codex-python 0.1.1

A minimal Python library scaffold for codex-python

codex-python

A minimal Python library scaffold using uv with Python 3.13+.

Quickstart

• Requires Python 3.13+.
• Package import name: codex.
• Distribution name (PyPI): codex-python.

Repo

• Git: [e-mail]:gersmann/codex-python.git
• URL: https://github.com/gersmann/codex-python

Usage

Basic non-interactive execution via Codex CLI:

from codex import run_exec

out = run_exec("explain this repo")
print(out)

Options:

• Choose model: run_exec("...", model="gpt-4.1")
• Full auto: run_exec("scaffold a cli", full_auto=True)
• Run in another dir: run_exec("...", cd="/path/to/project")

Using a client with defaults:

from codex import CodexClient

client = CodexClient(model="gpt-4.1", full_auto=True)
print(client.run("explain this repo"))

Install uv

• macOS (Homebrew): brew install uv
• Or via install script:
  • Unix/macOS: curl -LsSf https://astral.sh/uv/install.sh | sh
  • Windows (PowerShell): iwr https://astral.sh/uv/install.ps1 -UseBasicParsing | iex

See: https://docs.astral.sh/uv/

Create a virtual env (optional)

uv python install 3.13
uv venv --python 3.13
. .venv/bin/activate  # or .venv\Scripts\activate on Windows

Build

uv build

Artifacts appear in dist/ (.whl and .tar.gz).

Publish to PyPI

• Manual:

export PYPI_API_TOKEN="pypi-XXXX"  # create at https://pypi.org/manage/account/token/
uv publish --token "$PYPI_API_TOKEN"

• GitHub Actions (Trusted Publishing): enable PyPI Trusted Publishing for gersmann/codex-python and push a tag like v0.1.0. No token is needed. The workflow at .github/workflows/publish.yml requests an OIDC token and runs uv publish --trusted-publishing=always on v* tags.

Dev tooling

• Lint: make lint (ruff + mypy)
• Tests: make test (pytest)
• Format: make fmt (ruff formatter)
• Pre-commit: uvx pre-commit install && uvx pre-commit run --all-files

Project Layout

.
├── codex/              # package root (import name: codex)
│   └── __init__.py     # version lives here
├── pyproject.toml      # PEP 621 metadata, hatchling build backend
├── README.md
└── .gitignore

Versioning

Version is managed via codex/__init__.py and exposed as __version__. The build uses Hatch’s version source.

Python Compatibility

• Requires Python >=3.13.