bblocks-projects 1.0.2

ONE Data Python project utilities - templates, tools, and workflows

bblocks-projects

A CLI for creating, updating, and auditing Python projects against a shared Copier template.

bblocks-projects is the ONE Campaign's tool for keeping its Python projects consistent. It scaffolds new repos, pulls template improvements into existing ones, and adds individual pieces (pre-commit, CI, ruff config) to projects that didn't start from the template.

Why this exists

Every Python project starts with the same decisions and boilerplate: which linter, which type checker, how to wire CI, where tests live, how publishing works. bblocks-projects keeps the choices in one place (the template repo) and gives you four commands to apply them: create, update, add, and doctor.

Install

Requires Python 3.11+, git, and uv on PATH.

# One-off, no install (recommended)
uvx bblocks-projects create my-project

# Install with uv
uv tool install bblocks-projects

# Or pipx / pip
pipx install bblocks-projects
pip install bblocks-projects

If you're signed into the GitHub CLI (gh auth login), create prefills your name, email, and username from your GitHub profile.

Usage

Create a new project. Interactive by default — answer a handful of questions about name, license, and project type:

bblocks-projects create my-analysis        # new subdirectory
bblocks-projects create --here             # current directory

You'll be asked to pick package (publishable to PyPI; src layout, pytest, py.typed) or project (analysis, pipeline, internal tool — adds data/raw/, data/processed/, optional notebooks, and ruff/ty configs that tolerate pandas/numpy stub gaps). To skip the prompts:

bblocks-projects create my-package --type package --defaults

Update an existing project when the template changes upstream:

cd my-analysis
uvx bblocks-projects update
git diff       # review what changed
uv sync

Previously-answered questions are reused silently. Pass --no-skip-answered to re-prompt for any of them, or --ref v1.2.0 to update to a specific template version instead of main.

Add a single component to any Python project — including ones that didn't start from the template:

uvx bblocks-projects add --list             # see what's available
uvx bblocks-projects add pre-commit         # one component
uvx bblocks-projects add ruff ci pytest     # several at once

Available components: pre-commit, ruff, ci, pytest, and pypi-publish (packages only). add merges into your existing pyproject.toml rather than overwriting it, and shows the diff before applying.

Run the doctor to check a project against the current standards:

uvx bblocks-projects doctor                 # diagnose the current directory
uvx bblocks-projects doctor --strict        # exit non-zero on findings (CI-friendly)
uvx bblocks-projects doctor --check-updates # is the template newer upstream?

When not to use this

• You're not at ONE Data. The template encodes our specific choices (uv, ruff, ty, GitHub Actions, src-layout). If your team has different ones, fork the template repo — don't fight ours.
• You don't want a Copier-managed project. create and update rely on .copier-answers.yml to track which template version generated the project. If you'd rather not have that file, use add to take a one-shot copy of the components you want and walk away.
• Your project predates Copier and you want full template adoption. update requires .copier-answers.yml. Either start from add for the pieces you need, or follow the adoption guide.

What you get

Every generated project includes uv, ruff, ty, pre-commit, and a GitHub Actions CI workflow. Packages also get pytest with coverage, a py.typed marker, and a PyPI trusted-publishing workflow. Projects also get a data/ tree (raw/ gitignored, processed/ tracked) and an optional notebooks/ directory.

Next

• Getting Started — a 15-minute walkthrough from zero to a generated project you've made a real change to.

---

Source · Issues · MIT — see LICENSE.