Skip to content

Idea (parked): run the dlcs-docs-client samples against a deployment on a schedule #11

Description

@tomcrane

Parked by the hygiene sprint's session 0 (2026-08-06) — recorded for a later conversation, deliberately not commissioned now.

The idea. The Python samples in dlcs-docs-client/ exercise most of the documented API surface. Running them on a schedule (e.g. a GitHub Actions cron) against a real deployment, and treating failures as documentation bugs, would catch docs/API drift continuously instead of via periodic hand audits.

Roughly how it would work:

  • a runner script invokes each runnable sample module (python -m ... from the client root, per the existing README) against a target environment, in a dependency-safe order, with cleanup;
  • host + key/secret come from repo secrets (IIIF_CS_API_HOST etc. via settings.py); a dedicated docs customer/space keeps it isolated;
  • samples that intentionally demonstrate errors (e.g. the p02_registering 405 demos) are marked and asserted as such;
  • a failure opens/updates an issue rather than failing silently.

Why parked: the Playwright tests are the platform's proper regression suite; these samples are educational. The sample-parity rule now applies to main only, and docs branches may carry samples targeting unreleased or not-yet-existing API surface — a scheduled runner would need to respect that (run from main only). Needs a target environment, credentials, and an owner before it's worth building.

Register trail: scratch/hygiene-sprint/session-0-cross-cutting.md, card XC-10.

🤖 Generated with Claude Code

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions