pcbjam/tests/tools/screenshots/promote.ts
Viktor Vaczi 13551f4f22 feat(tests): screenshot regression + Discord review, perf-tracked
New tooling in tests/tools/screenshots/ (TypeScript via tsx):
- compare.ts: one pixelmatch engine (AA-excluded), connected-component
  "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors.
- promote.ts: churn-free updater — overwrite a baseline only when decoded
  pixels differ beyond the floor, copying CI bytes verbatim (no re-encode
  churn); pulls a CI run via `gh run download` or a local --from dir.
- post-discord.ts: always-on CI-on-main report (SHA + e2e status + the
  track-only runtime-perf table), then screenshot triptychs, batched +
  size-capped + flood-collapsed + 429-aware.
- perf-report.ts: perf table with Δ vs the previous main run (via gh).
- changelog.ts: no-build git-history baseline differ (Discord trigger B).
- noise.ts / gen-manifest.ts: calibration + manifest generation.

CI wiring:
- wasm-build.yml: post-test step runs the gate + report on the already-
  produced test-results (no extra build); report-only (continue-on-error),
  posts only on push to main, inert without DISCORD_WEBHOOK_URL.
- ci-ubicloud.yml: secrets: inherit (pass the webhook through).
- screenshot-changelog.yml: ~30s no-build changelog on baseline changes.

screenshot-manifest.json: canonical 354-name set + best-effort engine tags
(313 chromium-swiftshader / 41 firefox-llvmpipe).

Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1)
so committed baselines are uniformly css-scaled.

Design: CI's Linux render is the single source of truth; no pinned
container (accept rare env drift -> re-promote); dev commits via promote.
Replaces the byte-cmp compare-screenshots.sh + file-size-proxy
update-baseline-screenshots.sh (kept for now until the first re-baseline).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:17:20 +02:00

149 lines
6.2 KiB
TypeScript

/**
* Churn-free baseline updater — "promote CI's render".
*
* CI's x86 render is the source of truth. This pulls a CI run's screenshots
* (`gh run download`) — or a local dir via --from — and, for each one, overwrites
* the committed baseline ONLY when the decoded pixels differ beyond the per-engine
* floor. Unchanged baselines are left byte-identical (never re-encoded), so git
* sees no churn. New shots are added; baselines with no render are reported as
* removal candidates and only deleted with --prune.
*
* CLI (from tests/):
* tsx tools/screenshots/promote.ts --run <ci-run-id> [--repo owner/repo] [--prune] [--dry-run]
* tsx tools/screenshots/promote.ts --from <dir> [--prune] [--dry-run]
*/
import * as fs from 'fs';
import * as os from 'os';
import * as path from 'path';
import { execFileSync } from 'child_process';
import { BASELINE_DIRS, MANIFEST_PATH, floorFor, type Manifest } from './config';
import { diffImages, loadPng } from './image-ops';
function listPngs(dir: string): string[] {
if (!fs.existsSync(dir)) return [];
return fs.readdirSync(dir).filter((f) => f.toLowerCase().endsWith('.png'));
}
/** basename → absolute committed baseline path (first BASELINE_DIRS entry wins). */
function baselineIndex(root: string): Map<string, string> {
const index = new Map<string, string>();
for (const dir of BASELINE_DIRS) {
for (const name of listPngs(path.join(root, dir))) {
if (!index.has(name)) index.set(name, path.join(root, dir, name));
}
}
return index;
}
function loadManifest(root: string): Manifest | undefined {
const p = path.join(root, MANIFEST_PATH);
if (!fs.existsSync(p)) return undefined;
try {
return JSON.parse(fs.readFileSync(p, 'utf8')) as Manifest;
} catch {
return undefined;
}
}
/** Download a CI run's artifact and return the top-level test-results dir holding the shots. */
function downloadRun(runId: string, repo?: string): string {
const tmp = fs.mkdtempSync(path.join(os.tmpdir(), 'promote-'));
const repoArgs = repo ? ['--repo', repo] : [];
execFileSync('gh', ['run', 'download', runId, '-D', tmp, ...repoArgs], { stdio: 'inherit' });
// The artifact stores test-results/** — find that dir; its immediate *.png are the shots
// (exclude the nested screenshot-diff/ triptychs).
const stack = [tmp];
while (stack.length) {
const dir = stack.pop()!;
const entries = fs.readdirSync(dir, { withFileTypes: true });
if (path.basename(dir) === 'test-results') return dir;
for (const e of entries) if (e.isDirectory()) stack.push(path.join(dir, e.name));
}
throw new Error(`no test-results/ dir found in the downloaded artifact under ${tmp}`);
}
type Plan = { updated: string[]; added: string[]; unchanged: string[]; removedCandidates: string[] };
function buildPlan(root: string, renderDir: string, manifest?: Manifest): { plan: Plan; apply: () => void } {
const baselines = baselineIndex(root);
const rendered = new Set(listPngs(renderDir));
const plan: Plan = { updated: [], added: [], unchanged: [], removedCandidates: [] };
const actions: Array<() => void> = [];
for (const name of rendered) {
const src = path.join(renderDir, name);
const existing = baselines.get(name);
if (!existing) {
const dest = path.join(root, BASELINE_DIRS[0], name);
plan.added.push(name);
actions.push(() => fs.copyFileSync(src, dest)); // verbatim bytes
continue;
}
const d = diffImages(loadPng(existing), loadPng(src));
const floor = floorFor(name, manifest);
if (!d.dimsMatch || d.changedRatio > floor.changedRatio) {
plan.updated.push(name);
actions.push(() => fs.copyFileSync(src, existing)); // verbatim bytes, no re-encode → no churn
} else {
plan.unchanged.push(name); // leave the committed file untouched
}
}
// Removal candidates: a committed baseline the manifest expects but this render didn't produce.
for (const [name, abs] of baselines) {
if (rendered.has(name)) continue;
if (manifest && !manifest.screenshots.some((e) => e.name === name)) continue;
plan.removedCandidates.push(name);
actions.push(() => {}); // pruning is opt-in (see main)
void abs;
}
return { plan, apply: () => actions.forEach((a) => a()) };
}
function parseArgs(argv: string[]): Record<string, string | boolean> {
const out: Record<string, string | boolean> = {};
for (let i = 0; i < argv.length; i++) {
const a = argv[i];
if (a === '--run') out.run = argv[++i];
else if (a === '--from') out.from = argv[++i];
else if (a === '--repo') out.repo = argv[++i];
else if (a === '--prune') out.prune = true;
else if (a === '--dry-run') out.dryRun = true;
}
return out;
}
function main(): void {
const args = parseArgs(process.argv.slice(2));
const root = process.cwd();
if (!args.run && !args.from) {
console.error('usage: promote.ts --run <ci-run-id> [--repo owner/repo] | --from <dir> [--prune] [--dry-run]');
process.exitCode = 2;
return;
}
const renderDir = args.from ? (args.from as string) : downloadRun(args.run as string, args.repo as string);
const manifest = loadManifest(root);
const { plan, apply } = buildPlan(root, renderDir, manifest);
console.log(
`[promote] updated=${plan.updated.length} added=${plan.added.length} ` +
`unchanged=${plan.unchanged.length} removal-candidates=${plan.removedCandidates.length}`
);
for (const n of plan.updated) console.log(` UPDATE ${n}`);
for (const n of plan.added) console.log(` ADD ${n}`);
for (const n of plan.removedCandidates) console.log(` REMOVE? ${n}${args.prune ? ' (pruning)' : ' (use --prune to delete)'}`);
if (args.dryRun) {
console.log('[promote] dry-run — no files written');
return;
}
apply();
if (args.prune) {
const baselines = baselineIndex(root);
for (const n of plan.removedCandidates) fs.rmSync(baselines.get(n)!, { force: true });
}
console.log('[promote] done — review `git status` and commit the changed baselines');
}
if (require.main === module) main();