from pathlib import Path
import pathspec
import argparse
import mimetypes

def get_file_description(path: Path):
    mime_type, _ =  mimetypes.guess_type(path)
    return mime_type or "Unknown mimetype"

def output_metadata(path: Path, root: Path):
    desc = get_file_description(path)
    relative_path = path.relative_to(root)
    print("#################")
    print(f"## Filename: {relative_path}")
    print(f"## Mimetype: {desc}")
    print("#################")

def get_ignore_spec(root: Path):
    gitignore_path = root / ".gitignore"
    if gitignore_path.exists():
        with gitignore_path.open("r") as f:
            return pathspec.PathSpec.from_lines('gitwildmatch', f)
    return None

def should_include(path: Path, spec: pathspec.PathSpec | None, root: Path, include_git_dir: bool = False):
    relative_path = path.relative_to(root)
    if ".git" in path.parts and not include_git_dir:
        return False

    if spec is None:
        return True
    
    return not spec.match_file(str(relative_path))

def walk_filesystem(ignore_gitignore: bool, include_git_dir: bool = False):
    root = Path(".")

    spec = None
    if not ignore_gitignore:
        spec = get_ignore_spec(root)        

    for path in root.rglob("*"):
        if path.is_file():
            # print the filename (for now)
            if should_include(path, spec, root, include_git_dir):
                output_metadata(path, root)
                success, content = get_file_contents(path)
                if success:
                    print("## File Contents:")
                    print(content)
                    print("\n")

def get_file_contents(path: Path) -> tuple[bool, str]:

    try:
        with path.open("r", encoding="utf-8") as f:
            return (True, f.read())
    except (UnicodeDecodeError, PermissionError):
        return (False, "")

def main():

    parser = argparse.ArgumentParser(description="Dumpy: A tool for providing a text representation of a project formatted in a way that LLMs will understand.")

    parser.add_argument(
        "--no-gitignore",
        action="store_true",
        help="Ignore the .gitignore file and include all files"
    )

    parser.add_argument(
        "--include-git-dir",
        action="store_true",
        help="Include the .git directory in the output"
    )

    args = parser.parse_args()

    walk_filesystem(ignore_gitignore=args.no_gitignore)

if __name__ == "__main__":
    main()