benchmark.py

#!/usr/bin/env python3
'''
Performance Benchmark for SOMA Compression Algorithms
Measures speed and compression for all 15 algorithms
'''

import time
import json
import sys
from pathlib import Path

def load_sample_data():
    '''Load sample messages for testing'''
    return [
        {
            'role': 'user',
            'content': 'Analyze this code for bugs and suggest improvements. The code processes user input.'
        },
        {
            'role': 'assistant',
            'content': 'I found several issues: 1) No input validation 2) Missing error handling 3) No type checking. Here is the corrected version with proper validation and error handling included.'
        }
    ]

def benchmark_compressor(name, compress_func, messages, iterations=100):
    '''Benchmark a single compressor'''
    original_size = len(str(messages))

    # Warmup
    compress_func(messages.copy())

    # Benchmark
    start = time.time()
    for _ in range(iterations):
        result = compress_func(messages.copy())
    elapsed = time.time() - start

    compressed_size = len(str(result))

    return {
        'name': name,
        'original_size': original_size,
        'compressed_size': compressed_size,
        'compression_ratio': compressed_size / original_size,
        'compression_percent': (1 - compressed_size / original_size) * 100,
        'avg_time_ms': (elapsed / iterations) * 1000,
        'throughput_per_sec': iterations / elapsed
    }

def main():
    print("SOMA Compression Algorithms - Performance Benchmark")
    print("=" * 70)

    messages = load_sample_data()

    # Add current directory to path
    sys.path.insert(0, str(Path(__file__).parent))

    compressors = []

    # Try to import all compressors
    try:
        from performance_focused_compressor import compress_messages as perf
        compressors.append(('Performance', perf))
    except ImportError as e:
        print(f"Could not import performance_focused: {e}")

    try:
        from aggressive_hybrid_compressor import compress_messages as agg
        compressors.append(('Aggressive', agg))
    except ImportError:
        pass

    try:
        from code_focused_compressor import compress_messages as code
        compressors.append(('Code-Focused', code))
    except ImportError:
        pass

    if not compressors:
        print("No compressors found! Ensure .py files are in same directory.")
        return

    print(f"\nTesting {len(compressors)} compressors...")
    print(f"Original size: {len(str(messages))} bytes\n")

    results = []
    for name, func in compressors:
        print(f"Benchmarking {name}...", end=' ', flush=True)
        result = benchmark_compressor(name, func, messages)
        results.append(result)
        print(f"Done: {result['compression_percent']:.1f}% compression")

    # Sort by compression
    results.sort(key=lambda x: x['compression_percent'], reverse=True)

    print("\n" + "=" * 70)
    print("RESULTS")
    print("=" * 70)

    for r in results:
        print(f"{r['name']}: {r['compression_percent']:.1f}% compression, "
              f"{r['avg_time_ms']:.2f}ms avg")

    print("\nBenchmark complete!")

if __name__ == '__main__':
    main()