QVAC-17810 test[skiplog]: add img2img integration tests for diffusion

[Victor-Rodzko]

🎯 What problem does this PR solve?

• img2img was shipped to the SDK in QVAC-17304 feat[api]: add img2img support to SDK diffusion API #1662 but tests-qvac only had unit/mock coverage; real integration coverage against loaded diffusion models was missing.
• tests-qvac/tests/shared/executors/diffusion-executor.ts had drifted: heavy if (testId === ...) branching, unknown/any params, ad-hoc PNG-size byte checks that produce false positives on compressed images.

📝 How does it solve it?

• New e2e cases in diffusion-tests.ts exercising the img2img path against real loaded models:
  • diffusion-img2img-vs-txt2img-baseline — proves init_image actually changes output (byte-delta + IHDR-dimension comparison vs txt2img baseline).
  • diffusion-img2img-img-cfg-scale — img_cfg_scale parameter accepted/rendered.
  • diffusion-img2img-invalid-strength — Zod rejects out-of-range strength.
  • Reuses existing diffusion-basic-img2img.
• Platform split (matches vision-test pattern): asset filename → Uint8Array resolution lives in desktop/executors/diffusion-executor.ts (Node fs); shared executor stays React Native-clean and only sees bytes.
• Mobile: all diffusion tests skipped (SD 2.1 1B Q8_0 cold-load OOMs Device Farm devices, ~3GB). SkipExecutor message updated; mobile/executors/diffusion-executor.ts removed as dead code.
• Refactor of shared/executors/diffusion-executor.ts to be a typed reference implementation:
  • Removed execute() override; replaced with a strongly-typed handlers map.
  • Required<{ [K in testId]: HandlerFn<…> }> annotation makes the map exhaustive at compile time — adding a new test without a handler is a TS error.
  • New DiffusionParams interface (no more unknown/any); buildParams/resolveParams typed end-to-end.
  • Consolidated 4 near-duplicate handlers into a single runBasic(resourceKey, …) via bind.
  • Extracted compareWithBaseline helper for img2img-vs-txt2img and fusion-vs-flux2 comparisons.
  • Extracted readPngDims / assertEqualPngDimensions (parse IHDR) so we no longer false-positive on compressed-byte length differences.
• New minimal asset assets/images/diffusion-img2img-source-256.png (562 B, 256×256 RGB) — keeps SD 2.1 output dimensions matching requested 256×256 and minimizes resource cost.

🧪 How was it tested?

• Desktop: npm run install:build:full → full diffusion suite green locally (FLUX 2 Klein).
• iOS (single dev device): img2img cases green locally (full Device Farm run still skipped at consumer level due to OOM).
• tsc --noEmit clean. Exhaustiveness check verified by removing a handler entry and confirming TS error: Property '"diffusion-standalone-upscaler-x4"' is missing in type … but required in type 'Required<…>'.
• CI run: https://github.com/tetherto/qvac/actions/runs/26229867278

[github-actions]

QVAC E2E — android — ✅ all tests passed (83/91, 2826s)

Config: suite=smoke · filter=(none) · exclude=(none)
View run · Artifacts: reports · Device Farm logs

[github-actions]

QVAC E2E — ios — ✅ all tests passed (82/91, 1200s)

Config: suite=smoke · filter=(none) · exclude=(none)
View run · Artifacts: reports · Device Farm logs · Device Farm logs

[github-actions]

QVAC E2E — windows — ✅ all tests passed (91/91, 429s)

Config: suite=smoke · filter=(none) · exclude=(none)
View run · Artifacts: reports

[github-actions]

QVAC E2E — linux — ✅ all tests passed (91/91, 236s)

Config: suite=smoke · filter=(none) · exclude=(none)
View run · Artifacts: reports

[github-actions]

QVAC E2E — macos — ✅ all tests passed (91/91, 309s)

Config: suite=smoke · filter=(none) · exclude=(none)
View run · Artifacts: reports

[simon-iribarren]

CI Status

Green. PR validation, SDK checks, build, desktop GPU lanes, Android, and iOS e2e are passing. Release/publish jobs are skipped as expected for PR context.

API Surface & Tagging

No public API changes detected. This is test coverage only, so test[skiplog] is appropriate and no [api], [bc], or [mod] tag is required.

Recommendation

Approve. The img2img coverage is scoped to desktop because mobile diffusion remains intentionally skipped for Device Farm memory pressure, which is fine for this PR.

[github-actions]

Tier-based Approval Status

**PR Tier:** TIER1

**Current Status:** ✅ APPROVED

**Requirements:**
- 1 Team Member approval ✅ (1/1)
- 1 Team Lead OR Management approval ✅ (1/1)



---
*This comment is automatically updated when reviews change.*

[kinsta]

Preview deployments for qvac-docs-staging ⚡️

Status	Branch preview	Commit preview
🔁 Deploying...	N/A	N/A

Commit: ebb024e63b7c26b99e2c99ccd0614012b1ce8850

Deployment ID: 71bc0f90-16b6-4317-ad3c-3d102afb26b2

Static site name: qvac-docs-staging-fazwv

[Victor-Rodzko]

/review