Core Concepts

This section covers the foundational building blocks of Laurus. Understanding these concepts will help you design effective schemas and configure your search engine.

Topics

Schema & Fields

How to define the structure of your documents. Covers:

• Schema and SchemaBuilder
• Lexical field types (Text, Integer, Float, Boolean, DateTime, Geo, Bytes)
• Vector field types (Flat, HNSW, IVF)
• Document and DocumentBuilder
• DataValue — the unified value type

Text Analysis

How text is processed before indexing. Covers:

• The Analyzer trait and the analysis pipeline
• Built-in analyzers (Standard, Japanese, Keyword, Pipeline)
• PerFieldAnalyzer — different analyzers for different fields
• Tokenizers and token filters

Embeddings

How text and images are converted to vectors. Covers:

• The Embedder trait
• Built-in embedders (Candle BERT, OpenAI, CLIP, Precomputed)
• PerFieldEmbedder — different embedders for different fields

Storage

Where index data is stored. Covers:

• The Storage trait
• Storage backends (Memory, File, Mmap)
• PrefixedStorage for component isolation
• Choosing the right backend for your use case