Account

Test matrices & artifact registries

P36.pipelines-as-code.02 · Audience: guest, it-ml, language-pro · Prerequisites: Pipelines as code & quality gates

Real LLM grading for this pageLLM grading (this page):

One pipeline usually isn't one job — it's many. A library that claims to support Python 3.11–3.13 on Linux and macOS has to prove it on every combination, and a fast pipeline caches what it can and ships only what earned it. This module builds the three primitives behind that: matrix expansion, caching, and artifact promotion.

Step 1 / 3 — A matrix fans out
ⓘ Concept: One job, every combination
A matrix takes lists of parameters — Python versions, operating systems — and expands them into one job per combination. Support three versions on two OSes and you run six jobs from one declaration. You'll build expand_matrix, including exclude for the combination that never works.

Why it matters — A matrix is how you catch 'works on 3.11, breaks on 3.13' before a user does — the same job, run once per environment.

Ask the mentor about this module

Ask a question about this content. The mentor explains and grounds its answer in what you are studying; asking is recorded as a learning signal, not a grade.

Ctrl/Cmd + Enter to send

🎓 Practice ladder

1 graded rung · ~14 min

Build the matrix-expansion, caching and promotion helpers a real pipeline runs on.

Rung 1 — test matrices, caching & promotion

Loading exercise…

Your pipeline gates, fans out, and promotes. The last track packages the result into the artifact you actually deploy — a container image.

Where next?

Continue

Docker for Python & release automation →

Containers & Delivery · P36

Go up a level