autusm/test_metadata_detection.py

#!/usr/bin/env python3
"""
Test script to reproduce the metadata detection issue with GCC.
"""

import os
import sys
import tempfile
import shutil
import urllib.request
import tarfile
from pathlib import Path

# Add src to path
sys.path.insert(0, 'src')

from autusm.metadata import MetadataExtractor
from autusm.models import PackageInfo

def test_gcc_metadata_extraction():
    """Test metadata extraction with GCC source code."""
    
    # GCC URL
    gcc_url = "https://mirrors.middlendian.com/gnu//gcc/gcc-15.1.0/gcc-15.1.0.tar.xz"
    
    # Create a temporary directory for extraction
    with tempfile.TemporaryDirectory() as temp_dir:
        temp_path = Path(temp_dir)
        archive_path = temp_path / "gcc.tar.xz"
        extract_path = temp_path / "extracted"
        
        print(f"Downloading GCC from {gcc_url}...")
        try:
            # Download the archive
            urllib.request.urlretrieve(gcc_url, archive_path)
            print("Download completed.")
            
            # Extract the archive
            print("Extracting archive...")
            extract_path.mkdir(exist_ok=True)
            with tarfile.open(archive_path, "r:xz") as tar:
                tar.extractall(extract_path)
            print("Extraction completed.")
            
            # Find the extracted directory
            extracted_dirs = [d for d in extract_path.iterdir() if d.is_dir()]
            if not extracted_dirs:
                print("Error: No extracted directory found")
                return False
                
            source_dir = extracted_dirs[0]
            print(f"Source directory: {source_dir}")
            
            # Test metadata extraction
            print("\nTesting metadata extraction...")
            extractor = MetadataExtractor()
            package_info = extractor.extract(source_dir, gcc_url)
            
            # Print the results
            print(f"\nExtracted package name: {package_info.name}")
            print(f"Version: {package_info.version}")
            print(f"Summary: {package_info.summary}")
            print(f"URL: {package_info.url}")
            print(f"Metadata files found: {package_info.metadata_files}")
            
            # Check if it correctly identified GCC
            if package_info.name and "gcc" in package_info.name.lower():
                print("\n✅ SUCCESS: Correctly identified GCC as the main package")
                return True
            else:
                print(f"\n❌ FAILURE: Identified '{package_info.name}' instead of GCC")
                return False
                
        except Exception as e:
            print(f"Error during test: {e}")
            import traceback
            traceback.print_exc()
            return False

if __name__ == "__main__":
    success = test_gcc_metadata_extraction()
    sys.exit(0 if success else 1)