feat: package as gh2gl on PyPI with gh2gl init workflow

[paladini]

Summary

• Adds pyproject.toml (hatchling build backend) replacing requirements.txt
• Renames src/ → src/gh2gl/ as a proper installable Python package named gh2gl
• Adds src/gh2gl/cli.py with backup (default) and init subcommands
• gh2gl init creates a github-backup/ workspace folder in the current directory with config.yaml and .env ready to edit
• Updates backup.py to a thin dev shim (python backup.py still works for development)
• Updates README.md with pip install / init / run flow
• Adds github-backup/ to .gitignore

New user flow

pip install gh2gl

gh2gl init        # creates ./github-backup/ with config.yaml and .env
cd github-backup
# edit config.yaml — set github.username and gitlab.username
# edit .env — set GITHUB_TOKEN and GITLAB_TOKEN

gh2gl --dry-run   # preview
gh2gl             # run backup

Backwards compatibility

All existing CLI flags work without a subcommand (gh2gl --dry-run, gh2gl --filter "...", etc.). Running from a directory with config.yaml and .env in place works the same as before.

Tested

• pip install -e . installs the gh2gl entry point
• gh2gl init creates github-backup/config.yaml and github-backup/.env
• gh2gl --dry-run from inside the workspace listed 189 repos with correct visibility — exit code 0

Test plan

• pip install gh2gl (once published to PyPI) installs the gh2gl command
• pip install -e . works from the repo root (editable install)
• gh2gl --help shows backup args and {backup,init} subcommands
• gh2gl init creates github-backup/config.yaml and github-backup/.env
• Running gh2gl init again says folder already exists (no overwrite)
• gh2gl --dry-run from inside github-backup/ reads the local config.yaml
• python backup.py --dry-run still works (dev shim)

🤖 Generated with Claude Code