Setup Python Repository

This documentation was generated with the assistance of AI. Please report any inaccuracies.

End-to-end bootstrap for a Python web repository of one of two flavors — a FastAPI service or a Django project — by orchestrating the flavor’s scaffold skill and the shared repository-level skills in one pass, collecting every shared parameter once.

Purpose

iru-setup-python-repository is the Python stack orchestrator, run stand-alone or from the iru-setup-repository front door. It resolves the flavor (from args, an iru-explore Project type: line of python-fastapi or python-django, or on-disk markers such as a root manage.py or fastapi/django in pyproject.toml, otherwise asked), the project identity (name, import package, description, per-flavor license, developer details), the Python version (checked live against devguide.python.org or endoflife.date, catalog default 3.13, never below 3.12), the flavor’s shape (database, plus orm for FastAPI or api for Django), and the shared pipeline parameters (open-source, visibility, ci, ci-tests, pages, distribution as a GHCR container image, sonar and its detail keys, branches, and — Django only — accessibility, the level of the browser UI tests' automated accessibility checks written to .ui-quality.yml), each with visibility-aware cost guidance. It then runs, each through the iru-isolated-skill-executor agent so only a short summary returns: the scaffold skill, iru-setup-antora, iru-setup-python-gitignore, iru-setup-python-github-workflows (skipped entirely when ci is no), iru-setup-changelog, and iru-setup-readme.

It announces the setup mode first (from scratch vs updating an existing repository, with a per-component inventory marked absent/present/partial); each sub-skill then asks Update / Keep as is / Replace for what it finds. Plain Python libraries (python-library) and other web frameworks such as Flask are out of scope and stop the run. Existing Poetry, PDM or pipenv projects stay in scope: the scaffold keeps the existing manager instead of silently migrating to uv.

Purpose group

Python Bootstrap

Plugin

Shipped by the iru-python plugin of the ai-catalog marketplace (install with /plugin install iru-python@ai-catalog; see Plugins).

Invocation

/iru-setup-python-repository
/iru-setup-python-repository <args as key: value lines>

Inputs

Argument Required Description Default

flavor

No

fastapi or django.

Detected from an iru-explore report or on-disk markers, otherwise asked.

mode

No

new or existing.

existing when a manifest or flavor marker exists (confirmed), else new.

open-source, visibility

No

Whether the project is open source; public, private or internal.

visibility detected via gh repo view, else asked; open source plus private is flagged as contradictory.

ci, ci-tests, pages

No

Generate workflows; unit or full CI tests (flavor-specific meaning); publish the Antora site to Pages.

ci asked (Yes recommended); ci-tests and pages asked only with ci: yes; pages defaults to yes when public.

distribution

No

none, internal (private GHCR image) or store (public GHCR image). There is no PyPI publishing.

store for public open source; otherwise asked with internal recommended.

sonar, sonar-organization, sonar-project-key, sonar-host-url

No

cloud, self-hosted or none, plus its detail keys.

cloud recommended only for public open source, otherwise none.

python-version

No

CPython minor the project runs on (.python-version, Docker base image, CI); requires-python floor stays >=3.12.

Live lookup; catalog default 3.13.

database

No

postgresql, sqlite or none (none is not offered for Django).

Asked; PostgreSQL recommended.

orm

No

sqlmodel or sqlalchemy; FastAPI only, and only when database is not none.

Asked; SQLModel recommended.

api

No

drf, ninja or none; Django only.

Asked; Django REST Framework recommended.

accessibility

No

enforce, report, off or none (opt out entirely) — whether accessibility and assistive-technology support is wanted and the automated level of the Django scaffold’s browser UI tests, written to .ui-quality.yml (see Building user interfaces); Django only, since the FastAPI flavor renders no UI.

enforce recommended; in an existing repository pre-selected from iru-explore’s `Accessibility: line (present enforce, partial report with a mixed-compliance warning, absent none), never silently flipping an existing .ui-quality.yml.

integration-branch, stable-branch

No

Branches used by the workflows.

develop and main.

Identity keys

No

name, package, description, license, developer-name, developer-email, organization-url.

Asked; skipped in existing mode where pyproject.toml already carries them.

on-existing

No

ask, keep, update or replace; forwarded to every file-writing sub-skill only when supplied.

ask (inside each sub-skill).

Unknown keys (for example publish) are ignored, as is orm for Django and api and accessibility for FastAPI.

Outputs

  • A scaffolded FastAPI or Django project with tests, quality configuration, Sphinx API docs sources and container files, via the scaffold skill — for Django also .ui-quality.yml and a Playwright UI tier (tests/ui/, .ui-screenshots/ captures).

  • An Antora docs site, a root .gitignore, CHANGELOG.md and README.md, via the delegated skills.

  • Unless ci: no, GitHub Actions workflows (build, release, sync, security) and Dependabot.

  • A final report: the setup mode, resolved flavor and mode, identity and license, Python version and its source, database/ORM/API shape, branches, a per-component action table (including .ui-quality.yml for Django; created / updated / replaced / kept as is / skipped (ci: no) / not written), the verification results (including the Django UI tier), what was wired or omitted and why, the required-secrets table, and explicit warnings.

Execution flow

flowchart TD A["Start /iru-setup-python-repository"] --> B["Step 0: resolve inputs from args"] B --> C["Step 1: resolve flavor and mode,\nannounce setup mode and inventory"] C --> C1{"Python library or other framework?"} C1 -- yes --> C2["Stop - out of scope"] C1 -- no --> D["Step 2: collect identity and license"] D --> E["Step 3: collect Python version\nand flavor shape"] E --> F["Step 4: collect shared pipeline parameters"] F --> G["Step 5: flavor scaffold skill"] G --> G1{"User stopped with no usable project?"} G1 -- yes --> G2["Stop"] G1 -- no --> H["Step 6: iru-setup-antora"] H --> I["Step 7: iru-setup-python-gitignore"] I --> J{"ci: no?"} J -- yes --> K["Step 8 skipped"] J -- no --> L["Step 8: iru-setup-python-github-workflows"] K --> M["Step 9: iru-setup-changelog"] L --> M M --> N["Step 10: iru-setup-readme"] N --> O["Step 11: report per-component actions,\nsecrets table and warnings"]

Dependencies

Invokes

  • Setup Python FastAPI — Step 5, scaffold for the fastapi flavor (via iru-isolated-skill-executor).

  • Setup Python Django — Step 5, scaffold for the django flavor (via iru-isolated-skill-executor).

  • Setup Antora — Step 6, Antora documentation site (via iru-isolated-skill-executor).

  • Setup Python Gitignore — Step 7, root .gitignore (via iru-isolated-skill-executor).

  • Setup Python GitHub Workflows — Step 8, CI/CD and security workflows, skipped when ci: no (via iru-isolated-skill-executor).

  • Setup Changelog — Step 9, root CHANGELOG.md (via iru-isolated-skill-executor).

  • Setup Readme — Step 10, root README.md (via iru-isolated-skill-executor).

Invoked by

Related agents

Source

SKILL.md on GitHub — the file this page was generated from.