deploy-pkg 0.0.10

One-shot bootstrap & deploy script for Python packages (git+GitHub+PyPI).

deploy_pkg

Zero-click Python package deployment.
Build → GitHub → PyPI in a single command, without prompts.

---

✨ What is it?

deploy_pkg is a one-file deployment tool that:

1. Bootstraps a project (creates pyproject.toml, README.md, LICENSE, etc.).
2. Cleans build artifacts.
3. Bumps the patch version (X.Y.Z → X.Y.(Z+1)).
4. Builds an optional frontend (npm install && npm run build if a package.json is found).
5. Builds your Python package (python -m build) and installs it locally in editable mode.
6. Initializes a Git repository, or re-uses the existing one.
7. Creates a GitHub repo via the API (if none exists), adds it as origin, commits, tags, and pushes.
8. Uploads the freshly built distribution to PyPI with twine.

All of that without asking a single question.

---

🚀 Quick Start

pip install deploy_pkg        # install the script
deploy-me                    # run it in any project directory

• First run in an empty folder?
  → deploy_pkg scaffolds a minimal package and publishes it immediately.

• Run it again later?
  → deploy_pkg just bumps the version, rebuilds, commits, tags and republishes.

---

📦 Installation

From PyPI

pip install deploy_pkg

From source

git clone https://github.com/<you>/deploy_pkg.git
cd deploy_pkg
pip install -e .

---

🔧 Prerequisites

Tool / Env var	Purpose
git	version control, pushing to GitHub
Python ≥ 3.8	runtime
GITHUB_TOKEN	Personal Access Token with repo + user:email scopes
(optional) npm	builds /frontend if present

Create a .env file in your project root:

GITHUB_TOKEN=ghp_XXXXXXXXXXXXXXXXXXXXXXXXXXXX
# PKG_NAME=my_custom_name   # (optional) overrides the default package name

---

🖥️ Usage

deploy-me          # or:  python -m deploy_pkg

Works in any directory:

• Blank directory → bootstraps + publishes a first 0.0.1 release.
• Existing package → bumps patch version, rebuilds, commits & tags.
• Dirty git status → aborts with a clear error (keeps you safe).

---

How it works – Step by Step

#	Action	Details
1	Requirements	Checks for git, installs missing Python libs (build, twine, toml, requests, python-dotenv).
2	Scaffolding	Creates pyproject.toml, README.md, LICENSE, MANIFEST.in, .gitignore if absent.
3	Clean	Removes build/, dist/, *.egg-info, __pycache__.
4	Version bump	Reads project.version from pyproject, increments patch.
5	Frontend	If a package.json exists anywhere: bumps its version, runs npm install & npm run build.
6	Build & install	python -m build then pip install -U -e .
7	Git / GitHub	Init repo, create GitHub repo (via REST API), add remote, commit "patch update #<ver>", tag v<ver>, push.
8	Publish	twine upload dist/* – your new package is live on PyPI.

Any failure aborts the pipeline with a clear, human-friendly error.

---

CLI Flags

No flags yet. deploy_pkg is intentionally minimal – but a --dry-run or --verbose flag is planned (see Roadmap).

---

Examples

First-time release

mkdir awesome_pkg
cd awesome_pkg
deploy-me
# → 0.0.1 built, repo created on GitHub, uploaded to PyPI, installed locally

Routine patch release

cd awesome_pkg
git status   # should be clean
deploy-me
# → 0.0.2 built, commit, tag v0.0.2, push, PyPI upload

Using a custom package name

echo "PKG_NAME=super_lib" >> .env
deploy-me

---

🛡️ Security

• Access tokens are never printed.
• Git working tree must be clean or the run aborts (prevents accidental commits).
• Network operations (requests) use a 10-second timeout and explicit error handling.

---

🛠 Contributing

Pull Requests welcome!
Clone the repo, install dev deps, run tests (to be added):

pip install -e ".[dev]"
pytest

Feel free to open issues for feature requests or bug reports.

---

🗺️ Roadmap

• --dry-run flag (show all steps, no side-effects)
• Colorised logging (rich)
• Pre-/post-deploy hooks
• Config file for per-project overrides
• Verbose mode with full traceback

---

🙋 Author

Baptiste – math teacher, Python developer, and despiser of tedious release checklists.

---

📝 License

deploy_pkg is released under the MIT License – do whatever you want, just keep the copyright.

---

❤️ Acknowledgements

Inspired by the hundreds of times we forgot one of the following:

• bumping the version number
• tagging the commit
• uploading with Twine
• pushing the tag
• building the frontend first…

deploy_pkg never forgets.