#!/usr/bin/env python3
"""
Simple LLM Benchmark
"""

import time
import json
from datetime import datetime

def benchmark_enhanced_tokenizer():
    """Test our enhanced tokenizer."""
    print("Testing Enhanced Tokenizer...")
    
    try:
        from enhanced_tokenizer_minimal import MinimalEnhancedTokenizer
        import asyncio
        
        tokenizer = MinimalEnhancedTokenizer()
        
        test_texts = [
            "Hello world! This is a test.",
            "The equation x^2 + y^2 = z^2 is fundamental.",
            "def fibonacci(n): return n if n <= 1 else fibonacci(n-1) + fibonacci(n-2)",
        ]
        
        results = []
        total_time = 0
        total_tokens = 0
        
        for text in test_texts:
            start_time = time.time()
            
            loop = asyncio.new_event_loop()
            asyncio.set_event_loop(loop)
            result = loop.run_until_complete(tokenizer.tokenize(text))
            loop.close()
            
            end_time = time.time()
            processing_time = end_time - start_time
            
            results.append({
                "text": text[:50] + "...",
                "tokens": result.token_count,
                "time": processing_time,
                "type": result.semantic_features.get("content_type", "unknown"),
                "entities": len(result.entities),
                "math": len(result.math_expressions)
            })
            
            total_time += processing_time
            total_tokens += result.token_count
        
        return {
            "name": "Enhanced Advanced Tokenizer",
            "total_tokens": total_tokens,
            "total_time": total_time,
            "tokens_per_second": total_tokens / total_time if total_time > 0 else 0,
            "features": {
                "semantic_embeddings": True,
                "mathematical_processing": True,
                "named_entity_recognition": True,
                "fractal_analysis": True,
                "content_type_detection": True
            },
            "tests": results
        }
        
    except Exception as e:
        return {"name": "Enhanced Tokenizer", "error": str(e)}

def benchmark_basic_tokenizer():
    """Test basic tokenization."""
    print("Testing Basic Tokenizer...")
    
    test_texts = [
        "Hello world! This is a test.",
        "The equation x^2 + y^2 = z^2 is fundamental.",
        "def fibonacci(n): return n if n <= 1 else fibonacci(n-1) + fibonacci(n-2)",
    ]
    
    results = []
    total_time = 0
    total_tokens = 0
    
    for text in test_texts:
        start_time = time.time()
        
        tokens = text.split()
        token_count = len(tokens)
        
        end_time = time.time()
        processing_time = end_time - start_time
        
        results.append({
            "text": text[:50] + "...",
            "tokens": token_count,
            "time": processing_time,
            "type": "basic",
            "entities": 0,
            "math": 0
        })
        
        total_time += processing_time
        total_tokens += token_count
    
    return {
        "name": "Basic Python Tokenizer",
        "total_tokens": total_tokens,
        "total_time": total_time,
        "tokens_per_second": total_tokens / total_time if total_time > 0 else 0,
        "features": {
            "semantic_embeddings": False,
            "mathematical_processing": False,
            "named_entity_recognition": False,
            "fractal_analysis": False,
            "content_type_detection": False
        },
        "tests": results
    }

def main():
    print("🚀 LLM Benchmark Comparison")
    print("=" * 40)
    
    # Run benchmarks
    enhanced_results = benchmark_enhanced_tokenizer()
    basic_results = benchmark_basic_tokenizer()
    
    # Create comparison
    comparison = {
        "timestamp": datetime.now().isoformat(),
        "benchmarks": [enhanced_results, basic_results],
        "summary": {
            "enhanced_speed": enhanced_results.get("tokens_per_second", 0),
            "basic_speed": basic_results.get("tokens_per_second", 0),
            "enhanced_features": sum(enhanced_results.get("features", {}).values()),
            "basic_features": sum(basic_results.get("features", {}).values())
        }
    }
    
    # Print results
    print("\n📊 Results:")
    print(f"Enhanced Tokenizer: {enhanced_results.get('tokens_per_second', 0):.1f} tokens/sec")
    print(f"Basic Tokenizer: {basic_results.get('tokens_per_second', 0):.1f} tokens/sec")
    print(f"Enhanced Features: {comparison['summary']['enhanced_features']}/5")
    print(f"Basic Features: {comparison['summary']['basic_features']}/5")
    
    # Save results
    with open("benchmark_results.json", "w") as f:
        json.dump(comparison, f, indent=2)
    
    print("\n✅ Benchmark complete! Results saved to benchmark_results.json")
    
    # Recommendations
    print("\n💡 Recommendations:")
    if enhanced_results.get("tokens_per_second", 0) > basic_results.get("tokens_per_second", 0):
        print("• Enhanced tokenizer is faster than basic")
    else:
        print("• Basic tokenizer is faster (expected due to complexity)")
    
    if comparison['summary']['enhanced_features'] > comparison['summary']['basic_features']:
        print("• Enhanced tokenizer offers significantly more features")
    
    print("• Enhanced tokenizer is best for production AI applications")
    print("• Use enhanced tokenizer for multi-modal content processing")

if __name__ == "__main__":
    main()