docs: docs-tests harness for Console UI drift detection#2669
Draft
marcel-rbro wants to merge 5 commits into
Draft
docs: docs-tests harness for Console UI drift detection#2669marcel-rbro wants to merge 5 commits into
marcel-rbro wants to merge 5 commits into
Conversation
Docs-as-tests for the Apify Console: extract UI claims (routes, tabs, buttons, headings) from platform docs with an LLM, store them as a reviewed baseline under assertions/, and verify them against Console staging with Playwright. Failures point back to source_file:line. - pages.json: adjustable list of docs pages to cover - scripts/extract*.sh: LLM extraction (run locally, commit the result) - tests/from-doc.spec.ts: evaluate the stored assertions (CI-friendly) - reporters/issues-reporter.ts: machine-readable drift report No secrets committed; auth.json and .env are gitignored. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
github-actions
Bot
added
t-docs
Contributor
|
✅ Preview for this PR (commit |
- portable read loop in extract-all.sh (macOS bash 3.2 has no mapfile) - detach claude stdin so it doesn't drain the page list when looped - slice the <output> block with perl so single-line tag+JSON also parses Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Narrow the manifest to four fixture-free landing pages (console index, settings, billing, store) and commit the AI-extracted baseline: 53 assertions (21 route, 15 text, 11 tab, 6 button). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
6 tasks
Replace the interactive auth.setup.ts + auth.json storageState handoff with a worker-scoped fixture that logs in fresh each run from CONSOLE_STAGING_USER_EMAIL/_PASSWORD (.env locally, GitHub Secrets in CI) and keeps the session in memory. Nothing is written to or read from disk, so no auth file has to pre-exist in the GitHub Action. Seeded user has no 2FA, so it's a plain email+password submit; drop the setup project and pnpm auth. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The worker auth fixture assumed a single combined sign-in form; Apify's sign-in is two-step (email -> Next -> password -> Log in), both steps on /sign-in. Pin the real selectors, avoid the SSO buttons, and wait on domcontentloaded (the Console SPA never reaches networkidle). Add timestamped step logging to the login fixture so a slow or stuck login is visible instead of a silent hang before any test reports. Also adds the pnpm workspace + lockfile so docs-tests installs in isolation, and documents the staging-user vars in .env.example. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Part of #2671
What
Adds
docs-tests/— a docs-as-tests harness that catches when the platform docs claim something the Console UI no longer does (a renamed tab, a moved route, a relabeled button).Model: an adjustable list of docs pages → AI-extracted assertions → a committed, reviewed baseline → evaluated against Console staging with Playwright. Failures point back to
source_file:lineso the offending prose is one click away.How it's split
The pipeline splits on determinism, which also keeps the public CI clean:
scripts/extract*.sh) — non-deterministic, run locally / on-demand; a human reviews the diff and commitsassertions/. No API key in CI.pnpm test) — deterministic given the stored assertion set; this is the part intended to run as a scheduled GitHub Action (modeled onlychee.yml), reading the committed baseline and reporting drift.Assertion kinds
route(path reachable),element_tab,element_button,element_text(documented label/heading visible on the page named inat).Notes for review
assertions/baseline is generated in a follow-up commit so you can review the extraction output as a separate diff.🤖 Generated with Claude Code