AI Agent Generator

"""Smart Context module for AI-enhanced decision making""" import json from typing import Dict, Any, List from datetime import datetime class SmartContext: """Inteligentny kontekst do uczenia się wzorców i podpowiadania""" def __init__(self): self.learned_patterns = {} self.component_performance = {} self.domain_insights = {} async def get_smart_component_suggestions(self, description: str, domain: str, existing_component_ids: List[str] = None) -> List[Dict[str, Any]]: """Inteligentne sugestie komponentów na podstawie opisu i domeny""" if existing_component_ids is None: existing_component_ids = [] suggestions = [] # Podstawowe sugestie na podstawie słów kluczowych keywords = description.lower().split() if any(word in keywords for word in ['chat', 'conversation', 'talk', 'rozmowa', 'czat']): suggestions.extend([ {'component_id': 'llm_text_generator', 'reason': 'Conversation needs LLM', 'confidence': 90}, {'component_id': 'chat_interface', 'reason': 'Chat functionality', 'confidence': 85}, {'component_id': 'conversation_memory', 'reason': 'Context retention', 'confidence': 80} ]) if any(word in keywords for word in ['email', 'mail', 'wiadomość', 'newsletter', 'poczta']): suggestions.extend([ {'component_id': 'gmail_integration', 'reason': 'Gmail email handling', 'confidence': 90}, {'component_id': 'sendgrid_integration', 'reason': 'Bulk email sending', 'confidence': 85}, {'component_id': 'email_template_manager', 'reason': 'Template management', 'confidence': 80} ]) if any(word in keywords for word in ['calendar', 'schedule', 'kalendarz', 'terminarz']): suggestions.extend([ {'component_id': 'google_calendar_integration', 'reason': 'Calendar access', 'confidence': 90}, {'component_id': 'scheduling_system', 'reason': 'Appointment scheduling', 'confidence': 85} ]) # Filter out existing components filtered_suggestions = [s for s in suggestions if s['component_id'] not in existing_component_ids] # Return max 10 suggestions return filtered_suggestions[:10] async def get_intelligence_insights(self) -> Dict[str, Any]: """Zwraca aktualne insights AI""" return { "learned_patterns": list(self.learned_patterns.keys()), "success_metrics": { "total_patterns": len(self.learned_patterns), "component_performance_count": len(self.component_performance) }, "optimization_suggestions": [ "Use components with high success rates", "Learn from successful agent patterns", "Optimize based on domain-specific insights" ], "background_intelligence": "Active", "status": "full" } async def learn_from_successful_agent(self, agent: Dict[str, Any]): """Uczenie się z udanych agentów""" domain = agent.get('domain', 'general') components = agent.get('components', []) # Zapisz wzorzec sukcesu pattern_key = f"{domain}_{len(components)}_components" if pattern_key not in self.learned_patterns: self.learned_patterns[pattern_key] = { 'count': 0, 'components': {}, 'descriptions': [] } pattern = self.learned_patterns[pattern_key] pattern['count'] += 1 pattern['descriptions'].append(agent.get('description', '')) # Zlicz komponenty for comp in components: comp_id = comp.get('component_id', 'unknown') if comp_id not in pattern['components']: pattern['components'][comp_id] = 0 pattern['components'][comp_id] += 1 # Aktualizuj performance komponentów for comp in components: comp_id = comp.get('component_id', 'unknown') if comp_id not in self.component_performance: self.component_performance[comp_id] = { 'usage_count': 0, 'success_rate': 0.0, 'domains': [] } perf = self.component_performance[comp_id] perf['usage_count'] += 1 perf['success_rate'] = min(1.0, perf['success_rate'] + 0.1) if domain not in perf['domains']: perf['domains'].append(domain) async def get_domain_insights(self, domain: str) -> Dict[str, Any]: """Zwraca insights dla konkretnej domeny""" insights = { 'popular_components': [], 'success_patterns': [], 'recommendations': [] } # Znajdź popularne komponenty w tej domenie domain_components = {} for comp_id, perf in self.component_performance.items(): if domain in perf['domains']: domain_components[comp_id] = perf['usage_count'] # Sortuj po popularności popular = sorted(domain_components.items(), key=lambda x: x[1], reverse=True) insights['popular_components'] = [comp[0] for comp in popular[:5]] # Wzorce sukcesu for pattern_key, pattern in self.learned_patterns.items(): if domain in pattern_key: insights['success_patterns'].append({ 'pattern': pattern_key, 'count': pattern['count'], 'top_components': sorted(pattern['components'].items(), key=lambda x: x[1], reverse=True)[:3] }) # Rekomendacje insights['recommendations'] = [ "Używaj sprawdzonych komponentów z wysokim success_rate", "Rozważ komponenty popularne w tej domenie", "Testuj nowe kombinacje na podstawie udanych wzorców" ] return insights # Singleton instance _smart_context_instance = None def get_smart_context() -> SmartContext: """Zwraca singleton instance SmartContext""" global _smart_context_instance if _smart_context_instance is None: _smart_context_instance = SmartContext() return _smart_context_instance