debug_tool_integration.py

#!/usr/bin/env python3
"""
GAIA Agent Tool Integration Debugging Script
Phase 2: Tool Integration Validation

This script systematically tests and debugs each of the 11 tools in the GAIA Agent
to identify and resolve the issues causing evaluation failures.

Critical Issues to Debug:
1. Image Processing Failures: "I'm sorry, I am unable to process the image at the moment"
2. File Handling Issues: Missing file references and incorrect file path handling
3. Tool Selection Logic: Inappropriate tool selection for specific question types
4. API Integration: Ensure all API keys and endpoints are working correctly
"""

import os
import sys
import logging
import traceback
import tempfile
import json
from pathlib import Path
from typing import Dict, Any, List, Optional
import requests
from PIL import Image
import io

# Add deployment-ready to path
sys.path.insert(0, str(Path(__file__).parent))

# Configure logging
logging.basicConfig(
    level=logging.INFO,
    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
)
logger = logging.getLogger(__name__)

class GAIAToolDebugger:
    """Comprehensive debugger for GAIA Agent tools."""
    
    def __init__(self):
        """Initialize the debugger."""
        logger.info("🐛 Initializing GAIA Tool Debugger...")
        
        # Load environment variables
        self._load_env_file()
        
        # Initialize test results
        self.test_results = {
            'tool_initialization': {},
            'tool_functionality': {},
            'api_integrations': {},
            'multimodal_capabilities': {},
            'error_handling': {},
            'overall_status': 'UNKNOWN'
        }
        
        # Test data
        self.test_data = self._prepare_test_data()
        
        logger.info("✅ GAIA Tool Debugger initialized")
    
    def _load_env_file(self):
        """Load environment variables from .env file."""
        env_file = Path('.env')
        if env_file.exists():
            with open(env_file, 'r') as f:
                for line in f:
                    line = line.strip()
                    if line and not line.startswith('#') and '=' in line:
                        key, value = line.split('=', 1)
                        os.environ[key.strip()] = value.strip()
            logger.info("✅ Environment variables loaded from .env")
        else:
            logger.warning("⚠️ No .env file found")
    
    def _prepare_test_data(self) -> Dict[str, Any]:
        """Prepare test data for debugging."""
        # Create a simple test image
        test_image = Image.new('RGB', (100, 100), color='red')
        test_image_path = tempfile.mktemp(suffix='.png')
        test_image.save(test_image_path)
        
        # Create test audio file (placeholder)
        test_audio_path = tempfile.mktemp(suffix='.wav')
        with open(test_audio_path, 'wb') as f:
            f.write(b'RIFF\x24\x00\x00\x00WAVEfmt \x10\x00\x00\x00\x01\x00\x01\x00\x44\xac\x00\x00\x88X\x01\x00\x02\x00\x10\x00data\x00\x00\x00\x00')
        
        # Create test document
        test_document = "This is a test document. It contains information about testing. The answer to the test question is 42."
        
        return {
            'image_path': test_image_path,
            'audio_path': test_audio_path,
            'document_text': test_document,
            'test_questions': {
                'math': "What is 25 * 17?",
                'python': "Calculate the sum of numbers from 1 to 10",
                'wikipedia': "What is the capital of France?",
                'arxiv': "Find papers about machine learning",
                'web_search': "What is the current weather?",
                'file_operation': "List files in current directory",
                'shell_command': "echo 'Hello World'",
                'image_analysis': "What do you see in this image?",
                'audio_transcription': "Transcribe this audio",
                'document_qa': "What is the answer mentioned in the document?"
            }
        }
    
    def debug_tool_initialization(self) -> Dict[str, Any]:
        """Debug tool initialization process."""
        logger.info("🔧 Debugging tool initialization...")
        
        results = {}
        
        # Test core AGNO tools
        core_tools = [
            ('calculator', 'agno.tools.calculator', 'CalculatorTools'),
            ('python', 'agno.tools.python', 'PythonTools'),
            ('wikipedia', 'agno.tools.wikipedia', 'WikipediaTools'),
            ('arxiv', 'agno.tools.arxiv', 'ArxivTools'),
            ('file', 'agno.tools.file', 'FileTools'),
            ('shell', 'agno.tools.shell', 'ShellTools'),
        ]
        
        # Test API-dependent tools
        api_tools = [
            ('firecrawl', 'agno.tools.firecrawl', 'FirecrawlTools', 'FIRECRAWL_API_KEY'),
            ('exa', 'agno.tools.exa', 'ExaTools', 'EXA_API_KEY'),
        ]
        
        # Test core tools
        for tool_name, module_path, class_name in core_tools:
            results[tool_name] = self._test_tool_initialization(
                tool_name, module_path, class_name
            )
        
        # Test API tools
        for tool_name, module_path, class_name, api_key in api_tools:
            results[tool_name] = self._test_tool_initialization(
                tool_name, module_path, class_name, api_key
            )
        
        # Test multimodal tools
        results['multimodal'] = self._test_multimodal_initialization()
        
        self.test_results['tool_initialization'] = results
        return results
    
    def _test_tool_initialization(self, tool_name: str, module_path: str, 
                                 class_name: str, required_api_key: str = None) -> Dict[str, Any]:
        """Test individual tool initialization."""
        result = {
            'status': 'UNKNOWN',
            'error': None,
            'api_key_present': None,
            'instance_created': False
        }
        
        try:
            # Check API key if required
            if required_api_key:
                api_key_value = os.getenv(required_api_key)
                result['api_key_present'] = bool(api_key_value)
                if not api_key_value:
                    result['status'] = 'MISSING_API_KEY'
                    result['error'] = f"Missing {required_api_key}"
                    return result
            
            # Try to import and instantiate
            module = __import__(module_path, fromlist=[class_name])
            tool_class = getattr(module, class_name)
            
            # Initialize with appropriate parameters
            if tool_name == 'exa':
                tool_instance = tool_class(api_key=os.getenv('EXA_API_KEY'))
            elif tool_name == 'firecrawl':
                tool_instance = tool_class(api_key=os.getenv('FIRECRAWL_API_KEY'))
            else:
                tool_instance = tool_class()
            
            result['instance_created'] = True
            result['status'] = 'SUCCESS'
            logger.info(f"✅ {tool_name} initialized successfully")
            
        except ImportError as e:
            result['status'] = 'IMPORT_ERROR'
            result['error'] = str(e)
            logger.error(f"❌ {tool_name} import failed: {e}")
            
        except Exception as e:
            result['status'] = 'INITIALIZATION_ERROR'
            result['error'] = str(e)
            logger.error(f"❌ {tool_name} initialization failed: {e}")
        
        return result
    
    def _test_multimodal_initialization(self) -> Dict[str, Any]:
        """Test multimodal tools initialization."""
        result = {
            'status': 'UNKNOWN',
            'error': None,
            'mistral_available': False,
            'transformers_available': False,
            'whisper_available': False,
            'capabilities': {}
        }
        
        try:
            # Test Mistral availability
            try:
                from mistralai.client import MistralClient
                result['mistral_available'] = True
                logger.info("✅ Mistral client available")
            except ImportError:
                try:
                    from mistralai import Mistral as MistralClient
                    result['mistral_available'] = True
                    logger.info("✅ Mistral client available (alternative import)")
                except ImportError:
                    logger.warning("⚠️ Mistral client not available")
            
            # Test transformers availability
            try:
                from transformers import pipeline
                result['transformers_available'] = True
                logger.info("✅ Transformers available")
            except ImportError:
                logger.warning("⚠️ Transformers not available")
            
            # Test Faster-Whisper availability
            try:
                import faster_whisper
                result['whisper_available'] = True
                logger.info("✅ Faster-Whisper available")
            except ImportError:
                logger.warning("⚠️ Faster-Whisper not available")
            
            # Try to initialize multimodal tools
            from agents.mistral_multimodal_agent import OpenSourceMultimodalTools
            multimodal_tools = OpenSourceMultimodalTools()
            result['capabilities'] = multimodal_tools.get_capabilities_status()
            result['status'] = 'SUCCESS'
            logger.info("✅ Multimodal tools initialized")
            
        except Exception as e:
            result['status'] = 'ERROR'
            result['error'] = str(e)
            logger.error(f"❌ Multimodal tools initialization failed: {e}")
        
        return result
    
    def debug_tool_functionality(self) -> Dict[str, Any]:
        """Debug individual tool functionality."""
        logger.info("🧪 Debugging tool functionality...")
        
        results = {}
        
        # Test each tool with appropriate test cases
        test_cases = [
            ('calculator', self._test_calculator),
            ('python', self._test_python),
            ('wikipedia', self._test_wikipedia),
            ('arxiv', self._test_arxiv),
            ('firecrawl', self._test_firecrawl),
            ('exa', self._test_exa),
            ('file', self._test_file),
            ('shell', self._test_shell),
            ('image_analysis', self._test_image_analysis),
            ('audio_transcription', self._test_audio_transcription),
            ('document_analysis', self._test_document_analysis),
        ]
        
        for tool_name, test_func in test_cases:
            try:
                results[tool_name] = test_func()
            except Exception as e:
                results[tool_name] = {
                    'status': 'ERROR',
                    'error': str(e),
                    'traceback': traceback.format_exc()
                }
                logger.error(f"❌ {tool_name} test failed: {e}")
        
        self.test_results['tool_functionality'] = results
        return results
    
    def _test_calculator(self) -> Dict[str, Any]:
        """Test calculator tool."""
        try:
            from agno.tools.calculator import CalculatorTools
            calc = CalculatorTools()
            
            # Test basic calculation using correct AGNO method
            result = calc.multiply(25, 17)
            expected = 425
            
            # Extract result from JSON response if needed
            actual_result = result
            if isinstance(result, dict) and 'result' in result:
                actual_result = result['result']
            elif isinstance(result, str) and 'result' in result:
                import json
                try:
                    parsed = json.loads(result)
                    actual_result = parsed.get('result', result)
                except:
                    actual_result = result
            
            return {
                'status': 'SUCCESS' if actual_result == expected else 'FAILED',
                'test_input': "multiply(25, 17)",
                'expected': expected,
                'actual': actual_result,
                'raw_result': result,
                'error': None
            }
        except Exception as e:
            return {
                'status': 'ERROR',
                'error': str(e),
                'traceback': traceback.format_exc()
            }
    
    def _test_python(self) -> Dict[str, Any]:
        """Test Python tool."""
        try:
            from agno.tools.python import PythonTools
            python_tool = PythonTools()
            
            # Test simple Python execution
            code = "result = sum(range(1, 11))\nprint(result)"
            result = python_tool.run_python_code(code)
            
            # Extract actual output from result
            if isinstance(result, dict):
                # If result is a dict, look for output or stdout keys
                actual_output = result.get('output', result.get('stdout', str(result)))
            elif isinstance(result, str):
                if "successfully" in result.lower() and "55" not in result:
                    # If it's just a success message, indicate that execution worked
                    actual_output = "Python execution completed successfully (output may be captured elsewhere)"
                else:
                    actual_output = result
            else:
                actual_output = str(result)
            
            return {
                'status': 'SUCCESS' if '55' in str(result) or '55' in str(actual_output) else 'FAILED',
                'test_input': code,
                'expected': "55",
                'actual': actual_output,
                'raw_result': result,
                'error': None
            }
        except Exception as e:
            return {
                'status': 'ERROR',
                'error': str(e),
                'traceback': traceback.format_exc()
            }
    
    def _test_wikipedia(self) -> Dict[str, Any]:
        """Test Wikipedia tool."""
        try:
            from agno.tools.wikipedia import WikipediaTools
            wiki = WikipediaTools()
            
            # Test Wikipedia search
            result = wiki.search_wikipedia("Paris France capital")
            
            return {
                'status': 'SUCCESS' if 'Paris' in str(result) else 'FAILED',
                'test_input': "Paris France capital",
                'actual': str(result)[:200] + "..." if len(str(result)) > 200 else str(result),
                'error': None
            }
        except Exception as e:
            return {
                'status': 'ERROR',
                'error': str(e),
                'traceback': traceback.format_exc()
            }
    
    def _test_arxiv(self) -> Dict[str, Any]:
        """Test ArXiv tool."""
        try:
            from agno.tools.arxiv import ArxivTools
            arxiv = ArxivTools()
            
            # Test ArXiv search using correct AGNO method (without max_results parameter)
            result = arxiv.search_arxiv_and_return_articles("machine learning")
            
            return {
                'status': 'SUCCESS' if result and len(str(result)) > 10 else 'FAILED',
                'test_input': "machine learning",
                'actual': str(result)[:200] + "..." if len(str(result)) > 200 else str(result),
                'error': None
            }
        except Exception as e:
            return {
                'status': 'ERROR',
                'error': str(e),
                'traceback': traceback.format_exc()
            }
    
    def _test_firecrawl(self) -> Dict[str, Any]:
        """Test Firecrawl tool."""
        api_key = os.getenv('FIRECRAWL_API_KEY')
        if not api_key:
            return {
                'status': 'SKIPPED',
                'error': 'FIRECRAWL_API_KEY not found'
            }
        
        try:
            from agno.tools.firecrawl import FirecrawlTools
            firecrawl = FirecrawlTools(api_key=api_key)
            
            # Test simple web scraping using correct AGNO method
            result = firecrawl.scrape_website("https://httpbin.org/json")
            
            return {
                'status': 'SUCCESS' if result and len(str(result)) > 10 else 'FAILED',
                'test_input': "https://httpbin.org/json",
                'actual': str(result)[:200] + "..." if len(str(result)) > 200 else str(result),
                'error': None
            }
        except Exception as e:
            return {
                'status': 'ERROR',
                'error': str(e),
                'traceback': traceback.format_exc()
            }
    
    def _test_exa(self) -> Dict[str, Any]:
        """Test Exa tool."""
        api_key = os.getenv('EXA_API_KEY')
        if not api_key:
            return {
                'status': 'SKIPPED',
                'error': 'EXA_API_KEY not found'
            }
        
        try:
            from agno.tools.exa import ExaTools
            exa = ExaTools(api_key=api_key)
            
            # Test search using correct AGNO method
            result = exa.search_exa("Python programming", num_results=1)
            
            return {
                'status': 'SUCCESS' if result and len(str(result)) > 10 else 'FAILED',
                'test_input': "Python programming",
                'actual': str(result)[:200] + "..." if len(str(result)) > 200 else str(result),
                'error': None
            }
        except Exception as e:
            return {
                'status': 'ERROR',
                'error': str(e),
                'traceback': traceback.format_exc()
            }
    
    def _test_file(self) -> Dict[str, Any]:
        """Test File tool."""
        try:
            from agno.tools.file import FileTools
            file_tool = FileTools()
            
            # Test file listing (without parameters - check if method accepts no args)
            result = file_tool.list_files()
            
            return {
                'status': 'SUCCESS' if result and len(str(result)) > 10 else 'FAILED',
                'test_input': "current directory",
                'actual': str(result)[:200] + "..." if len(str(result)) > 200 else str(result),
                'error': None
            }
        except Exception as e:
            return {
                'status': 'ERROR',
                'error': str(e),
                'traceback': traceback.format_exc()
            }
    
    def _test_shell(self) -> Dict[str, Any]:
        """Test Shell tool."""
        try:
            from agno.tools.shell import ShellTools
            shell = ShellTools()
            
            # Test simple command
            result = shell.run_shell_command("echo 'Hello World'")
            
            return {
                'status': 'SUCCESS' if 'Hello World' in str(result) else 'FAILED',
                'test_input': "echo 'Hello World'",
                'expected': "Hello World",
                'actual': result,
                'error': None
            }
        except Exception as e:
            return {
                'status': 'ERROR',
                'error': str(e),
                'traceback': traceback.format_exc()
            }
    
    def _test_image_analysis(self) -> Dict[str, Any]:
        """Test image analysis capability."""
        try:
            from agents.mistral_multimodal_agent import OpenSourceMultimodalTools
            multimodal = OpenSourceMultimodalTools()
            
            # Test with our test image
            result = multimodal.analyze_image(
                self.test_data['image_path'], 
                "What color is this image?"
            )
            
            # Check if we get a proper response (not an error message)
            is_error = any(error_word in result.lower() for error_word in [
                'unable', 'cannot', 'error', 'failed', 'sorry'
            ])
            
            return {
                'status': 'FAILED' if is_error else 'SUCCESS',
                'test_input': "Red color image analysis",
                'actual': result,
                'error': result if is_error else None
            }
        except Exception as e:
            return {
                'status': 'ERROR',
                'error': str(e),
                'traceback': traceback.format_exc()
            }
    
    def _test_audio_transcription(self) -> Dict[str, Any]:
        """Test audio transcription capability."""
        try:
            from agents.mistral_multimodal_agent import OpenSourceMultimodalTools
            multimodal = OpenSourceMultimodalTools()
            
            # Test with our test audio file
            result = multimodal.transcribe_audio(self.test_data['audio_path'])
            
            # Check if we get a proper response (not an error message)
            is_error = any(error_word in result.lower() for error_word in [
                'unable', 'cannot', 'error', 'failed', 'sorry', 'not available'
            ])
            
            return {
                'status': 'FAILED' if is_error else 'SUCCESS',
                'test_input': "Test audio file",
                'actual': result,
                'error': result if is_error else None
            }
        except Exception as e:
            return {
                'status': 'ERROR',
                'error': str(e),
                'traceback': traceback.format_exc()
            }
    
    def _test_document_analysis(self) -> Dict[str, Any]:
        """Test document analysis capability."""
        try:
            from agents.mistral_multimodal_agent import OpenSourceMultimodalTools
            multimodal = OpenSourceMultimodalTools()
            
            # Test document Q&A
            result = multimodal.analyze_document(
                self.test_data['document_text'],
                "What is the answer mentioned in the document?"
            )
            
            # Check if we get a proper response
            is_error = any(error_word in result.lower() for error_word in [
                'unable', 'cannot', 'error', 'failed', 'sorry'
            ])
            
            return {
                'status': 'FAILED' if is_error else 'SUCCESS',
                'test_input': "Document Q&A about answer 42",
                'expected': "42",
                'actual': result,
                'error': result if is_error else None
            }
        except Exception as e:
            return {
                'status': 'ERROR',
                'error': str(e),
                'traceback': traceback.format_exc()
            }
    
    def debug_api_integrations(self) -> Dict[str, Any]:
        """Debug API integrations."""
        logger.info("🔑 Debugging API integrations...")
        
        results = {}
        
        # Check API keys
        api_keys = {
            'MISTRAL_API_KEY': os.getenv('MISTRAL_API_KEY'),
            'EXA_API_KEY': os.getenv('EXA_API_KEY'),
            'FIRECRAWL_API_KEY': os.getenv('FIRECRAWL_API_KEY'),
        }
        
        for key_name, key_value in api_keys.items():
            results[key_name] = {
                'present': bool(key_value),
                'length': len(key_value) if key_value else 0,
                'valid_format': self._validate_api_key_format(key_name, key_value)
            }
        
        # Test API connectivity
        results['mistral_connectivity'] = self._test_mistral_api()
        results['exa_connectivity'] = self._test_exa_api()
        results['firecrawl_connectivity'] = self._test_firecrawl_api()
        
        self.test_results['api_integrations'] = results
        return results
    
    def _validate_api_key_format(self, key_name: str, key_value: str) -> bool:
        """Validate API key format."""
        if not key_value:
            return False
        
        # Basic format validation
        if key_name == 'MISTRAL_API_KEY':
            return len(key_value) > 20 and key_value.startswith(('sk-', 'ms-'))
        elif key_name == 'EXA_API_KEY':
            return len(key_value) > 10
        elif key_name == 'FIRECRAWL_API_KEY':
            return len(key_value) > 10
        
        return True
    
    def _test_mistral_api(self) -> Dict[str, Any]:
        """Test Mistral API connectivity."""
        api_key = os.getenv('MISTRAL_API_KEY')
        if not api_key:
            return {'status': 'SKIPPED', 'error': 'API key not found'}
        
        try:
            from mistralai.client import MistralClient
            client = MistralClient(api_key=api_key)
            
            # Simple test call
            response = client.chat(
                model="mistral-large-latest",
                messages=[{"role": "user", "content": "Hello"}],
                max_tokens=10
            )
            
            return {
                'status': 'SUCCESS',
                'response_length': len(str(response)),
                'error': None
            }
        except Exception as e:
            return {
                'status': 'ERROR',
                'error': str(e)
            }
    
    def _test_exa_api(self) -> Dict[str, Any]:
        """Test Exa API connectivity."""
        api_key = os.getenv('EXA_API_KEY')
        if not api_key:
            return {'status': 'SKIPPED', 'error': 'API key not found'}
        
        try:
            # Simple HTTP test to Exa API
            headers = {'Authorization': f'Bearer {api_key}'}
            response = requests.get(
                'https://api.exa.ai/search',
                headers=headers,
                params={'query': 'test', 'num_results': 1},
                timeout=10
            )
            
            return {
                'status': 'SUCCESS' if response.status_code == 200 else 'FAILED',
                'status_code': response.status_code,
                'error': None if response.status_code == 200 else response.text
            }
        except Exception as e:
            return {
                'status': 'ERROR',
                'error': str(e)
            }
    
    def _test_firecrawl_api(self) -> Dict[str, Any]:
        """Test Firecrawl API connectivity."""
        api_key = os.getenv('FIRECRAWL_API_KEY')
        if not api_key:
            return {'status': 'SKIPPED', 'error': 'API key not found'}
        
        try:
            # Simple HTTP test to Firecrawl API
            headers = {'Authorization': f'Bearer {api_key}'}
            response = requests.post(
                'https://api.firecrawl.dev/v0/scrape',
                headers=headers,
                json={'url': 'https://httpbin.org/json'},
                timeout=10
            )
            
            return {
                'status': 'SUCCESS' if response.status_code == 200 else 'FAILED',
                'status_code': response.status_code,
                'error': None if response.status_code == 200 else response.text
            }
        except Exception as e:
            return {
                'status': 'ERROR',
                'error': str(e)
            }
    
    def debug_agent_integration(self) -> Dict[str, Any]:
        """Debug the full agent integration."""
        logger.info("🤖 Debugging full agent integration...")
        
        try:
            from agents.fixed_enhanced_unified_agno_agent import FixedGAIAAgent
            
            # Initialize agent
            agent = FixedGAIAAgent()
            
            # Test agent status
            status = agent.get_tool_status()
            
            # Test with sample questions
            test_questions = [
                "What is 25 * 17?",
                "What is the capital of France?",
            ]
            
            results = {
                'agent_status': status,
                'test_responses': {}
            }
            
            for question in test_questions:
                try:
                    response = agent(question)
                    results['test_responses'][question] = {
                        'response': response,
                        'status': 'SUCCESS' if response != 'unknown' else 'FAILED'
                    }
                except Exception as e:
                    results['test_responses'][question] = {
                        'response': None,
                        'status': 'ERROR',
                        'error': str(e)
                    }
            
            return results
            
        except Exception as e:
            return {
                'status': 'ERROR',
                'error': str(e),
                'traceback': traceback.format_exc()
            }
    
    def run_comprehensive_debug(self) -> Dict[str, Any]:
        """Run comprehensive debugging of all tools."""
        logger.info("🚀 Starting comprehensive GAIA Agent debugging...")
        
        # Run all debug phases
        self.debug_tool_initialization()
        self.debug_tool_functionality()
        self.debug_api_integrations()
        
        # Test full agent integration
        self.test_results['agent_integration'] = self.debug_agent_integration()
        
        # Determine overall status
        self._determine_overall_status()
        
        # Generate summary
        self._generate_debug_summary()
        
        return self.test_results
    
    def _determine_overall_status(self):
        """Determine overall debugging status."""
        failed_tools = []
        error_tools = []
        
        # Check tool functionality
        for tool_name, result in self.test_results['tool_functionality'].items():
            if result.get('status') == 'FAILED':
                failed_tools.append(tool_name)
            elif result.get('status') == 'ERROR':
                error_tools.append(tool_name)
        
        # Check API integrations
        api_issues = []
        for api_name, result in self.test_results['api_integrations'].items():
            if isinstance(result, dict) and result.get('status') in ['FAILED', 'ERROR']:
                api_issues.append(api_name)
        
        if error_tools or api_issues:
            self.test_results['overall_status'] = 'CRITICAL_ISSUES'
        elif failed_tools:
            self.test_results['overall_status'] = 'SOME_ISSUES'
        else:
            self.test_results['overall_status'] = 'HEALTHY'
    
    def _generate_debug_summary(self):
        """Generate debugging summary."""
        summary = {
            'total_tools_tested': len(self.test_results['tool_functionality']),
            'successful_tools': [],
            'failed_tools': [],
            'error_tools': [],
            'api_status': {},
            'critical_issues': [],
            'recommendations': []
        }
        
        # Analyze tool results
        for tool_name, result in self.test_results['tool_functionality'].items():
            status = result.get('status', 'UNKNOWN')
            if status == 'SUCCESS':
                summary['successful_tools'].append(tool_name)
            elif status == 'FAILED':
                summary['failed_tools'].append(tool_name)
            elif status == 'ERROR':
                summary['error_tools'].append(tool_name)
        
        # Analyze API status
        for api_name, result in self.test_results['api_integrations'].items():
            if isinstance(result, dict):
                summary['api_status'][api_name] = result.get('status', 'UNKNOWN')
        
        # Identify critical issues
        if 'image_analysis' in summary['failed_tools']:
            summary['critical_issues'].append("Image processing failures - multimodal capabilities compromised")
        
        if 'audio_transcription' in summary['failed_tools']:
            summary['critical_issues'].append("Audio transcription failures - multimodal capabilities compromised")
        
        if any('ERROR' in str(result) for result in self.test_results['api_integrations'].values()):
            summary['critical_issues'].append("API integration failures - external service access compromised")
        
        # Generate recommendations
        if summary['failed_tools']:
            summary['recommendations'].append(f"Fix failed tools: {', '.join(summary['failed_tools'])}")
        
        if summary['error_tools']:
            summary['recommendations'].append(f"Debug error tools: {', '.join(summary['error_tools'])}")
        
        if 'image_analysis' in summary['failed_tools'] or 'audio_transcription' in summary['failed_tools']:
            summary['recommendations'].append("Install missing multimodal dependencies (transformers, faster-whisper)")
        
        self.test_results['debug_summary'] = summary
    
    def save_results(self, output_file: str = "debug_results.json"):
        """Save debugging results to file."""
        with open(output_file, 'w') as f:
            json.dump(self.test_results, f, indent=2, default=str)
        logger.info(f"📄 Debug results saved to {output_file}")
    
    def cleanup(self):
        """Clean up test files."""
        try:
            if os.path.exists(self.test_data['image_path']):
                os.unlink(self.test_data['image_path'])
            if os.path.exists(self.test_data['audio_path']):
                os.unlink(self.test_data['audio_path'])
        except Exception as e:
            logger.warning(f"⚠️ Cleanup failed: {e}")


def main():
    """Main debugging function."""
    debugger = GAIAToolDebugger()
    
    try:
        # Run comprehensive debugging
        results = debugger.run_comprehensive_debug()
        
        # Save results
        debugger.save_results("debug_results.json")
        
        # Print summary
        print("\n" + "="*80)
        print("🐛 GAIA AGENT TOOL DEBUGGING SUMMARY")
        print("="*80)
        
        summary = results.get('debug_summary', {})
        
        print(f"📊 Overall Status: {results['overall_status']}")
        print(f"🔧 Total Tools Tested: {summary.get('total_tools_tested', 0)}")
        print(f"✅ Successful Tools: {len(summary.get('successful_tools', []))}")
        print(f"❌ Failed Tools: {len(summary.get('failed_tools', []))}")
        print(f"🚨 Error Tools: {len(summary.get('error_tools', []))}")
        
        if summary.get('failed_tools'):
            print(f"\n❌ Failed Tools: {', '.join(summary['failed_tools'])}")
        
        if summary.get('error_tools'):
            print(f"\n🚨 Error Tools: {', '.join(summary['error_tools'])}")
        
        if summary.get('critical_issues'):
            print(f"\n🚨 Critical Issues:")
            for issue in summary['critical_issues']:
                print(f"  - {issue}")
        
        if summary.get('recommendations'):
            print(f"\n💡 Recommendations:")
            for rec in summary['recommendations']:
                print(f"  - {rec}")
        
        print("\n" + "="*80)
        
    except Exception as e:
        logger.error(f"❌ Debugging failed: {e}")
        traceback.print_exc()
    
    finally:
        debugger.cleanup()


if __name__ == "__main__":
    main()