k-NN (or approximate k-NN for scale); the index returns top-k ids (and optionally the vectors or stored metadata). You then fetch the corresponding chunks and pass them to the LLM. Options include Pinecone, Weaviate, Chroma, pgvector, and others; choice depends on scale, latency, and whether you need metadata filtering.
应用场景
Vector stores are used whenever you need fast similarity search over many embeddings (RAG, recommendations, dedup).
- Storing and querying document embeddings for RAG
- Real-time similarity search at scale (例如 recommendations, dedup)
- Combining vector search with metadata filters (例如 by date, category)
外部文档
- Chroma – Get started
- Pinecone – Vector database docs
- pgvector — Vector similarity search in PostgreSQL