fix: cap benchmark confidence to measured evidence

[edhor1608]

Cap benchmark confidence to 0.75 and restrict 'high' label to ≥0.9

• computeConfidence in benchmark.ts now derives confidence solely from objective metrics: PSNR (0.35), SSIM (0.35), bitrate delta (0.2), and output colorspace (0.1); contributions from codec fields and output_matches_target are removed.
• The computed value is clamped to a maximum of 0.75, preventing automated runs from ever producing a 'high' label.
• confidenceLabel raises the 'high' threshold from ≥0.8 to ≥0.9, so only manually-supplied values above the cap could reach that label.
• Behavioral Change: all existing benchmarks that previously reported {label:'high', value:1} now report {label:'medium', value:0.75}; snapshots and integration test bounds updated accordingly.

^{Macroscope summarized 5ee7fdd.}

[coderabbitai]

Warning

Rate limit exceeded

@edhor1608 has exceeded the limit for the number of commits that can be reviewed per hour. Please wait 3 minutes and 48 seconds before requesting another review.

To keep reviews running without waiting, you can enable usage-based add-on for your organization. This allows additional reviews beyond the hourly cap. Account admins can enable it under billing.

⌛ How to resolve this issue?

After the wait time has elapsed, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

We recommend that you space out your commits to avoid hitting the rate limit.

🚦 How do rate limits work?

CodeRabbit enforces hourly rate limits for each developer per organization.

Our paid plans have higher rate limits than the trial, open-source and free plans. In all cases, we re-allow further reviews after a brief timeout.

Please see our FAQ for further information.

ℹ️ Review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: ASSERTIVE

Plan: Pro

Run ID: 2c13b724-e620-443f-b1d0-64532e064a37

📥 Commits

Reviewing files that changed from the base of the PR and between 3038201 and 5ee7fdd.

📒 Files selected for processing (12)

• docs/repo_refresh_audit.md
• src/domain/benchmark.ts
• tests/fixtures/e2e/snapshots/benchmark/benchmark-reliable-feed-landscape-white-classic.json
• tests/fixtures/e2e/snapshots/benchmark/benchmark-reliable-feed-portrait-png.json
• tests/fixtures/e2e/snapshots/benchmark/benchmark-reliable-reel-portrait-mp4-audio.json
• tests/fixtures/e2e/snapshots/benchmark/benchmark-reliable-reel-white-canvas.json
• tests/fixtures/e2e/snapshots/validate_matrix/validate-matrix-basic.json
• tests/fixtures/e2e/snapshots/validate_matrix/validate-matrix-max-cases-1.json
• tests/fixtures/e2e/snapshots/validate_matrix/validate-matrix-only-file-portrait.json
• tests/fixtures/e2e/snapshots/validate_matrix/validate-matrix-only-portrait.json
• tests/fixtures/e2e/snapshots/validate_matrix/validate-matrix-with-failure-fail-on-error.json
• tests/integration/benchmark.integration.test.ts

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch repo-refresh-benchmark-confidence-evidence

✨ Simplify code

• Create PR with simplified code
• Commit simplified code in branch repo-refresh-benchmark-confidence-evidence

---

_{Review rate limit: 0/5 reviews remaining, refill in 3 minutes and 48 seconds.}

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[edhor1608]

• test: update report workflow snapshots #55
• feat: add consistent cli help #54
• refactor: split media inspector parsers #53
• test: replace bulky pixel snapshots #52
• fix: cap benchmark confidence to measured evidence #51 👈 (View in Graphite)
• feat: add upload workflow report checks #50
• chore: quiet lint gate output #49
• test: isolate integration output files #48
• fix: validate config files deeply #47
• refactor: route media subprocesses through adapter #46
• refactor: share CLI argument parsing #45
• chore: replace legacy agent guide #44
• fix: load export profiles from module path #43
• chore: add repo refresh onboarding and CI #42
• fix: harden config and image helper loading #41
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.