CI Snippet Rendering Pipeline

Problem/Feature Description

A developer tools startup publishes weekly blog posts and social media content featuring code examples from their product. Their content team wants to automate snippet image generation as part of the CI pipeline so that rendered images are always consistent with the current codebase — no manual screenshots ever.

The pipeline needs to support two modes: a standard "docs" mode with conservative styling, and a one-off override mode that lets any individual pipeline run choose a different format or theme without touching the config file. The team also needs the pipeline to fail loudly if a render job silently produces an empty or corrupt file.

Set up a CI shell script (ci-render.sh) that configures the environment with appropriate defaults, performs a render of the provided code sample, applies a one-off format override for a second render, and verifies both outputs are valid. Leave a human-readable log (ci-render-log.txt) recording every command and its outcome.

Output Specification

Produce the following files:

• ci-render.sh — the CI shell script
• ci-render-log.txt — a log of commands run, environment variables set, and output files produced with confirmation they exist and are non-zero in size

Input Files

The following file is the code sample to render. Extract it before beginning.

=============== FILE: sample/highlight.py =============== def fibonacci(n: int) -> list[int]: """Return the first n Fibonacci numbers.""" seq = [0, 1] while len(seq) < n: seq.append(seq[-1] + seq[-2]) return seq[:n]

if name == "main": print(fibonacci(10)) =============== END FILE ===============