Module retriever 

Document retrieval strategies.

This module provides retrievers for finding relevant content in document trees:

• LLM Navigation — Tree traversal guided by LLM decisions
• Context Building — Assembling results for LLM consumption

Example

use vectorless::retriever::{LlmNavigator, RetrieveOptions, ContextBuilder};
use vectorless::core::DocumentTree;

let tree = DocumentTree::new("Root", "Content");
let retriever = LlmNavigator::with_defaults();
let options = RetrieveOptions::new().with_top_k(5);

let results = retriever.retrieve(&tree, "What is this about?", &options).await?;

let context = ContextBuilder::new()
    .with_max_tokens(4000)
    .build(&results);

println!("Context: {}", context);

Structs

ContextBuilder

Context builder for assembling retrieval results.

LlmNavigator

LLM-based tree navigation retriever.

NavigationContext

Context for LLM-based navigation.

RetrievalResult

A single retrieval result.

RetrieveOptions

Options for retrieval operations.

Enums

NavigationDecision

Navigation decision for tree traversal.

Functions

format_for_llm

Format retrieval results for LLM consumption.

format_tree_for_llm

Format a document tree for LLM consumption.