Extract the path of git URLs (ssh, git@, https) in a file system-friendly representation.

git-url-path-extract.py

"""Extract the path of git URLs in a file system-friendly representation.

For example:

- `[email protected]:user/some/nested/repo.git` -> `user/some/nested/repo`
- `https://example.com/user/repo` -> `user/repo`

The output can be used to `git clone` into directories corresponding to the server path
automatically, e.g. `git clone $1 $(this-script $1)`, where $1 is some git URL.

This script handles nested repos (more path elements than just `owner/repo.git`), which
https://pypi.org/project/giturlparse/ doesn't.
"""

import argparse
import sys
from pathlib import PurePath, PurePosixPath
from urllib.parse import ParseResult, urlparse


def extract_path(raw_url: str) -> str:
    """Accepts a raw git URL and extracts its file-system friendly path."""

    url = urlparse(raw_url)

    found_path = None
    match url:
        case ParseResult(scheme="https" | "ssh", path=str(path)):
            # These schemes have native concepts of paths, so just reuse.
            found_path = path
        case ParseResult(scheme="", path=str(path)) if path.startswith("git@"):
            # The git-specific format,
            # https://git-scm.com/book/en/v2/Git-on-the-Server-The-Protocols
            _username, path = path.split("@")
            _host, path = path.split(":")
            found_path = path
        case _:
            pass

    if found_path is not None:
        return str(strip_suffix(force_relative(PurePath(found_path))))

    raise ValueError(f"Cannot process {raw_url=} (parsed as {url=}). Invalid format?")


def force_relative[P: PurePath](path: P) -> P:
    if path.is_absolute():
        path = path.relative_to(PurePosixPath("/").root)

    return path


def strip_suffix[P: PurePath](path: P) -> P:
    return path.with_suffix("")


def test():
    """Performs self-testing."""

    urls_to_result = {
        "git": ValueError(),  # Not a valid URL
        "git@": ValueError(),  # Missing stuff
        "[email protected]": ValueError(),  # Missing user etc.
        "[email protected]:repo": "repo",
        "[email protected]:owner/repo": "owner/repo",
        "[email protected]:owner/repo.git": "owner/repo",
        "[email protected]:owner/deeper/repo": "owner/deeper/repo",
        "[email protected]:owner/deeper/repo.git": "owner/deeper/repo",
        "ssh://[email protected]/repo": "repo",
        "ssh://[email protected]/owner/repo": "owner/repo",
        "ssh://[email protected]/owner/repo.git": "owner/repo",
        "ssh://[email protected]/owner/deeper/repo": "owner/deeper/repo",
        "ssh://[email protected]/owner/deeper/repo.git": "owner/deeper/repo",
        "https://example.com/repo": "repo",
        "https://example.com/owner/repo": "owner/repo",
        "https://example.com/owner/repo.git": "owner/repo",
        "https://example.com/owner/deeper/repo": "owner/deeper/repo",
        "https://example.com/owner/deeper/repo.git": "owner/deeper/repo",
    }

    for url, expected in urls_to_result.items():
        try:
            res = extract_path(url)
        except Exception as e:
            if type(e) is not type(expected):
                raise
        else:
            assert res == expected, f"{res=} and {expected=} deviate"

    # Print our own file name so if this output occurs in some intermingled stderr log,
    # its origin is clear.
    print(f"{__file__}: Self-tests OK", file=sys.stderr)


def main():
    parser = argparse.ArgumentParser(
        description=__doc__,
    )
    parser.add_argument("git_url")
    parser.add_argument(
        "--no-test",
        help="Skip running unit tests",
        action="store_true",
    )
    args = parser.parse_args()

    if not args.no_test:
        test()

    print(extract_path(args.git_url))


if __name__ == "__main__":
    main()

pyproject.toml

[project]
name = "git-url-path-extract.py"
version = "0.1.0"
requires-python = ">=3.12"

[build-system]
requires = ["setuptools >= 61.0"]
build-backend = "setuptools.build_meta"