synth-ai 0.2.13.dev1__py3-none-any.whl → 0.2.13.dev2__py3-none-any.whl

examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/tests/run_tests.py

@@ -0,0 +1,122 @@
+#!/usr/bin/env python3
+"""
+Test runner for Pokemon Emerald emulator tests
+
+Usage:
+    python tests/run_tests.py                    # Run all tests
+    python tests/run_tests.py test_fps_adjustment # Run specific test
+"""
+
+import os
+import sys
+import subprocess
+import importlib.util
+
+def run_test(test_name):
+    """Run a specific test"""
+    test_file = f"tests/{test_name}.py"
+    
+    if not os.path.exists(test_file):
+        print(f"❌ Test file not found: {test_file}")
+        return False
+    
+    print(f"🧪 Running test: {test_name}")
+    print("=" * 50)
+    
+    try:
+        # Run the test
+        result = subprocess.run([
+            sys.executable, test_file
+        ], capture_output=False, text=True)
+        
+        if result.returncode == 0:
+            print(f"✅ Test {test_name} passed!")
+            return True
+        else:
+            print(f"❌ Test {test_name} failed!")
+            return False
+            
+    except Exception as e:
+        print(f"❌ Error running test {test_name}: {e}")
+        return False
+
+def run_all_tests():
+    """Run all tests in the tests directory"""
+    print("🧪 Pokemon Emerald Emulator Test Suite")
+    print("=" * 50)
+    
+    # Find all test files
+    test_files = []
+    for file in os.listdir("tests"):
+        if file.startswith("test_") and file.endswith(".py"):
+            test_name = file[:-3]  # Remove .py extension
+            test_files.append(test_name)
+    
+    if not test_files:
+        print("❌ No test files found in tests/ directory")
+        return False
+    
+    print(f"Found {len(test_files)} test(s):")
+    for test in test_files:
+        print(f"  - {test}")
+    print()
+    print("💡 Note: For pytest-style tests, run:")
+    print("   python -m pytest tests/test_fps_adjustment_pytest.py -v")
+    print("   python -m pytest tests/test_server_map_validation.py -v")
+    print("💡 Note: Test state files are located in tests/states/")
+    print("💡 Note: Map reference files are saved in tests/map_references/")
+    print()
+    
+    # Run each test
+    results = []
+    for test in test_files:
+        success = run_test(test)
+        results.append((test, success))
+        print()  # Add spacing between tests
+    
+    # Print summary
+    print("📋 Test Summary")
+    print("=" * 30)
+    
+    passed_count = sum(1 for _, success in results if success)
+    total_count = len(results)
+    
+    for test, success in results:
+        status = "✅ PASS" if success else "❌ FAIL"
+        print(f"{test}: {status}")
+    
+    print(f"\nResults: {passed_count}/{total_count} tests passed")
+    
+    if passed_count == total_count:
+        print("🎉 All tests passed!")
+        return True
+    else:
+        print("⚠️  Some tests failed.")
+        return False
+
+def main():
+    """Main function"""
+    # Check if we're in the right directory
+    if not os.path.exists("server/app.py"):
+        print("❌ Error: This test runner must be run from the project root directory")
+        print("Please run: python tests/run_tests.py")
+        sys.exit(1)
+    
+    # Check if tests directory exists
+    if not os.path.exists("tests"):
+        print("❌ Error: tests/ directory not found")
+        sys.exit(1)
+    
+    # Parse command line arguments
+    if len(sys.argv) > 1:
+        # Run specific test
+        test_name = sys.argv[1]
+        success = run_test(test_name)
+        sys.exit(0 if success else 1)
+    else:
+        # Run all tests
+        success = run_all_tests()
+        sys.exit(0 if success else 1)
+
+if __name__ == "__main__":
+    main() 

examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/tests/test_agent_direct.py

@@ -0,0 +1,76 @@
+#!/usr/bin/env python3
+"""
+Pytest for agent functionality (legacy test file)
+"""
+
+import pytest
+import requests
+import json
+import time
+
+class TestAgentDirectAPI:
+    """Test class for agent API endpoints"""
+    
+    base_url = "http://localhost:8080"
+    
+    def test_imports_work(self):
+        """Test that our imports are working"""
+        import agent
+        assert hasattr(agent, 'app')
+        assert hasattr(agent, 'agent_mode')
+        assert hasattr(agent, 'websocket_connections')
+        
+    def test_global_state_initialized(self):
+        """Test that global state variables are properly initialized"""
+        import agent
+        assert agent.agent_mode == True   # Should start in agent mode by default
+        assert agent.agent_auto_enabled == False  # Should start with auto disabled
+        assert isinstance(agent.websocket_connections, set)
+        assert len(agent.websocket_connections) == 0  # Should start empty
+        
+    def test_broadcast_function_exists(self):
+        """Test that broadcast function exists and is callable"""
+        import agent
+        assert hasattr(agent, 'broadcast_state_update')
+        assert callable(agent.broadcast_state_update)
+    
+    @pytest.mark.skip(reason="Requires running server")
+    def test_status_endpoint(self):
+        """Test the /status endpoint"""
+        try:
+            response = requests.get(f"{self.base_url}/status", timeout=2)
+            assert response.status_code == 200
+            data = response.json()
+            assert "step" in data
+            assert "agent_initialized" in data
+        except requests.exceptions.ConnectionError:
+            pytest.skip("agent server not running")
+    
+    @pytest.mark.skip(reason="Requires running server") 
+    def test_toggle_mode_endpoint(self):
+        """Test the /toggle_mode endpoint"""
+        try:
+            response = requests.post(f"{self.base_url}/toggle_mode", timeout=2)
+            assert response.status_code == 200
+            data = response.json()
+            assert "mode" in data
+            assert "agent_mode" in data
+            assert data["mode"] in ["MANUAL", "AGENT"]
+        except requests.exceptions.ConnectionError:
+            pytest.skip("agent server not running")
+    
+    @pytest.mark.skip(reason="Requires running server")
+    def test_toggle_auto_endpoint(self):
+        """Test the /toggle_auto endpoint"""
+        try:
+            response = requests.post(f"{self.base_url}/toggle_auto", timeout=2)
+            assert response.status_code == 200
+            data = response.json()
+            assert "auto_enabled" in data
+            assert "status" in data
+            assert data["status"] in ["ENABLED", "DISABLED"]
+        except requests.exceptions.ConnectionError:
+            pytest.skip("agent server not running")
+
+if __name__ == "__main__":
+    pytest.main([__file__, "-v"])

examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/tests/test_agent_prompts.py

@@ -0,0 +1,413 @@
+#!/usr/bin/env python3
+"""
+Pytest for Agent Prompts Validation
+
+Tests that validate the actual prompt outputs from agent modules:
+- action.py: Action decision prompts
+- memory.py: Memory context generation
+- perception.py: Observation and scene analysis
+- planning.py: Strategic planning prompts
+
+This test validates that the agent modules generate proper prompts without "Unknown" values.
+"""
+
+import pytest
+import json
+import sys
+import os
+import requests
+import time
+import subprocess
+from typing import Dict, Any, List, Set
+from unittest.mock import Mock, patch
+from pathlib import Path
+
+# Add the project root to the path
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+# Import agent modules
+from agent.action import action_step
+from agent.memory import memory_step, extract_key_state_info
+from agent.perception import perception_step
+from agent.planning import planning_step
+from utils.vlm import VLM
+from utils.state_formatter import format_state_for_llm, format_state_summary
+
+
+class TestAgentPrompts:
+    """Test class for agent prompts validation"""
+    
+    @pytest.fixture
+    def server_url(self):
+        """Server URL for testing"""
+        return "http://localhost:8000"
+    
+    @pytest.fixture
+    def mock_vlm(self):
+        """Create a mock VLM that captures prompts and returns reasonable responses"""
+        mock_vlm = Mock(spec=VLM)
+        
+        def mock_get_query(frame, prompt, context=""):
+            # Capture the prompt for analysis
+            if not hasattr(mock_vlm, 'captured_prompts'):
+                mock_vlm.captured_prompts = []
+            mock_vlm.captured_prompts.append({
+                'context': context,
+                'prompt': prompt,
+                'frame': frame is not None
+            })
+            
+            # Return reasonable responses based on context
+            if "PERCEPTION" in context:
+                return "I can see the player character on a grassy route with trees and paths."
+            elif "ACTION" in context:
+                return "UP"
+            elif "PLANNING" in context:
+                return "Continue exploring the route and look for items or trainers."
+            elif "MEMORY" in context:
+                return "Updated memory context with current observations."
+            else:
+                return "Default response"
+        
+        def mock_get_text_query(prompt, context=""):
+            # Capture the prompt for analysis
+            if not hasattr(mock_vlm, 'captured_prompts'):
+                mock_vlm.captured_prompts = []
+            mock_vlm.captured_prompts.append({
+                'context': context,
+                'prompt': prompt,
+                'frame': False
+            })
+            
+            # Return reasonable responses based on context
+            if "ACTION" in context:
+                return "UP"
+            elif "PLANNING" in context:
+                return "Continue exploring the route and look for items or trainers."
+            elif "MEMORY" in context:
+                return "Updated memory context with current observations."
+            else:
+                return "Default response"
+        
+        mock_vlm.get_query = mock_get_query
+        mock_vlm.get_text_query = mock_get_text_query
+        
+        return mock_vlm
+    
+    def find_state_files(self):
+        """Find all .state files in the tests/states directory"""
+        states_dir = Path(__file__).parent / "states"
+        if not states_dir.exists():
+            return []
+        
+        state_files = list(states_dir.glob("*.state"))
+        return sorted(state_files)
+    
+    def start_server_with_state(self, state_file_path: str):
+        """Start the server with a specific state file"""
+        import subprocess
+        
+        # Kill any existing server processes
+        try:
+            subprocess.run(["pkill", "-f", "server.app"], check=False)
+            time.sleep(1)
+        except:
+            pass
+        
+        # Start new server
+        server_process = subprocess.Popen([
+            "conda", "run", "-n", "mgba", "python", "-m", "server.app", 
+            "--load-state", state_file_path
+        ], stdout=subprocess.PIPE, stderr=subprocess.PIPE)
+        
+        # Wait for server to start
+        time.sleep(3)
+        
+        return server_process
+    
+    def stop_server(self, server_process):
+        """Stop the server"""
+        if server_process:
+            try:
+                server_process.terminate()
+                server_process.wait(timeout=5)
+            except:
+                server_process.kill()
+    
+    def get_state_from_server(self, server_url: str) -> Dict[str, Any]:
+        """Get state data from the server"""
+        try:
+            response = requests.get(f"{server_url}/state", timeout=10)
+            if response.status_code == 200:
+                return response.json()
+            else:
+                return {}
+        except Exception:
+            return {}
+    
+    def test_action_module_prompts(self, mock_vlm, server_url):
+        """Test that action module generates proper prompts without 'Unknown' values"""
+        state_files = self.find_state_files()
+        
+        for state_file in state_files:
+            print(f"\nTesting ACTION module with {state_file}")
+            
+            # Start server
+            server_process = self.start_server_with_state(str(state_file))
+            
+            try:
+                # Get state data
+                state_data = self.get_state_from_server(server_url)
+                if not state_data:
+                    pytest.skip(f"Could not get state data for {state_file}")
+                
+                # Mock inputs for action_step
+                memory_context = "Test memory context"
+                current_plan = "Test plan"
+                latest_observation = "Test observation"
+                frame = None
+                recent_actions = ["UP", "A", "RIGHT"]
+                
+                # Call action_step
+                actions = action_step(memory_context, current_plan, latest_observation, frame, state_data, recent_actions, mock_vlm)
+                
+                # Check captured prompts
+                action_prompts = [p for p in mock_vlm.captured_prompts if "ACTION" in p['context']]
+                
+                assert action_prompts, f"No action prompts captured for {state_file}"
+                
+                # Analyze the action prompt for issues
+                action_prompt = action_prompts[0]['prompt']
+                
+                # Check for "Unknown" values in the prompt
+                assert "Unknown" not in action_prompt, f"Action prompt contains 'Unknown' values in {state_file}"
+                
+                # Check for required sections
+                required_sections = ["COMPREHENSIVE GAME STATE DATA", "ENHANCED ACTION CONTEXT", "ACTION DECISION TASK"]
+                for section in required_sections:
+                    assert section in action_prompt, f"Action prompt missing section '{section}' in {state_file}"
+                
+                # Check that actions were returned
+                assert actions, f"Action module returned no actions for {state_file}"
+                assert isinstance(actions, list), f"Action module returned non-list actions for {state_file}"
+                
+            finally:
+                self.stop_server(server_process)
+    
+    def test_memory_module_prompts(self, mock_vlm, server_url):
+        """Test that memory module generates proper prompts without 'Unknown' values"""
+        state_files = self.find_state_files()
+        
+        for state_file in state_files:
+            print(f"\nTesting MEMORY module with {state_file}")
+            
+            # Start server
+            server_process = self.start_server_with_state(str(state_file))
+            
+            try:
+                # Get state data
+                state_data = self.get_state_from_server(server_url)
+                if not state_data:
+                    pytest.skip(f"Could not get state data for {state_file}")
+                
+                # Test extract_key_state_info function
+                key_info = extract_key_state_info(state_data)
+                
+                # Check for "Unknown" values in key info
+                assert "Unknown" not in str(key_info), f"Memory key_info contains 'Unknown' values in {state_file}"
+                
+                # Check for required fields
+                required_fields = ['state_summary', 'player_name', 'money', 'current_map', 'in_battle', 'party_health']
+                for field in required_fields:
+                    assert field in key_info, f"Memory key_info missing field '{field}' in {state_file}"
+                
+                # Test memory_step function
+                memory_context = "Test memory context"
+                current_plan = "Test plan"
+                recent_actions = ["UP", "A", "RIGHT"]
+                observation_buffer = [
+                    {
+                        "frame_id": 1,
+                        "observation": "Test observation",
+                        "state": state_data
+                    }
+                ]
+                
+                # Call memory_step
+                updated_memory = memory_step(memory_context, current_plan, recent_actions, observation_buffer, mock_vlm)
+                
+                # Check for "Unknown" values in memory context
+                assert "Unknown" not in updated_memory, f"Memory context contains 'Unknown' values in {state_file}"
+                
+                # Check for required sections
+                required_sections = ["COMPREHENSIVE MEMORY CONTEXT", "CURRENT STATE", "CURRENT PLAN", "KEY EVENTS", "RECENT MEMORY"]
+                for section in required_sections:
+                    assert section in updated_memory, f"Memory context missing section '{section}' in {state_file}"
+                
+                # Check that memory context is not empty
+                assert len(updated_memory.strip()) > 100, f"Memory context seems too short in {state_file}"
+                
+            finally:
+                self.stop_server(server_process)
+    
+    def test_perception_module_prompts(self, mock_vlm, server_url):
+        """Test that perception module generates proper prompts without 'Unknown' values"""
+        state_files = self.find_state_files()
+        
+        for state_file in state_files:
+            print(f"\nTesting PERCEPTION module with {state_file}")
+            
+            # Start server
+            server_process = self.start_server_with_state(str(state_file))
+            
+            try:
+                # Get state data
+                state_data = self.get_state_from_server(server_url)
+                if not state_data:
+                    pytest.skip(f"Could not get state data for {state_file}")
+                
+                # Mock frame
+                frame = None
+                
+                # Call perception_step
+                observation, slow_thinking = perception_step(frame, state_data, mock_vlm)
+                
+                # Check captured prompts
+                perception_prompts = [p for p in mock_vlm.captured_prompts if "PERCEPTION" in p['context']]
+                
+                assert perception_prompts, f"No perception prompts captured for {state_file}"
+                
+                # Analyze the perception prompt for issues
+                perception_prompt = perception_prompts[0]['prompt']
+                
+                # Check for "Unknown" values in the prompt
+                assert "Unknown" not in perception_prompt, f"Perception prompt contains 'Unknown' values in {state_file}"
+                
+                # Check for required sections
+                required_sections = ["COMPREHENSIVE GAME STATE DATA", "VISUAL ANALYSIS TASK"]
+                for section in required_sections:
+                    assert section in perception_prompt, f"Perception prompt missing section '{section}' in {state_file}"
+                
+                # Check for analysis instructions
+                analysis_keywords = ["CUTSCENE", "MAP", "BATTLE", "DIALOGUE", "MENU"]
+                found_keywords = [kw for kw in analysis_keywords if kw in perception_prompt]
+                assert len(found_keywords) >= 3, f"Perception prompt missing analysis keywords in {state_file}. Found: {found_keywords}"
+                
+                # Check that observation was returned
+                assert observation, f"Perception module returned no observation for {state_file}"
+                assert isinstance(observation, dict), f"Perception module returned non-dict observation for {state_file}"
+                
+                # Check that slow_thinking is boolean
+                assert isinstance(slow_thinking, bool), f"Perception module returned non-boolean slow_thinking for {state_file}"
+                
+            finally:
+                self.stop_server(server_process)
+    
+    def test_planning_module_prompts(self, mock_vlm, server_url):
+        """Test that planning module generates proper prompts without 'Unknown' values"""
+        state_files = self.find_state_files()
+        
+        for state_file in state_files:
+            print(f"\nTesting PLANNING module with {state_file}")
+            
+            # Start server
+            server_process = self.start_server_with_state(str(state_file))
+            
+            try:
+                # Get state data
+                state_data = self.get_state_from_server(server_url)
+                if not state_data:
+                    pytest.skip(f"Could not get state data for {state_file}")
+                
+                # Mock inputs for planning_step
+                memory_context = "Test memory context"
+                current_plan = None  # Start with no plan
+                slow_thinking_needed = True
+                
+                # Call planning_step
+                plan = planning_step(memory_context, current_plan, slow_thinking_needed, state_data, mock_vlm)
+                
+                # Check captured prompts
+                planning_prompts = [p for p in mock_vlm.captured_prompts if "PLANNING" in p['context']]
+                
+                assert planning_prompts, f"No planning prompts captured for {state_file}"
+                
+                # Analyze the planning prompt for issues
+                planning_prompt = planning_prompts[0]['prompt']
+                
+                # Check for "Unknown" values in the prompt
+                assert "Unknown" not in planning_prompt, f"Planning prompt contains 'Unknown' values in {state_file}"
+                
+                # Check for required sections
+                required_sections = ["COMPREHENSIVE GAME STATE DATA", "STRATEGIC PLANNING TASK"]
+                for section in required_sections:
+                    assert section in planning_prompt, f"Planning prompt missing section '{section}' in {state_file}"
+                
+                # Check for planning instructions
+                planning_keywords = ["IMMEDIATE GOAL", "SHORT-TERM OBJECTIVES", "LONG-TERM STRATEGY", "EFFICIENCY NOTES"]
+                found_keywords = [kw for kw in planning_keywords if kw in planning_prompt]
+                assert len(found_keywords) >= 3, f"Planning prompt missing planning keywords in {state_file}. Found: {found_keywords}"
+                
+                # Check that plan was returned
+                assert plan, f"Planning module returned no plan for {state_file}"
+                assert isinstance(plan, str), f"Planning module returned non-string plan for {state_file}"
+                
+            finally:
+                self.stop_server(server_process)
+    
+    def test_all_modules_integration(self, mock_vlm, server_url):
+        """Test that all modules work together without 'Unknown' values"""
+        state_files = self.find_state_files()
+        
+        for state_file in state_files:
+            print(f"\nTesting ALL MODULES integration with {state_file}")
+            
+            # Start server
+            server_process = self.start_server_with_state(str(state_file))
+            
+            try:
+                # Get state data
+                state_data = self.get_state_from_server(server_url)
+                if not state_data:
+                    pytest.skip(f"Could not get state data for {state_file}")
+                
+                # Test all modules in sequence
+                memory_context = "Test memory context"
+                current_plan = None
+                latest_observation = "Test observation"
+                frame = None
+                recent_actions = ["UP", "A", "RIGHT"]
+                observation_buffer = [
+                    {
+                        "frame_id": 1,
+                        "observation": "Test observation",
+                        "state": state_data
+                    }
+                ]
+                
+                # 1. Perception
+                observation, slow_thinking = perception_step(frame, state_data, mock_vlm)
+                assert observation and isinstance(observation, dict)
+                
+                # 2. Memory
+                updated_memory = memory_step(memory_context, current_plan, recent_actions, observation_buffer, mock_vlm)
+                assert "Unknown" not in updated_memory
+                
+                # 3. Planning
+                plan = planning_step(updated_memory, current_plan, slow_thinking, state_data, mock_vlm)
+                assert plan and isinstance(plan, str)
+                
+                # 4. Action
+                actions = action_step(updated_memory, plan, observation, frame, state_data, recent_actions, mock_vlm)
+                assert actions and isinstance(actions, list)
+                
+                # Check that no prompts contain "Unknown"
+                for prompt_data in mock_vlm.captured_prompts:
+                    assert "Unknown" not in prompt_data['prompt'], f"Found 'Unknown' in {prompt_data['context']} prompt for {state_file}"
+                
+            finally:
+                self.stop_server(server_process)
+
+
+if __name__ == "__main__":
+    pytest.main([__file__])