d14150da75
Implement intelligent agent learning from Knowledge Graph execution history with per-task-type expertise tracking, recency bias, and learning curves. ## Phase 5.3 Implementation ### Learning Infrastructure (✅ Complete) - LearningProfileService with per-task-type expertise metrics - TaskTypeExpertise model tracking success_rate, confidence, learning curves - Recency bias weighting: recent 7 days weighted 3x higher (exponential decay) - Confidence scoring prevents overfitting: min(1.0, executions / 20) - Learning curves computed from daily execution windows ### Agent Scoring Service (✅ Complete) - Unified AgentScore combining SwarmCoordinator + learning profiles - Scoring formula: 0.3*base + 0.5*expertise + 0.2*confidence - Rank agents by combined score for intelligent assignment - Support for recency-biased scoring (recent_success_rate) - Methods: rank_agents, select_best, rank_agents_with_recency ### KG Integration (✅ Complete) - KGPersistence::get_executions_for_task_type() - query by agent + task type - KGPersistence::get_agent_executions() - all executions for agent - Coordinator::load_learning_profile_from_kg() - core KG→Learning integration - Coordinator::load_all_learning_profiles() - batch load for multiple agents - Convert PersistedExecution → ExecutionData for learning calculations ### Agent Assignment Integration (✅ Complete) - AgentCoordinator uses learning profiles for task assignment - extract_task_type() infers task type from title/description - assign_task() scores candidates using AgentScoringService - Fallback to load-based selection if no learning data available - Learning profiles stored in coordinator.learning_profiles RwLock ### Profile Adapter Enhancements (✅ Complete) - create_learning_profile() - initialize empty profiles - add_task_type_expertise() - set task-type expertise - update_profile_with_learning() - update swarm profiles from learning ## Files Modified ### vapora-knowledge-graph/src/persistence.rs (+30 lines) - get_executions_for_task_type(agent_id, task_type, limit) - get_agent_executions(agent_id, limit) ### vapora-agents/src/coordinator.rs (+100 lines) - load_learning_profile_from_kg() - core KG integration method - load_all_learning_profiles() - batch loading for agents - assign_task() already uses learning-based scoring via AgentScoringService ### Existing Complete Implementation - vapora-knowledge-graph/src/learning.rs - calculation functions - vapora-agents/src/learning_profile.rs - data structures and expertise - vapora-agents/src/scoring.rs - unified scoring service - vapora-agents/src/profile_adapter.rs - adapter methods ## Tests Passing - learning_profile: 7 tests ✅ - scoring: 5 tests ✅ - profile_adapter: 6 tests ✅ - coordinator: learning-specific tests ✅ ## Data Flow 1. Task arrives → AgentCoordinator::assign_task() 2. Extract task_type from description 3. Query KG for task-type executions (load_learning_profile_from_kg) 4. Calculate expertise with recency bias 5. Score candidates (SwarmCoordinator + learning) 6. Assign to top-scored agent 7. Execution result → KG → Update learning profiles ## Key Design Decisions ✅ Recency bias: 7-day half-life with 3x weight for recent performance ✅ Confidence scoring: min(1.0, total_executions / 20) prevents overfitting ✅ Hierarchical scoring: 30% base load, 50% expertise, 20% confidence ✅ KG query limit: 100 recent executions per task-type for performance ✅ Async loading: load_learning_profile_from_kg supports concurrent loads ## Next: Phase 5.4 - Cost Optimization Ready to implement budget enforcement and cost-aware provider selection.
66 lines
3.7 KiB
Plaintext
66 lines
3.7 KiB
Plaintext
-- Migration 004: RAG (Retrieval-Augmented Generation)
|
|
-- Creates tables for document storage and semantic search
|
|
|
|
-- Documents table
|
|
DEFINE TABLE documents SCHEMAFULL
|
|
PERMISSIONS
|
|
FOR select WHERE tenant_id = $auth.tenant_id
|
|
FOR create, update, delete WHERE tenant_id = $auth.tenant_id;
|
|
|
|
DEFINE FIELD id ON TABLE documents TYPE record<documents>;
|
|
DEFINE FIELD tenant_id ON TABLE documents TYPE string ASSERT $value != NONE;
|
|
DEFINE FIELD project_id ON TABLE documents TYPE option<string>;
|
|
DEFINE FIELD title ON TABLE documents TYPE string ASSERT $value != NONE;
|
|
DEFINE FIELD content ON TABLE documents TYPE string ASSERT $value != NONE;
|
|
DEFINE FIELD content_type ON TABLE documents TYPE string ASSERT $value INSIDE ["markdown", "code", "text", "json"] DEFAULT "text";
|
|
DEFINE FIELD metadata ON TABLE documents TYPE object DEFAULT {};
|
|
DEFINE FIELD embedding ON TABLE documents TYPE option<array<float>>;
|
|
DEFINE FIELD source_path ON TABLE documents TYPE option<string>;
|
|
DEFINE FIELD tags ON TABLE documents TYPE array<string> DEFAULT [];
|
|
DEFINE FIELD created_at ON TABLE documents TYPE datetime DEFAULT time::now();
|
|
DEFINE FIELD updated_at ON TABLE documents TYPE datetime DEFAULT time::now() VALUE time::now();
|
|
|
|
DEFINE INDEX idx_documents_tenant ON TABLE documents COLUMNS tenant_id;
|
|
DEFINE INDEX idx_documents_project ON TABLE documents COLUMNS project_id;
|
|
DEFINE INDEX idx_documents_content_type ON TABLE documents COLUMNS content_type;
|
|
DEFINE INDEX idx_documents_tags ON TABLE documents COLUMNS tags;
|
|
|
|
-- Vector index for semantic search (HNSW)
|
|
-- Note: SurrealDB 2.x+ supports vector search with MTREE indexes
|
|
DEFINE INDEX idx_documents_embedding ON TABLE documents FIELDS embedding MTREE DIMENSION 1536;
|
|
|
|
-- Document chunks table (for large documents split into chunks)
|
|
DEFINE TABLE document_chunks SCHEMAFULL
|
|
PERMISSIONS
|
|
FOR select WHERE $parent.tenant_id = $auth.tenant_id
|
|
FOR create, update, delete WHERE $parent.tenant_id = $auth.tenant_id;
|
|
|
|
DEFINE FIELD id ON TABLE document_chunks TYPE record<document_chunks>;
|
|
DEFINE FIELD document_id ON TABLE document_chunks TYPE string ASSERT $value != NONE;
|
|
DEFINE FIELD chunk_index ON TABLE document_chunks TYPE int ASSERT $value >= 0;
|
|
DEFINE FIELD content ON TABLE document_chunks TYPE string ASSERT $value != NONE;
|
|
DEFINE FIELD embedding ON TABLE document_chunks TYPE option<array<float>>;
|
|
DEFINE FIELD token_count ON TABLE document_chunks TYPE option<int>;
|
|
DEFINE FIELD created_at ON TABLE document_chunks TYPE datetime DEFAULT time::now();
|
|
|
|
DEFINE INDEX idx_document_chunks_document ON TABLE document_chunks COLUMNS document_id;
|
|
DEFINE INDEX idx_document_chunks_document_index ON TABLE document_chunks COLUMNS document_id, chunk_index UNIQUE;
|
|
DEFINE INDEX idx_document_chunks_embedding ON TABLE document_chunks FIELDS embedding MTREE DIMENSION 1536;
|
|
|
|
-- Search history table (for analytics and improvement)
|
|
DEFINE TABLE search_history SCHEMAFULL
|
|
PERMISSIONS
|
|
FOR select WHERE tenant_id = $auth.tenant_id
|
|
FOR create WHERE tenant_id = $auth.tenant_id;
|
|
|
|
DEFINE FIELD id ON TABLE search_history TYPE record<search_history>;
|
|
DEFINE FIELD tenant_id ON TABLE search_history TYPE string ASSERT $value != NONE;
|
|
DEFINE FIELD query ON TABLE search_history TYPE string ASSERT $value != NONE;
|
|
DEFINE FIELD results_count ON TABLE search_history TYPE int DEFAULT 0;
|
|
DEFINE FIELD top_result_id ON TABLE search_history TYPE option<string>;
|
|
DEFINE FIELD search_time_ms ON TABLE search_history TYPE int;
|
|
DEFINE FIELD created_at ON TABLE search_history TYPE datetime DEFAULT time::now();
|
|
|
|
DEFINE INDEX idx_search_history_tenant ON TABLE search_history COLUMNS tenant_id;
|
|
DEFINE INDEX idx_search_history_created ON TABLE search_history COLUMNS created_at;
|