weaviate-rag

Semantic search with Weaviate involves configuring a vector store to embed and retrieve documents by meaning rather than keyword matching. This Skill configures a Weaviate client with embedding models, defines a schema for document storage, and deploys the system via Docker to enable similarity-based retrieval across your document corpus.

Weaviate supports multiple embedding models including nomic-embed-text, bge-m3, and mxbai-embed-large. This Skill configures the text2vec-transformers vectorizer to automatically embed documents during ingestion, allowing you to choose the model that best fits your semantic search performance and latency requirements.

Yes. This Skill provides Docker Compose configuration to deploy Weaviate locally with HTTP endpoint at localhost:8095 and gRPC at localhost:50055, enabling you to develop and test RAG workflows on your machine before moving to production environments.

Configure the Weaviate client by providing an OPENAI API key header for authentication. This Skill handles client setup with the required credentials, allowing your RAG system to authenticate embedding and retrieval requests within MoodleNRW workflows and Chainlit UI interfaces.

Keyword search matches exact terms; semantic search understands meaning and context to find relevant documents even when phrasing differs. Weaviate-powered RAG uses vector embeddings to perform semantic search, returning contextually similar results that keyword systems might miss.

Yes. This Skill defines a MoodleDocument schema with the text2vec-transformers vectorizer, specifying how documents are structured and embedded. The schema ensures consistent document processing and enables efficient semantic search across stored entries.