scripts/check_license_headers.py

#!/usr/bin/env python3
# SPDX-FileCopyrightText: 2026 Team Centurions
# SPDX-License-Identifier: AGPL-3.0-or-later

import sys
from pathlib import Path

ROOT_DIR = Path(__file__).resolve().parent.parent

EXCLUDED_DIRS = {
    ".git",
    ".github",
    ".gitlab",
    ".venv",
    ".pytest_cache",
    ".mypy_cache",
    "node_modules",
    "dist",
    "build",
    "tmp",
    "models",
    ".specify",
}

EXCLUDED_FILES = {
    "pageparse.db",
    "package-lock.json",
    ".secrets.baseline",
}

REQUIRED_KEYWORDS = [
    "SPDX-FileCopyrightText",
    "SPDX-License-Identifier"
]

def check_file(file_path: Path) -> bool:
    # Skip binary or very large files
    if file_path.suffix in [".png", ".jpg", ".jpeg", ".db", ".onnx", ".gguf", ".zip", ".tar", ".gz"]:
        return True
    if file_path.name in EXCLUDED_FILES:
        return True
        
    try:
        content = file_path.read_text(encoding="utf-8", errors="ignore")
        # Check if the keywords exist in the file
        missing = [kw for kw in REQUIRED_KEYWORDS if kw not in content]
        if missing:
            print(f"FAIL: Missing license keywords {missing} in {file_path.relative_to(ROOT_DIR)}")
            return False
        return True
    except Exception as e:
        print(f"ERROR reading {file_path}: {e}")
        return False

def main() -> None:
    print("Checking license compliance in files...")
    all_passed = True
    count = 0
    
    # Check all files under source and configurations
    for path in ROOT_DIR.rglob("*"):
        if path.is_file():
            # Check if any parent dir is excluded
            parts = path.relative_to(ROOT_DIR).parts
            if any(p in EXCLUDED_DIRS for p in parts):
                continue
            
            # Check only code/config/markdown files
            if path.suffix in [".py", ".ts", ".tsx", ".yml", ".yaml", ".md", ".gbnf"]:
                count += 1
                if not check_file(path):
                    all_passed = False
                    
    print(f"Checked {count} files.")
    if all_passed:
        print("SUCCESS: All checked files have correct license headers.")
        sys.exit(0)
    else:
        print("FAILURE: Some files are missing copyright or license headers.")
        sys.exit(1)

if __name__ == "__main__":
    main()