🌊 The original agent harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, federation, vector RAG integration, and native Claude Code / Codex / Hermes and many more Integrated
TypeScript agentic-aiagentic-frameworkagentic-workflowagentsai-agentsai-assistantai-skillsautonomous-agentsclaudecodexdsh-pluginharnessmcp-servermulti-agentmulti-agent-systemsnpmskillsswarmswarm-intelligencetypescript
Checkpoint: File edits
rUv4b0ad3ce90c0
diff --git a/docs/ruvector-postgres/scripts/init-db.sql b/docs/ruvector-postgres/scripts/init-db.sqlindex e98e26e30c..f2715cd7ac 100644--- a/docs/ruvector-postgres/scripts/init-db.sql+++ b/docs/ruvector-postgres/scripts/init-db.sql@@ -2,18 +2,15 @@-- RUVECTOR POSTGRESQL INITIALIZATION SCRIPT-- ============================================----- This script initializes the RuVector PostgreSQL extension--- from ruvnet/ruvector with full V3 Claude-Flow integration.+-- This script initializes RuVector PostgreSQL extension+-- from ruvnet/ruvector-postgres with Claude-Flow V3 integration.----- RuVector provides 77+ SQL functions including:--- - Vector similarity search (HNSW/IVFFlat)--- - 39 Attention mechanisms--- - GNN layers for graph operations+-- RuVector 2.0.0 provides 77+ SQL functions including:+-- - Vector similarity search (HNSW with SIMD)-- - Hyperbolic embeddings (Poincaré/Lorentz)--- - Sparse vectors (BM25/TF-IDF)--- - SPARQL support (50+ RDF functions)--- - Local embeddings (6 fastembed models)--- - Self-learning index optimization+-- - Graph operations (Cypher queries)+-- - RDF/SPARQL support+-- - Agent routing and learning---- Performance: ~61µs latency, 16,400 QPS@@ -21,8 +18,8 @@-- PART 1: EXTENSION AND SCHEMA SETUP-- ============================================--- Enable RuVector extension (includes vector type)-CREATE EXTENSION IF NOT EXISTS ruvector;+-- Enable RuVector extension (use VERSION to work around control file issue)+CREATE EXTENSION IF NOT EXISTS ruvector VERSION '0.1.0';-- Enable additional required extensionsCREATE EXTENSION IF NOT EXISTS pgcrypto;@@ -32,21 +29,19 @@ CREATE SCHEMA IF NOT EXISTS claude_flow;-- Grant permissionsGRANT ALL ON SCHEMA claude_flow TO claude;-GRANT USAGE ON SCHEMA ruvector TO claude;--- Set search path for convenience-SET search_path TO claude_flow, ruvector, public;+-- Set search path+SET search_path TO claude_flow, public;-- ============================================-- PART 2: CORE TABLES-- ============================================--- Embeddings table with RuVector vector type+-- Embeddings table with RuVector vector type (384-dim for all-MiniLM-L6-v2)CREATE TABLE IF NOT EXISTS claude_flow.embeddings (id UUID PRIMARY KEY DEFAULT gen_random_uuid(),content TEXT NOT NULL,- embedding ruvector.vector(384), -- RuVector vector type- sparse_embedding ruvector.sparse_vector, -- For hybrid BM25/TF-IDF search+ embedding ruvector(384),metadata JSONB DEFAULT '{}',created_at TIMESTAMPTZ DEFAULT NOW(),updated_at TIMESTAMPTZ DEFAULT NOW()@@ -57,12 +52,12 @@ CREATE TABLE IF NOT EXISTS claude_flow.patterns (id UUID PRIMARY KEY DEFAULT gen_random_uuid(),name VARCHAR(255) NOT NULL,description TEXT,- embedding ruvector.vector(384),+ embedding ruvector(384),pattern_type VARCHAR(50),confidence FLOAT DEFAULT 0.5,success_count INT DEFAULT 0,failure_count INT DEFAULT 0,- ewc_importance FLOAT DEFAULT 1.0, -- EWC++ weight+ ewc_importance FLOAT DEFAULT 1.0,metadata JSONB DEFAULT '{}',created_at TIMESTAMPTZ DEFAULT NOW());@@ -73,7 +68,7 @@ CREATE TABLE IF NOT EXISTS claude_flow.agents (agent_id VARCHAR(255) NOT NULL UNIQUE,agent_type VARCHAR(50),state JSONB DEFAULT '{}',- memory_embedding ruvector.vector(384),+ memory_embedding ruvector(384),last_active TIMESTAMPTZ DEFAULT NOW(),created_at TIMESTAMPTZ DEFAULT NOW());@@ -96,9 +91,8 @@ CREATE TABLE IF NOT EXISTS claude_flow.trajectories (CREATE TABLE IF NOT EXISTS claude_flow.hyperbolic_embeddings (id UUID PRIMARY KEY DEFAULT gen_random_uuid(),content TEXT NOT NULL,- euclidean_embedding ruvector.vector(384),- poincare_embedding ruvector.vector(384), -- Poincaré ball (||x|| < 1)- lorentz_embedding ruvector.vector(385), -- Lorentz model (+1 dim)+ euclidean_embedding ruvector(384),+ poincare_embedding real[], -- Array for hyperbolic operationscurvature FLOAT DEFAULT -1.0,hierarchy_level INT DEFAULT 0,parent_id UUID REFERENCES claude_flow.hyperbolic_embeddings(id),@@ -111,12 +105,12 @@ CREATE TABLE IF NOT EXISTS claude_flow.graph_nodes (id UUID PRIMARY KEY DEFAULT gen_random_uuid(),node_id VARCHAR(255) NOT NULL UNIQUE,node_type VARCHAR(50),- embedding ruvector.vector(384),+ embedding ruvector(384),features JSONB DEFAULT '{}',created_at TIMESTAMPTZ DEFAULT NOW());--- Graph edges for GNN message passing+-- Graph edges for message passingCREATE TABLE IF NOT EXISTS claude_flow.graph_edges (id UUID PRIMARY KEY DEFAULT gen_random_uuid(),source_id UUID REFERENCES claude_flow.graph_nodes(id),@@ -128,52 +122,46 @@ CREATE TABLE IF NOT EXISTS claude_flow.graph_edges ();-- ============================================--- PART 3: RUVECTOR INDICES (HNSW + IVFFlat)+-- PART 3: HNSW INDICES-- ============================================--- HNSW index for embeddings (~61µs search latency)+-- HNSW index for embeddings (cosine distance)CREATE INDEX IF NOT EXISTS idx_embeddings_hnswON claude_flow.embeddings-USING ruvector_hnsw (embedding ruvector.cosine_ops)+USING hnsw (embedding ruvector_cosine_ops)WITH (m = 16, ef_construction = 100);-- HNSW index for patternsCREATE INDEX IF NOT EXISTS idx_patterns_hnswON claude_flow.patterns-USING ruvector_hnsw (embedding ruvector.cosine_ops)+USING hnsw (embedding ruvector_cosine_ops)WITH (m = 16, ef_construction = 100);-- HNSW index for agent memoryCREATE INDEX IF NOT EXISTS idx_agents_hnswON claude_flow.agents-USING ruvector_hnsw (memory_embedding ruvector.cosine_ops)+USING hnsw (memory_embedding ruvector_cosine_ops)WITH (m = 16, ef_construction = 64);-- HNSW index for hyperbolic embeddingsCREATE INDEX IF NOT EXISTS idx_hyperbolic_hnswON claude_flow.hyperbolic_embeddings-USING ruvector_hnsw (euclidean_embedding ruvector.cosine_ops)+USING hnsw (euclidean_embedding ruvector_cosine_ops)WITH (m = 16, ef_construction = 100);-- HNSW index for graph nodesCREATE INDEX IF NOT EXISTS idx_graph_nodes_hnswON claude_flow.graph_nodes-USING ruvector_hnsw (embedding ruvector.cosine_ops)+USING hnsw (embedding ruvector_cosine_ops)WITH (m = 16, ef_construction = 64);--- IVFFlat index for memory-efficient search (optional)--- CREATE INDEX IF NOT EXISTS idx_embeddings_ivfflat--- ON claude_flow.embeddings--- USING ruvector_ivfflat (embedding ruvector.cosine_ops)--- WITH (lists = 100);--- ============================================--- PART 4: RUVECTOR CORE FUNCTIONS+-- PART 4: CORE SEARCH FUNCTIONS-- ============================================--- Semantic similarity search using RuVector+-- Semantic similarity search using RuVector HNSWCREATE OR REPLACE FUNCTION claude_flow.search_similar(- query_embedding ruvector.vector(384),+ query_embedding ruvector(384),limit_count INT DEFAULT 10,min_similarity FLOAT DEFAULT 0.5)@@ -188,247 +176,106 @@ BEGINSELECTe.id,e.content,- ruvector.cosine_similarity(e.embedding, query_embedding)::FLOAT AS similarity,+ (1 - (e.embedding <=> query_embedding))::FLOAT AS similarity,e.metadataFROM claude_flow.embeddings eWHERE e.embedding IS NOT NULL- AND ruvector.cosine_similarity(e.embedding, query_embedding) >= min_similarity- ORDER BY e.embedding <-> query_embedding -- RuVector distance operator+ AND (1 - (e.embedding <=> query_embedding)) >= min_similarity+ ORDER BY e.embedding <=> query_embeddingLIMIT limit_count;END;$$ LANGUAGE plpgsql STABLE;--- Hybrid search (vector + BM25 sparse)-CREATE OR REPLACE FUNCTION claude_flow.hybrid_search(- query_embedding ruvector.vector(384),- query_text TEXT,+-- Pattern search with type filtering+CREATE OR REPLACE FUNCTION claude_flow.search_patterns(+ query_embedding ruvector(384),+ pattern_type_filter VARCHAR(50) DEFAULT NULL,limit_count INT DEFAULT 10,- vector_weight FLOAT DEFAULT 0.7+ min_confidence FLOAT DEFAULT 0.5)RETURNS TABLE (id UUID,- content TEXT,- vector_score FLOAT,- sparse_score FLOAT,- combined_score FLOAT,- metadata JSONB-) AS $$-BEGIN- RETURN QUERY- SELECT- e.id,- e.content,- ruvector.cosine_similarity(e.embedding, query_embedding)::FLOAT AS v_score,- COALESCE(ruvector.bm25_score(e.sparse_embedding, query_text), 0)::FLOAT AS s_score,- (- vector_weight * ruvector.cosine_similarity(e.embedding, query_embedding) +- (1 - vector_weight) * COALESCE(ruvector.bm25_score(e.sparse_embedding, query_text), 0)- )::FLOAT AS c_score,- e.metadata- FROM claude_flow.embeddings e- WHERE e.embedding IS NOT NULL- ORDER BY c_score DESC- LIMIT limit_count;-END;-$$ LANGUAGE plpgsql STABLE;---- ============================================--- PART 5: ATTENTION MECHANISM FUNCTIONS--- ============================================---- Self-attention using RuVector's attention functions-CREATE OR REPLACE FUNCTION claude_flow.self_attention(- source_ids UUID[],- temperature FLOAT DEFAULT 1.0-)-RETURNS TABLE (- source_id UUID,- target_id UUID,- attention_weight FLOAT-) AS $$-BEGIN- RETURN QUERY- WITH sources AS (- SELECT id, embedding- FROM claude_flow.embeddings- WHERE id = ANY(source_ids)- AND embedding IS NOT NULL- )- SELECT- s1.id AS source_id,- s2.id AS target_id,- ruvector.softmax_attention(s1.embedding, s2.embedding, temperature)::FLOAT AS attention_weight- FROM sources s1- CROSS JOIN sources s2;-END;-$$ LANGUAGE plpgsql STABLE;---- Multi-head attention-CREATE OR REPLACE FUNCTION claude_flow.multihead_attention(- query_embedding ruvector.vector(384),- num_heads INT DEFAULT 8,- limit_count INT DEFAULT 10-)-RETURNS TABLE (- id UUID,- content TEXT,- head_attentions FLOAT[],- aggregated_attention FLOAT,+ name VARCHAR(255),+ description TEXT,+ similarity FLOAT,+ confidence FLOAT,metadata JSONB) AS $$BEGINRETURN QUERYSELECT- e.id,- e.content,- ruvector.multihead_attention_scores(query_embedding, e.embedding, num_heads)::FLOAT[] AS head_scores,- ruvector.multihead_attention_aggregate(query_embedding, e.embedding, num_heads)::FLOAT AS agg_score,- e.metadata- FROM claude_flow.embeddings e- WHERE e.embedding IS NOT NULL- ORDER BY agg_score DESC+ p.id,+ p.name,+ p.description,+ (1 - (p.embedding <=> query_embedding))::FLOAT AS similarity,+ p.confidence,+ p.metadata+ FROM claude_flow.patterns p+ WHERE p.embedding IS NOT NULL+ AND p.confidence >= min_confidence+ AND (pattern_type_filter IS NULL OR p.pattern_type = pattern_type_filter)+ ORDER BY p.embedding <=> query_embeddingLIMIT limit_count;END;$$ LANGUAGE plpgsql STABLE;--- Cross-attention between query and key-value sets-CREATE OR REPLACE FUNCTION claude_flow.cross_attention(- query_ids UUID[],- kv_filter JSONB DEFAULT NULL,- temperature FLOAT DEFAULT 1.0+-- Agent routing by expertise similarity+CREATE OR REPLACE FUNCTION claude_flow.find_agents(+ query_embedding ruvector(384),+ agent_type_filter VARCHAR(50) DEFAULT NULL,+ limit_count INT DEFAULT 5)RETURNS TABLE (- query_id UUID,- kv_id UUID,- attention_weight FLOAT-) AS $$-BEGIN- RETURN QUERY- WITH queries AS (- SELECT id, embedding- FROM claude_flow.embeddings- WHERE id = ANY(query_ids) AND embedding IS NOT NULL- ),- kv AS (- SELECT id, embedding- FROM claude_flow.embeddings- WHERE embedding IS NOT NULL- AND id != ALL(query_ids)- AND (kv_filter IS NULL OR metadata @> kv_filter)- )- SELECT- q.id AS query_id,- kv.id AS kv_id,- ruvector.cross_attention_score(q.embedding, kv.embedding, temperature)::FLOAT AS attention_weight- FROM queries q- CROSS JOIN kv;-END;-$$ LANGUAGE plpgsql STABLE;---- ============================================--- PART 6: GNN OPERATIONS--- ============================================---- Message passing for GNN-CREATE OR REPLACE FUNCTION claude_flow.gnn_message_pass(- node_id VARCHAR(255),- aggregation TEXT DEFAULT 'mean' -- 'mean', 'sum', 'max'-)-RETURNS ruvector.vector(384) AS $$-DECLARE- v_result ruvector.vector(384);- v_node_uuid UUID;-BEGIN- -- Get node UUID- SELECT id INTO v_node_uuid- FROM claude_flow.graph_nodes- WHERE graph_nodes.node_id = gnn_message_pass.node_id;-- -- Aggregate neighbor embeddings using RuVector GNN functions- SELECT ruvector.gnn_aggregate(- ARRAY_AGG(n.embedding),- aggregation- )- INTO v_result- FROM claude_flow.graph_edges e- JOIN claude_flow.graph_nodes n ON n.id = e.target_id- WHERE e.source_id = v_node_uuid;-- RETURN v_result;-END;-$$ LANGUAGE plpgsql STABLE;---- Graph attention network (GAT) scoring-CREATE OR REPLACE FUNCTION claude_flow.gat_attention(- source_node VARCHAR(255),- limit_count INT DEFAULT 10-)-RETURNS TABLE (- neighbor_id VARCHAR(255),- attention_weight FLOAT,- edge_type VARCHAR(50)+ agent_id VARCHAR(255),+ agent_type VARCHAR(50),+ similarity FLOAT,+ state JSONB) AS $$BEGINRETURN QUERY- WITH source AS (- SELECT id, embedding- FROM claude_flow.graph_nodes- WHERE node_id = source_node- )SELECT- n.node_id AS neighbor_id,- ruvector.gat_score(s.embedding, n.embedding)::FLOAT AS attention_weight,- e.edge_type- FROM source s- JOIN claude_flow.graph_edges e ON e.source_id = s.id- JOIN claude_flow.graph_nodes n ON n.id = e.target_id- ORDER BY attention_weight DESC+ a.agent_id,+ a.agent_type,+ (1 - (a.memory_embedding <=> query_embedding))::FLOAT AS similarity,+ a.state+ FROM claude_flow.agents a+ WHERE a.memory_embedding IS NOT NULL+ AND (agent_type_filter IS NULL OR a.agent_type = agent_type_filter)+ ORDER BY a.memory_embedding <=> query_embeddingLIMIT limit_count;END;$$ LANGUAGE plpgsql STABLE;-- ============================================--- PART 7: HYPERBOLIC OPERATIONS+-- PART 5: HYPERBOLIC OPERATIONS-- ============================================-- Convert Euclidean to Poincaré embeddingCREATE OR REPLACE FUNCTION claude_flow.to_poincare(- euclidean ruvector.vector(384),+ euclidean real[],curvature FLOAT DEFAULT -1.0)-RETURNS ruvector.vector(384) AS $$+RETURNS real[] AS $$BEGIN- RETURN ruvector.exp_map_poincare(euclidean, curvature);+ RETURN ruvector_exp_map(ARRAY_FILL(0.0::real, ARRAY[array_length(euclidean, 1)]), euclidean, curvature);END;$$ LANGUAGE plpgsql IMMUTABLE;-- Poincaré distance (geodesic)CREATE OR REPLACE FUNCTION claude_flow.poincare_distance(- x ruvector.vector(384),- y ruvector.vector(384),+ x real[],+ y real[],curvature FLOAT DEFAULT -1.0)RETURNS FLOAT AS $$BEGIN- RETURN ruvector.hyperbolic_distance(x, y, 'poincare', curvature);+ RETURN ruvector_poincare_distance(x, y, curvature);END;$$ LANGUAGE plpgsql IMMUTABLE;--- Hyperbolic midpoint (Möbius addition)-CREATE OR REPLACE FUNCTION claude_flow.poincare_midpoint(- x ruvector.vector(384),- y ruvector.vector(384),- curvature FLOAT DEFAULT -1.0-)-RETURNS ruvector.vector(384) AS $$-BEGIN- RETURN ruvector.mobius_add(x, y, curvature);-END;-$$ LANGUAGE plpgsql IMMUTABLE;---- Hierarchical search in hyperbolic space+-- Hyperbolic search in Poincaré ballCREATE OR REPLACE FUNCTION claude_flow.hyperbolic_search(- query ruvector.vector(384),+ query ruvector(384),limit_count INT DEFAULT 10,curvature FLOAT DEFAULT -1.0)@@ -441,82 +288,92 @@ RETURNS TABLE (metadata JSONB) AS $$DECLARE- v_query_poincare ruvector.vector(384);+ query_arr real[];+ query_poincare real[];BEGIN- v_query_poincare := ruvector.exp_map_poincare(query, curvature);+ -- Convert query to array and then to Poincaré+ SELECT array_agg(x ORDER BY ordinality) INTO query_arr+ FROM unnest(string_to_array(trim(both '[]' from query::text), ',')) WITH ORDINALITY AS t(x, ordinality);++ query_poincare := claude_flow.to_poincare(query_arr, curvature);RETURN QUERYSELECThe.id,he.content,(he.euclidean_embedding <-> query)::FLOAT AS euc_dist,- ruvector.hyperbolic_distance(he.poincare_embedding, v_query_poincare, 'poincare', curvature)::FLOAT AS hyp_dist,+ COALESCE(ruvector_poincare_distance(he.poincare_embedding, query_poincare, curvature), 999.0)::FLOAT AS hyp_dist,he.hierarchy_level,he.metadataFROM claude_flow.hyperbolic_embeddings heWHERE he.euclidean_embedding IS NOT NULL- ORDER BY hyp_dist+ ORDER BY he.euclidean_embedding <-> queryLIMIT limit_count;END;$$ LANGUAGE plpgsql STABLE;-- ============================================--- PART 8: LOCAL EMBEDDING GENERATION+-- PART 6: UTILITY FUNCTIONS-- ============================================--- Generate embedding using RuVector's local fastembed models-CREATE OR REPLACE FUNCTION claude_flow.embed_text(- text_content TEXT,- model_name TEXT DEFAULT 'all-MiniLM-L6-v2'-)-RETURNS ruvector.vector(384) AS $$+-- Get RuVector version info+CREATE OR REPLACE FUNCTION claude_flow.ruvector_info()+RETURNS TABLE (+ version TEXT,+ simd_info TEXT+) AS $$BEGIN- RETURN ruvector.fastembed(text_content, model_name);+ RETURN QUERY+ SELECT ruvector_version(), ruvector_simd_info();END;$$ LANGUAGE plpgsql STABLE;--- Batch embed multiple texts-CREATE OR REPLACE FUNCTION claude_flow.embed_texts_batch(- texts TEXT[],- model_name TEXT DEFAULT 'all-MiniLM-L6-v2'+-- Cosine similarity helper (converts cosine distance to similarity)+CREATE OR REPLACE FUNCTION claude_flow.cosine_similarity(+ a ruvector,+ b ruvector)-RETURNS TABLE (- idx INT,- embedding ruvector.vector(384)-) AS $$+RETURNS FLOAT AS $$BEGIN- RETURN QUERY- SELECT- ordinality::INT AS idx,- ruvector.fastembed(t, model_name) AS embedding- FROM unnest(texts) WITH ORDINALITY AS t;+ RETURN (1 - (a <=> b))::FLOAT;END;-$$ LANGUAGE plpgsql STABLE;+$$ LANGUAGE plpgsql IMMUTABLE;--- ============================================--- PART 9: SELF-LEARNING INDEX OPTIMIZATION--- ============================================+-- L2 distance helper+CREATE OR REPLACE FUNCTION claude_flow.l2_distance(+ a ruvector,+ b ruvector+)+RETURNS FLOAT AS $$+BEGIN+ RETURN (a <-> b)::FLOAT;+END;+$$ LANGUAGE plpgsql IMMUTABLE;--- Trigger self-learning optimization-CREATE OR REPLACE FUNCTION claude_flow.optimize_indices()+-- Update embedding with auto-timestamp+CREATE OR REPLACE FUNCTION claude_flow.update_embedding(+ embedding_id UUID,+ new_content TEXT DEFAULT NULL,+ new_embedding ruvector(384) DEFAULT NULL,+ new_metadata JSONB DEFAULT NULL+)RETURNS void AS $$BEGIN- -- RuVector self-learning index optimization- PERFORM ruvector.learn_optimize('claude_flow.embeddings', 'embedding');- PERFORM ruvector.learn_optimize('claude_flow.patterns', 'embedding');- PERFORM ruvector.learn_optimize('claude_flow.agents', 'memory_embedding');+ UPDATE claude_flow.embeddings+ SET+ content = COALESCE(new_content, content),+ embedding = COALESCE(new_embedding, embedding),+ metadata = COALESCE(new_metadata, metadata),+ updated_at = NOW()+ WHERE id = embedding_id;END;$$ LANGUAGE plpgsql;--- Schedule periodic optimization (every hour)--- Note: Requires pg_cron extension or external scheduler--- SELECT cron.schedule('ruvector-optimize', '0 * * * *', 'SELECT claude_flow.optimize_indices()');--- ============================================--- PART 10: SAMPLE DATA+-- PART 7: SAMPLE DATA-- ============================================--- Insert sample embeddings for testing+-- Insert sample embeddings (without actual vectors - those need to come from model)INSERT INTO claude_flow.embeddings (content, metadata) VALUES('Implementing JWT authentication with refresh token rotation','{"category": "security", "language": "typescript", "agent": "security-architect"}'),@@ -563,26 +420,33 @@ ON CONFLICT DO NOTHING;-- ============================================DO $$+DECLARE+ v_version TEXT;+ v_simd TEXT;BEGIN+ SELECT ruvector_version() INTO v_version;+ SELECT ruvector_simd_info() INTO v_simd;+RAISE NOTICE '';RAISE NOTICE '============================================';RAISE NOTICE 'RuVector PostgreSQL Initialization Complete!';RAISE NOTICE '============================================';RAISE NOTICE '';- RAISE NOTICE 'Extension: ruvector (77+ SQL functions)';+ RAISE NOTICE 'RuVector Version: %', v_version;+ RAISE NOTICE 'SIMD: %', v_simd;+ RAISE NOTICE '';RAISE NOTICE 'Schema: claude_flow';- RAISE NOTICE 'Tables: embeddings, patterns, agents, trajectories, hyperbolic_embeddings, graph_nodes, graph_edges';+ RAISE NOTICE 'Tables: embeddings, patterns, agents, trajectories,';+ RAISE NOTICE ' hyperbolic_embeddings, graph_nodes, graph_edges';RAISE NOTICE 'Indices: 5 HNSW indices created';RAISE NOTICE '';- RAISE NOTICE 'Features enabled:';- RAISE NOTICE ' - HNSW/IVFFlat vector indexing';- RAISE NOTICE ' - 39 Attention mechanisms';- RAISE NOTICE ' - GNN operations';- RAISE NOTICE ' - Hyperbolic embeddings (Poincare/Lorentz)';- RAISE NOTICE ' - Sparse vectors (BM25/TF-IDF)';- RAISE NOTICE ' - Local embeddings (fastembed)';- RAISE NOTICE ' - Self-learning optimization';+ RAISE NOTICE 'Key Functions:';+ RAISE NOTICE ' - claude_flow.search_similar(embedding, limit, min_sim)';+ RAISE NOTICE ' - claude_flow.search_patterns(embedding, type, limit)';+ RAISE NOTICE ' - claude_flow.find_agents(embedding, type, limit)';+ RAISE NOTICE ' - claude_flow.hyperbolic_search(embedding, limit, curvature)';+ RAISE NOTICE ' - claude_flow.cosine_similarity(a, b)';RAISE NOTICE '';- RAISE NOTICE 'Performance: ~61us latency, 16,400 QPS';+ RAISE NOTICE 'Operators: <=> (cosine), <-> (L2), <#> (neg inner product)';RAISE NOTICE '';END $$;