test_library.py

#!/usr/bin/env python3
"""
Simple test script to verify RefineBench library functionality.
"""

import os
import sys

# Add the current directory to Python path
sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))

def test_imports():
    """Test that all modules can be imported"""
    try:
        from refinebench_eval import RefineBenchConfig, RefineBenchDataset
        print("✓ Core imports successful")
        
        from refinebench_eval.config import ModelConfig, ExperimentalConfig
        print("✓ Config imports successful")
        
        from refinebench_eval.dataset import RefineBenchDataset
        print("✓ Dataset imports successful")
        
        from refinebench_eval.utils import MessageBuilder
        print("✓ Utils imports successful")
        
        return True
    except ImportError as e:
        print(f"✗ Import error: {e}")
        return False

def test_config():
    """Test configuration creation"""
    try:
        from refinebench_eval import RefineBenchConfig
        
        config = RefineBenchConfig(
            agent=RefineBenchConfig.ModelConfig(
                model_name="test_model",
                model_path="test/path"
            ),
            evaluator=RefineBenchConfig.ModelConfig(
                model_name="test_evaluator",
                model_path="test/evaluator"
            )
        )
        
        print("✓ Configuration creation successful")
        return True
    except Exception as e:
        print(f"✗ Configuration error: {e}")
        return False

def test_dataset():
    """Test dataset loading"""
    try:
        from refinebench_eval import RefineBenchDataset
        
        dataset = RefineBenchDataset()
        print("✓ Dataset initialization successful")
        
        # Test getting data index list (this will try to load from HF)
        try:
            indices = dataset.get_data_index_list()
            print(f"✓ Dataset loading successful, found {len(indices)} instances")
        except Exception as e:
            print(f"⚠ Dataset loading failed (expected if no internet): {e}")
        
        return True
    except Exception as e:
        print(f"✗ Dataset error: {e}")
        return False

def main():
    print("Testing RefineBench Library...")
    print("=" * 40)
    
    tests = [
        ("Import Test", test_imports),
        ("Configuration Test", test_config),
        ("Dataset Test", test_dataset),
    ]
    
    passed = 0
    total = len(tests)
    
    for test_name, test_func in tests:
        print(f"\n{test_name}:")
        if test_func():
            passed += 1
        else:
            print(f"❌ {test_name} failed")
    
    print("\n" + "=" * 40)
    print(f"Tests passed: {passed}/{total}")
    
    if passed == total:
        print("🎉 All tests passed! Library is ready to use.")
        return 0
    else:
        print("⚠️  Some tests failed. Check the errors above.")
        return 1

if __name__ == "__main__":
    sys.exit(main())