Expand description
Knowledge base — document ingestion, chunking, and vector or BM25 retrieval.
Modules§
- chunker
- Document chunker — splits large documents into overlapping text chunks for indexing.
- in_
memory - In-memory knowledge base backed by BM25 scoring.
- loader
- Document loader — reads plain text, Markdown, and PDF files into the knowledge base.
- tools
- Knowledge base tool definitions for agent document retrieval.
Structs§
- Chunk
- Atomic search unit: a slice of a document.
- Document
Source - Provenance of a document chunk.
- Knowledge
Query - Query parameters for knowledge search.
- Search
Result - A single search result with relevance info.
Traits§
- Knowledge
Base - Trait for knowledge base implementations.