# QA Plan: Personalization Correctness Verification

## Scope

This feature is itself a verification suite. QA validates that:

1. All tests in `tidal/tests/pg1_personalization_correctness.rs` compile and pass.
2. No regressions are introduced in existing test suites.
3. The mathematical assertions are sound (not trivially true).

## Test Execution

### Step 1: Run the new test suite

```bash
cargo test --manifest-path tidal/Cargo.toml --test pg1_personalization_correctness -- --nocapture
```

**Pass criteria:** All tests pass. Zero failures, zero panics.

### Step 2: Run existing lib tests for regression

```bash
cargo test --manifest-path tidal/Cargo.toml --lib
```

**Pass criteria:** Same pass count as before (no regressions).

### Step 3: Run clippy and fmt

```bash
cargo fmt --manifest-path tidal/Cargo.toml -- --check
cargo clippy --manifest-path tidal/Cargo.toml --tests -- -D warnings
```

**Pass criteria:** No warnings, no formatting errors in our test file.

### Step 4: Verify test assertions are non-trivial

Manual review of each test to verify:

- Decay tests use different timestamps and weights (not all zeros or ones).
- Score ordering tests have items with meaningfully different signal histories.
- Reactivity tests actually measure wall-clock time (not just assert results).
- Gate tests have items both above and below the threshold.

**Pass criteria:** Reviewer confirms assertions are non-trivial.

## QA Checklist

| # | Check | Expected |
|---|-------|----------|
| 1 | `pg1_personalization_correctness` tests all pass | 0 failures |
| 2 | Lib tests pass (no regression) | Same count as baseline |
| 3 | `cargo fmt --check` clean for test file | No diffs |
| 4 | `cargo clippy --tests` clean for test file | No warnings |
| 5 | Decay tests use varied inputs | Confirmed by review |
| 6 | Ordering tests compare items with different signal profiles | Confirmed by review |
| 7 | Reactivity tests include timing assertions | Confirmed by review |