Congo River Compositional Intelligence

-- Congo River Compositional Intelligence Database Schema -- PostgreSQL with pgvector extension for vector embeddings -- Enable required extensions CREATE EXTENSION IF NOT EXISTS "uuid-ossp"; CREATE EXTENSION IF NOT EXISTS "pgvector"; -- ============================================================================ -- Knowledge Graph Storage -- ============================================================================ -- RDF-style triples for semantic knowledge representation CREATE TABLE IF NOT EXISTS triples ( id UUID PRIMARY KEY DEFAULT uuid_generate_v4(), subject TEXT NOT NULL, predicate TEXT NOT NULL, object TEXT NOT NULL, context TEXT, -- Named graph for organizing knowledge source TEXT, -- Which service/tool created this triple confidence FLOAT DEFAULT 1.0, -- Confidence score (0.0 - 1.0) created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP, updated_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP, metadata JSONB -- Additional properties ); -- Indexes for efficient triple queries CREATE INDEX IF NOT EXISTS idx_triples_subject ON triples(subject); CREATE INDEX IF NOT EXISTS idx_triples_predicate ON triples(predicate); CREATE INDEX IF NOT EXISTS idx_triples_object ON triples(object); CREATE INDEX IF NOT EXISTS idx_triples_context ON triples(context); CREATE INDEX IF NOT EXISTS idx_triples_source ON triples(source); CREATE INDEX IF NOT EXISTS idx_triples_spo ON triples(subject, predicate, object); -- ============================================================================ -- Proof Search and Reasoning -- ============================================================================ -- Store proof trees and inference traces CREATE TABLE IF NOT EXISTS proofs ( id UUID PRIMARY KEY DEFAULT uuid_generate_v4(), goal TEXT NOT NULL, premises JSONB, -- Array of premise statements proof_tree JSONB NOT NULL, -- Complete proof structure proof_method TEXT, -- forward_chaining, backward_chaining, resolution, etc. success BOOLEAN NOT NULL, error_message TEXT, duration_ms INTEGER, created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP, metadata JSONB ); -- Indexes for proof queries CREATE INDEX IF NOT EXISTS idx_proofs_goal ON proofs(goal); CREATE INDEX IF NOT EXISTS idx_proofs_success ON proofs(success); CREATE INDEX IF NOT EXISTS idx_proofs_created_at ON proofs(created_at DESC); -- ============================================================================ -- Reasoning History and Analytics -- ============================================================================ -- Track all tool invocations and their results CREATE TABLE IF NOT EXISTS reasoning_sessions ( id UUID PRIMARY KEY DEFAULT uuid_generate_v4(), tool_name TEXT NOT NULL, input JSONB NOT NULL, output JSONB NOT NULL, success BOOLEAN NOT NULL, error_message TEXT, duration_ms INTEGER, language_used TEXT, -- Which programming language service was used created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP, user_context TEXT, -- Optional user/session identifier metadata JSONB ); -- Indexes for reasoning history CREATE INDEX IF NOT EXISTS idx_reasoning_tool ON reasoning_sessions(tool_name); CREATE INDEX IF NOT EXISTS idx_reasoning_success ON reasoning_sessions(success); CREATE INDEX IF NOT EXISTS idx_reasoning_created_at ON reasoning_sessions(created_at DESC); CREATE INDEX IF NOT EXISTS idx_reasoning_user_context ON reasoning_sessions(user_context); -- ============================================================================ -- Vector Embeddings for Neuro-Symbolic Integration -- ============================================================================ -- Store semantic embeddings for hybrid reasoning CREATE TABLE IF NOT EXISTS embeddings ( id UUID PRIMARY KEY DEFAULT uuid_generate_v4(), content TEXT NOT NULL, embedding vector(384), -- sentence-transformers default dimension embedding_model TEXT DEFAULT 'all-MiniLM-L6-v2', content_type TEXT, -- concept, statement, code, etc. related_triple_id UUID REFERENCES triples(id) ON DELETE CASCADE, created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP, metadata JSONB ); -- Vector similarity search index (HNSW for fast approximate nearest neighbor) CREATE INDEX IF NOT EXISTS idx_embeddings_vector ON embeddings USING hnsw (embedding vector_cosine_ops); CREATE INDEX IF NOT EXISTS idx_embeddings_content_type ON embeddings(content_type); CREATE INDEX IF NOT EXISTS idx_embeddings_related_triple ON embeddings(related_triple_id); -- ============================================================================ -- Compositional Patterns and Meta-Learning -- ============================================================================ -- Learn and store successful decomposition patterns CREATE TABLE IF NOT EXISTS patterns ( id UUID PRIMARY KEY DEFAULT uuid_generate_v4(), pattern_type TEXT NOT NULL, -- decomposition, proof, abstraction, loop_discovery pattern_data JSONB NOT NULL, success_rate FLOAT DEFAULT 0.0, usage_count INTEGER DEFAULT 0, last_used_at TIMESTAMP, created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP, metadata JSONB ); -- Indexes for pattern queries CREATE INDEX IF NOT EXISTS idx_patterns_type ON patterns(pattern_type); CREATE INDEX IF NOT EXISTS idx_patterns_success_rate ON patterns(success_rate DESC); CREATE INDEX IF NOT EXISTS idx_patterns_usage_count ON patterns(usage_count DESC); -- ============================================================================ -- Lambda Abstractions -- ============================================================================ -- Store lambda calculus representations CREATE TABLE IF NOT EXISTS lambda_abstractions ( id UUID PRIMARY KEY DEFAULT uuid_generate_v4(), original_text TEXT NOT NULL, lambda_expression TEXT NOT NULL, expression_type TEXT, -- function, composition, application, etc. type_signature TEXT, -- Type in Hindley-Milner notation created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP, metadata JSONB ); CREATE INDEX IF NOT EXISTS idx_lambda_original ON lambda_abstractions(original_text); CREATE INDEX IF NOT EXISTS idx_lambda_type ON lambda_abstractions(expression_type); -- ============================================================================ -- Concept Graph (Meta-Level Knowledge) -- ============================================================================ -- Track conceptual relationships discovered through loop discovery CREATE TABLE IF NOT EXISTS concept_nodes ( id UUID PRIMARY KEY DEFAULT uuid_generate_v4(), concept TEXT UNIQUE NOT NULL, concept_type TEXT, -- domain, method, principle, etc. description TEXT, related_triples JSONB, -- Array of triple IDs created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP, metadata JSONB ); CREATE TABLE IF NOT EXISTS concept_edges ( id UUID PRIMARY KEY DEFAULT uuid_generate_v4(), source_concept_id UUID NOT NULL REFERENCES concept_nodes(id) ON DELETE CASCADE, target_concept_id UUID NOT NULL REFERENCES concept_nodes(id) ON DELETE CASCADE, relationship_type TEXT NOT NULL, -- analogous_to, implements, specializes, etc. strength FLOAT DEFAULT 1.0, -- Strength of connection (0.0 - 1.0) evidence JSONB, -- Supporting evidence for this connection created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP, UNIQUE(source_concept_id, target_concept_id, relationship_type) ); CREATE INDEX IF NOT EXISTS idx_concept_edges_source ON concept_edges(source_concept_id); CREATE INDEX IF NOT EXISTS idx_concept_edges_target ON concept_edges(target_concept_id); CREATE INDEX IF NOT EXISTS idx_concept_edges_type ON concept_edges(relationship_type); -- ============================================================================ -- Database Metadata and Versioning -- ============================================================================ CREATE TABLE IF NOT EXISTS schema_version ( version INTEGER PRIMARY KEY, applied_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP, description TEXT ); -- Insert initial version INSERT INTO schema_version (version, description) VALUES (1, 'Initial Congo River schema with triples, proofs, embeddings, and patterns') ON CONFLICT DO NOTHING; -- ============================================================================ -- Utility Functions -- ============================================================================ -- Function to update updated_at timestamp automatically CREATE OR REPLACE FUNCTION update_updated_at_column() RETURNS TRIGGER AS $$ BEGIN NEW.updated_at = CURRENT_TIMESTAMP; RETURN NEW; END; $$ language 'plpgsql'; -- Trigger for triples table CREATE TRIGGER update_triples_updated_at BEFORE UPDATE ON triples FOR EACH ROW EXECUTE FUNCTION update_updated_at_column(); -- Function for similarity search helper CREATE OR REPLACE FUNCTION find_similar_embeddings( query_embedding vector(384), similarity_threshold float DEFAULT 0.7, max_results int DEFAULT 10 ) RETURNS TABLE ( id UUID, content TEXT, similarity FLOAT, metadata JSONB ) AS $$ BEGIN RETURN QUERY SELECT e.id, e.content, 1 - (e.embedding <=> query_embedding) as similarity, e.metadata FROM embeddings e WHERE 1 - (e.embedding <=> query_embedding) >= similarity_threshold ORDER BY e.embedding <=> query_embedding LIMIT max_results; END; $$ LANGUAGE plpgsql; -- ============================================================================ -- Views for Common Queries -- ============================================================================ -- View for recent reasoning activity CREATE OR REPLACE VIEW recent_reasoning AS SELECT rs.id, rs.tool_name, rs.success, rs.duration_ms, rs.created_at, rs.error_message FROM reasoning_sessions rs ORDER BY rs.created_at DESC LIMIT 100; -- View for successful proof patterns CREATE OR REPLACE VIEW successful_proofs AS SELECT p.id, p.goal, p.proof_method, p.duration_ms, p.created_at FROM proofs p WHERE p.success = true ORDER BY p.created_at DESC; -- View for knowledge graph statistics CREATE OR REPLACE VIEW knowledge_stats AS SELECT COUNT(*) as total_triples, COUNT(DISTINCT subject) as unique_subjects, COUNT(DISTINCT predicate) as unique_predicates, COUNT(DISTINCT object) as unique_objects, COUNT(DISTINCT context) as unique_contexts FROM triples;