phago_rag/lib.rs
1//! # Phago RAG
2//!
3//! Biological Retrieval-Augmented Generation.
4//!
5//! Instead of vector similarity search, queries traverse the Hebbian knowledge
6//! graph following strongest connections. The graph learns from usage —
7//! frequently traversed paths strengthen, unused ones decay.
8//!
9//! ## How it differs from standard RAG
10//!
11//! | Standard RAG | Phago RAG |
12//! |-------------|-----------|
13//! | Chunk → embed → vector search | Digest → wire → graph traversal |
14//! | Static index | Self-reinforcing graph |
15//! | No learning from queries | Traversed paths strengthen |
16//! | No anomaly detection | Sentinels flag what doesn't fit |
17//! | Flat retrieval | Structured, weighted paths |
18
19pub mod query;
20pub mod scoring;
21pub mod baseline;
22pub mod code_query;
23pub mod hybrid;
24pub mod mcp;
25pub mod prelude;
26
27pub use query::{Query, QueryResult, QueryEngine};
28pub use hybrid::{hybrid_query, HybridConfig, HybridResult};
29pub use mcp::{phago_remember, phago_recall, phago_explore};