Crate sqlite_knowledge_graph 

SQLite-based Knowledge Graph Library

This library provides a knowledge graph implementation built on SQLite with support for:

• Entities with typed properties
• Relations between entities with weights
• Vector embeddings for semantic search
• Custom SQLite functions for direct SQL operations
• RAG (Retrieval-Augmented Generation) query functions
• Graph algorithms (PageRank, Louvain, Connected Components)

SQLite Extension

This crate can be compiled as a SQLite loadable extension:

cargo build --release
sqlite3 db.db ".load ./target/release/libsqlite_knowledge_graph.dylib"
sqlite3 db.db "SELECT kg_version();"

Re-exports

pub use algorithms::analyze_graph;

pub use algorithms::connected_components;

pub use algorithms::louvain_communities;

pub use algorithms::pagerank;

pub use algorithms::CommunityResult;

pub use algorithms::PageRankConfig;

pub use embed::check_dependencies;

pub use embed::get_entities_needing_embedding;

pub use embed::EmbeddingConfig;

pub use embed::EmbeddingGenerator;

pub use embed::EmbeddingStats;

pub use error::Error;

pub use error::Result;

pub use export::D3ExportGraph;

pub use export::D3ExportMetadata;

pub use export::D3Link;

pub use export::D3Node;

pub use export::DotConfig;

pub use extension::sqlite3_sqlite_knowledge_graph_init;

pub use functions::register_functions;

pub use graph::Direction;

pub use graph::GraphStats;

pub use graph::PathStep;

pub use graph::TraversalNode;

pub use graph::TraversalPath;

pub use graph::TraversalQuery;

pub use graph::Entity;

pub use graph::Neighbor;

pub use graph::Relation;

pub use graph::HigherOrderNeighbor;

pub use graph::HigherOrderPath;

pub use graph::HigherOrderPathStep;

pub use graph::Hyperedge;

pub use migrate::build_relationships;

pub use migrate::migrate_all;

pub use migrate::migrate_papers;

pub use migrate::migrate_skills;

pub use migrate::MigrationStats;

pub use rag::embedder::Embedder;

pub use rag::embedder::FixedEmbedder;

pub use rag::RagConfig;

pub use rag::RagEngine;

pub use rag::RagResult;

pub use schema::create_schema;

pub use schema::schema_exists;

pub use vector::cosine_similarity;

pub use vector::SearchResult;

pub use vector::VectorStore;

pub use vector::TurboQuantConfig;

pub use vector::TurboQuantIndex;

pub use vector::TurboQuantStats;

Modules

algorithms

embed

Vector embedding generation module for semantic search.

error

Error types for sqlite-knowledge-graph.

export

Graph visualization export module.

extension

SQLite extension entry point using sqlite-loadable

functions

SQLite custom functions for knowledge graph operations.

graph

Graph module for entity and relation storage.

migrate

Data migration module for importing external knowledge sources.

rag

Paper-driven two-stage RAG engine.

schema

Database schema creation and management.

vector

Vector module for semantic search with cosine similarity and TurboQuant indexing.

Structs

GraphContext

Graph context for an entity (root + neighbors).

HybridSearchResult

Hybrid search result combining semantic similarity and graph context.

KnowledgeGraph

Knowledge Graph Manager - main entry point for the library.

SearchResultWithEntity

Semantic search result with entity information.