Weaviate

Vector database used for semantic search and retrieval-augmented generation.

Repository

CROP-DB-Weaviate

Overview

Stores and serves vector embeddings for:

• Parts image embeddings (via CLIP)
• Parts description embeddings
• Semantic search queries
• RAG pipeline retrieval

Integration Points

• PDF Parser: Stores extracted parts vectors
• Embedding Service: Generates vectors for storage
• RAG Pipelines: Retrieves relevant context for LLM
• Search Service: Hybrid search (keyword + semantic)