src/gaia/utils/validate_gaia_environment.py

#!/usr/bin/env python3
"""
GAIA Environment Validation Script

This script checks all environment variables used by the GAIA system
and generates a comprehensive validation report.

Usage:
    python validate_gaia_environment.py [--generate-env] [--verbose]

Options:
    --generate-env  Generate a .env file from .env.example if not exists
    --verbose       Enable verbose output

Example:
    python validate_gaia_environment.py --generate-env --verbose
"""

import os
import sys
import argparse
import json
import logging
from pathlib import Path
from datetime import datetime
from typing import Dict, List, Any, Optional

# Add the project root to the path
sys.path.append('.')

# Import the config validator
from src.gaia.utils.config_validator import (
    validate_config,
    generate_env_example,
    initialize_env_from_example,
    get_documentation,
    REQUIRED_ENV_VARS,
    OPTIONAL_ENV_VARS
)

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

def parse_args():
    """Parse command line arguments."""
    parser = argparse.ArgumentParser(description="Validate GAIA environment configuration")
    parser.add_argument("--generate-env", action="store_true", help="Generate a .env file from .env.example if not exists")
    parser.add_argument("--verbose", action="store_true", help="Enable verbose output")
    return parser.parse_args()

def generate_markdown_report(validation_result: Dict[str, Any]) -> str:
    """
    Generate a comprehensive Markdown report of the validation results.
    
    Args:
        validation_result: Result from validate_config()
        
    Returns:
        Markdown report as string
    """
    config = validation_result["config"]
    errors = validation_result["errors"]
    warnings = validation_result["warnings"]
    
    # Start with report header
    report = f"# GAIA Environment Configuration Report\n\n"
    report += f"**Generated:** {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n\n"
    
    # Overall status
    if validation_result["valid"]:
        report += "## ✅ Configuration Status: VALID\n\n"
    else:
        report += "## ❌ Configuration Status: INVALID\n\n"
        report += f"Found {len(errors)} error(s) in configuration.\n\n"
    
    # Add errors section if there are any
    if errors:
        report += "## Errors\n\n"
        report += "The following errors must be fixed for the GAIA system to function properly:\n\n"
        for error in errors:
            report += f"- {error}\n"
        report += "\n"
    
    # Add warnings section if there are any
    if warnings:
        report += "## Warnings\n\n"
        report += "The following warnings should be reviewed but don't prevent the system from functioning:\n\n"
        for warning in warnings:
            report += f"- {warning}\n"
        report += "\n"
    
    # Required environment variables
    report += "## Required Environment Variables\n\n"
    report += "| Variable | Status | Description |\n"
    report += "|----------|--------|-------------|\n"
    
    for var in REQUIRED_ENV_VARS:
        var_name = var["name"]
        is_set = var_name in os.environ and os.environ[var_name] != ""
        
        var_error = None
        for error in errors:
            if var_name in error:
                var_error = error.split(": ", 1)[1] if ": " in error else error
                break
        
        status = "✅ Set" if is_set and not var_error else "❌ Missing or Invalid"
        if var_error:
            status += f" - {var_error}"
        
        # Mask sensitive values for security
        value = os.environ.get(var_name, "")
        if value and ("KEY" in var_name or "TOKEN" in var_name):
            value = value[:4] + "****" + value[-4:] if len(value) > 8 else "****"
        
        report += f"| {var_name} | {status} | {var['description']} |\n"
    
    # Optional environment variables
    report += "\n## Optional Environment Variables\n\n"
    report += "| Variable | Status | Description | Default |\n"
    report += "|----------|--------|-------------|--------|\n"
    
    for var in OPTIONAL_ENV_VARS:
        var_name = var["name"]
        is_set = var_name in os.environ and os.environ[var_name] != ""
        default = var["default"]
        
        var_error = None
        for error in errors:
            if var_name in error:
                var_error = error.split(": ", 1)[1] if ": " in error else error
                break
        
        if is_set:
            status = "✅ Custom" if not var_error else f"❌ Invalid - {var_error}"
        else:
            status = "⚠️ Using Default"
        
        # Mask sensitive values for security
        value = os.environ.get(var_name, default)
        if value and ("KEY" in var_name or "TOKEN" in var_name):
            value = value[:4] + "****" + value[-4:] if len(value) > 8 else "****"
        
        report += f"| {var_name} | {status} | {var['description']} | {default} |\n"
    
    # Configuration sections
    report += "\n## Configuration Sections\n\n"
    
    # Model settings
    report += "### Model Settings\n\n"
    report += f"- **Default Model:** {config.get('DEFAULT_MODEL', 'Not set')}\n"
    report += f"- **Fallback Model:** {config.get('FALLBACK_MODEL', 'Not set')}\n"
    report += f"- **Embedding Model:** {config.get('EMBEDDING_MODEL', 'Not set')}\n"
    report += f"- **Model Temperature:** {config.get('MODEL_TEMPERATURE', 'Not set')}\n"
    report += f"- **Model Max Tokens:** {config.get('MODEL_MAX_TOKENS', 'Not set')}\n"
    
    # Memory settings
    report += "\n### Memory Settings\n\n"
    report += f"- **Memory Enabled:** {config.get('MEMORY_ENABLED', 'Not set')}\n"
    report += f"- **Memory Table Name:** {config.get('MEMORY_TABLE_NAME', 'Not set')}\n"
    report += f"- **Memory TTL:** {config.get('MEMORY_TTL', 'Not set')} seconds\n"
    report += f"- **Memory Cache Size:** {config.get('MEMORY_CACHE_SIZE', 'Not set')}\n"
    
    # Web search settings
    report += "\n### Web Search Settings\n\n"
    report += f"- **Web Search Result Count:** {config.get('WEB_SEARCH_RESULT_COUNT', 'Not set')}\n"
    report += f"- **DuckDuckGo Timeout:** {config.get('DUCKDUCKGO_TIMEOUT', 'Not set')} seconds\n"
    report += f"- **DuckDuckGo Max Results:** {config.get('DUCKDUCKGO_MAX_RESULTS', 'Not set')}\n"
    
    # System settings
    report += "\n### System Settings\n\n"
    report += f"- **Max Iterations:** {config.get('MAX_ITERATIONS', 'Not set')}\n"
    report += f"- **Verbose:** {config.get('VERBOSE', 'Not set')}\n"
    report += f"- **Log Level:** {config.get('LOG_LEVEL', 'Not set')}\n"
    
    # Next steps
    report += "\n## Next Steps\n\n"
    
    if validation_result["valid"]:
        report += "Your GAIA environment configuration is valid. You can proceed with running the system.\n\n"
    else:
        report += "Please fix the errors listed above, then run this validation script again.\n\n"
        report += "To fix these issues:\n\n"
        report += "1. Edit your `.env` file and add/update the required variables\n"
        report += "2. Run this script again to verify your changes\n\n"
        report += "You can also regenerate a template `.env.example` file by running:\n\n"
        report += "```bash\n"
        report += "python -c \"from src.gaia.utils.config_validator import generate_env_example; generate_env_example()\"\n"
        report += "```\n"
    
    return report

def save_report(report: str, output_path: str = "environment_report.md") -> str:
    """
    Save the validation report to a file.
    
    Args:
        report: Markdown report to save
        output_path: Path to the output file
        
    Returns:
        Path to the saved report
    """
    with open(output_path, "w") as f:
        f.write(report)
    
    logger.info(f"Saved environment validation report to {output_path}")
    return output_path

def main():
    """Main function."""
    args = parse_args()
    
    if args.verbose:
        logger.setLevel(logging.DEBUG)
        logger.debug("Verbose logging enabled")
    
    logger.info("Starting GAIA environment validation")
    
    # Generate .env file if requested
    if args.generate_env:
        initialize_env_from_example()
    
    # Validate the configuration
    validation_result = validate_config()
    
    # Generate and save the report
    report = generate_markdown_report(validation_result)
    report_path = save_report(report)
    
    # Print summary to console
    if validation_result["valid"]:
        logger.info("✅ Environment configuration is valid!")
    else:
        logger.error(f"❌ Environment configuration has {len(validation_result['errors'])} errors!")
        for error in validation_result["errors"]:
            logger.error(f"  - {error}")
    
    if validation_result["warnings"]:
        logger.warning(f"⚠️ Environment configuration has {len(validation_result['warnings'])} warnings:")
        for warning in validation_result["warnings"]:
            logger.warning(f"  - {warning}")
    
    logger.info(f"See the full report at: {report_path}")

if __name__ == "__main__":
    main()