Lucene Rag Facet Provider
Lucene RAG facet with optional vector search support via an EmbeddingModel. Supports both in-memory and disk-based persistence. Implements WritableContentElementRepository so we can add to the store.
Parameters
Name of this RAG service
Optional embedding model for vector search; if null, only text search is supported
Weighting for vector similarity in hybrid search (0.0 to 1.0)
Configuration for content chunking
Optional path for disk-based index storage; if null, uses in-memory storage
Constructors
Functions
Get statistics about the current state
Manually trigger loading of existing chunks from disk. Useful for ensuring chunks are loaded immediately after startup.
The Retrievables have been saved to the store, but Retrievables are special, and we probably want to embed them
Will call save on the root and all descendants. The database only needs to store each descendant and link by id, rather than otherwise consider the entire structure.