Find Orphan Files Not Referenced Anywhere in Python

import os
from pathlib import Path
import re

def find_orphan_files(root_dir: str, extensions: set = None, ignore_patterns: list = None):
    """Find files not referenced by any other file in the project."""
    if extensions is None:
        extensions = {'.txt', '.md', '.py', '.html', '.css', '.js', '.json', '.yaml', '.yml'}
    if ignore_patterns is None:
        ignore_patterns = ['.git', '__pycache__', '.DS_Store']

    root = Path(root_dir)
    all_files = []
    references = set()

    for filepath in root.rglob('*'):
        if filepath.is_file() and filepath.suffix in extensions:
            if not any(part.startswith(pattern.rstrip('*')) for pattern in ignore_patterns for part in filepath.parts):
                all_files.append(filepath)
                try:
                    content = filepath.read_text(encoding='utf-8', errors='ignore')
                    # Find references in content (simple pattern: filename without path)
                    for ref_file in all_files[:-1]:  # Check against previously found files
                        if ref_file.name in content:
                            references.add(ref_file)
                except Exception:
                    pass

    orphan_files = [f for f in all_files if f not in references]
    return orphan_files

if __name__ == "__main__":
    example_dir = "."  # Current directory
    orphans = find_orphan_files(example_dir)
    print(f"Found {len(orphans)} orphan file(s):")
    for orphan in orphans:
        print(f"  {orphan}")