utility script for merging repos into a monorepo while preserving history

merge-repos.py

#!/usr/bin/env python
# merge-repos.py
# utility script for merging repos into a monorepo while preserving history


import argparse
from contextlib import contextmanager
import os
import shutil
import subprocess
import sys
import tempfile


def main():
    """ main function """
    parser = argparse.ArgumentParser(
        description="merge repos into a monorepo preseving history")

    parser.add_argument(
        "repository",
        type=str,
        help="repo address to merge e.g. git@bitbucket.org:org/dependency-one.git")
    parser.add_argument(
        "directory",
        type=str,
        help="directory in which to merge the repository e.g. "
             "dependency-one")
    parser.add_argument(
        "--branch",
        type=str,
        help="branch to merge",
        default="develop")
    parser.add_argument(
        "--dry-run", 
        dest="dryRun", 
        help="skip the git steps in this repo",
        action="store_true")

    args = parser.parse_args()

    print(args.repository)
    print(args.directory)

    with tempdir() as path:
        subprocess.call(["git", "clone", args.repository, path])
        subprocess.call(["git", "checkout", args.branch], cwd=path)

        newdir = os.path.join(path, args.directory)
        files = [(os.path.join(path, x), os.path.join(newdir, x))
                 for x in os.listdir(path) if x != '.git']
        os.makedirs(newdir)
        for src, dest in files:
            print("moving {} to {}".format(src, dest))
            os.rename(src, dest)
        subprocess.call(["git", "add", "-A"], cwd=path)
        subprocess.call(
            [
                "git",
                "commit",
                "-m",
                ":truck: Merging repositories - moving files to {}".format(
                    args.directory)],
            cwd=path)

        if not args.dryRun:
            remote = "merge-{}".format(args.directory)
            subprocess.call(["git", "remote", "add", remote, path])
            subprocess.call(["git",
                            "pull",
                            remote,
                            args.branch,
                            "--no-edit",
                            "--allow-unrelated-histories"])
            subprocess.call(["git", "remote", "rm", remote])
        else:
            print("dry-run: skipping git commands")

@contextmanager
def tempdir():
    """ creates a temporary directory that is automatically collected when finished with the
    context """
    path = tempfile.mkdtemp()
    try:
        yield path
    finally:
        try:
            shutil.rmtree(path)
        except IOError:
            print(
                'Failed to clean up temp dir {}'.format(path),
                file=sys.stderr)


main()