File size: 6,777 Bytes
25732fb
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164

import sys
import random
import time
from unittest.mock import MagicMock

# MOCK BROKEN DEPENDENCIES BEFORE IMPORTING APP
# This is necessary because google-generativeai depends on protobuf which is broken on Python 3.14
mock_genai = MagicMock()
sys.modules['google'] = MagicMock()
sys.modules['google.generativeai'] = mock_genai
sys.modules['google.ai'] = MagicMock()
sys.modules['google.api_core'] = MagicMock()

from app import app, db
from models import User, Topic, AgentKnowledge, AgentPerformance, KnowledgeState, LearningSession, QuizAttempt
from agents.coordinator_agent import CoordinatorAgent

# Mock LLM Service to avoid API calls during simulation
class MockLLMService:
    def generate_lesson_with_prompt(self, topic, difficulty, knowledge_level, custom_prompt=None):
        return f"Simulated lesson content for {topic} ({difficulty})"
        
    def generate_quiz_with_prompt(self, topic, num_questions, custom_prompt=None):
        return [{"question": "Simulated Q", "options": ["A", "B"], "answer": "A"}] * num_questions

    def generate_content(self, prompt):
        return "Simulated content"

class SimulatedStudent:
    def __init__(self, profile_type):
        self.profile_type = profile_type
        self.id = int(time.time()) + random.randint(1000, 9999) # Fake ID
        self.knowledge = {} # topic_id -> level
        
        # Define characteristics
        if profile_type == "fast_visual":
            self.learning_rate = 1.5
            self.preferred_style = "visual"
            self.patience = 3
        elif profile_type == "slow_detailed":
            self.learning_rate = 0.8
            self.preferred_style = "reading"
            self.patience = 8
        else: # average
            self.learning_rate = 1.0
            self.preferred_style = "mixed"
            self.patience = 5
            
    def attempt_quiz(self, difficulty, content_quality):
        """Simulate quiz result based on difficulty vs learning match"""
        base_success = 0.7 * self.learning_rate
        
        # Adjust based on content quality match
        if self.preferred_style in content_quality.get('style', ''):
            base_success += 0.2
            
        # Adjust based on difficulty
        if difficulty == "advanced": base_success -= 0.3
        elif difficulty == "intermediate": base_success -= 0.1
        
        # Random variance
        success_prob = min(max(base_success + random.uniform(-0.1, 0.1), 0.1), 0.95)
        return random.random() < success_prob

class TrainingSimulator:
    def __init__(self):
        self.coordinator = CoordinatorAgent()
        
        # Patch agents with Mock Service
        mock_llm = MockLLMService()
        self.coordinator.teaching_agent.llm_service = mock_llm
        self.coordinator.assessment_agent.llm_service = mock_llm
        self.coordinator.tutor_agent.llm_service = mock_llm
        self.coordinator.recommendation_agent.llm_service = mock_llm
        
        self.strategies_updated = 0
        
    def run_simulation(self, iterations=10):
        print(f"Starting simulation with {iterations} cycles...")
        
        with app.app_context():
            # Ensure we have topics
            topics = Topic.query.all()
            if not topics:
                print("No topics found. Run init_db first.")
                return

            for i in range(iterations):
                student_type = random.choice(["fast_visual", "slow_detailed", "average"])
                student = SimulatedStudent(student_type)
                topic = random.choice(topics)
                
                print(f"\nCycle {i+1}: Student {student.profile_type} tackling '{topic.name}'")
                
                # 1. Generate Lesson
                result = self.coordinator.perceive({
                    'task': 'generate_lesson',
                    'user_id': student.id, # Mock ID
                    'context': {
                        'topic_id': topic.id,
                        'knowledge_level': student.knowledge.get(topic.id, 0),
                        'learning_style': student.preferred_style
                    }
                }).decide().act()
                
                if not result['success']:
                    print("  - Lesson generation failed")
                    continue
                    
                teaching_data = result['results'].get('TeachingAgent', {})
                strategy_name = teaching_data.get('metadata', {}).get('teaching_style', 'unknown')
                
                # 2. Simulate Learning & Quiz
                # We assume the lesson content has some style metadata
                # For simulation, we assume the agent picked a strategy name
                
                # 3. Assessment
                passed = student.attempt_quiz(
                    teaching_data.get('metadata', {}).get('complexity', 'beginner'),
                    {'style': strategy_name} # Simplified matching
                )
                
                outcome = "PASSED" if passed else "FAILED"
                print(f"  - Strategy: {strategy_name} -> {outcome}")
                
                # 4. Reinforcement (Update AgentKnowledge)
                self.update_agent_knowledge(strategy_name, passed)
                
    def update_agent_knowledge(self, strategy_name, success):
        """Reinforce the strategy in the database"""
        # Find the strategy
        # Note: In a real system, we'd query by ID. 
        # Here we search JSON content for the name.
        
        # We need to find the AgentKnowledge row where content->>'name' == strategy_name
        # SQLite JSON support is limited, so we iterate (slow but fine for simulation script)
        knowledge_items = AgentKnowledge.query.filter_by(agent_type='teaching').all()
        
        target_item = None
        for item in knowledge_items:
            if item.content.get('name') == strategy_name:
                target_item = item
                break
        
        if target_item:
            # Update Score (Moving Average)
            alpha = 0.1
            reward = 1.0 if success else 0.0
            new_score = (1 - alpha) * target_item.effectiveness_score + (alpha * reward)
            
            target_item.effectiveness_score = new_score
            target_item.usage_count += 1
            db.session.commit()
            self.strategies_updated += 1
            print(f"    [Updated] Score: {target_item.effectiveness_score:.2f} (Uses: {target_item.usage_count})")
        else:
            print(f"    [Warning] Strategy '{strategy_name}' not found in DB")

if __name__ == '__main__':
    sim = TrainingSimulator()
    # Run a batch for verification
    sim.run_simulation(iterations=20)