[2G-Bug-02]: Credit habit_completions rows toward evaluate_graduation "Already done" rate on expired nudges

[WilliM233]

Verification

• ruff check . — ✅ Pass (All checks passed!)
• pytest -v — ✅ Pass (1292 passed, 4 skipped)
• Migration applied locally on brain3-dev: N/A (no schema change)
• Postgres-backed test confirmed: N/A (pure ORM logic; no Postgres-specific behavior)

Ran locally against develop HEAD at 90d6989 immediately before opening this PR.

Summary

evaluate_graduation now credits a same-date habit_completions row when an expired nudge has no response. The result: a user who completes a habit through any path (direct complete, routine cascade, reconciliation) gets credit toward graduation — not only when they explicitly tap "Already done" on a nudge. This is the ADHD quality lens in action: multiple low-friction completion paths work better than a single path, and all of them feel like "I did it" to the user.

Explicit responses still win. An explicit "Already done" counts on its own (back-compat preserved). An explicit non-"Already done" response (e.g., "Skip today") is never overridden by a completion row — a user saying "no" to a nudge means no, even if reconciliation later recorded a completion.

Changes

• app/services/graduation.py — evaluate_graduation now queries habit_completions over the evaluation window and builds a set of completion dates. The already_done_count derivation iterates notifications: "Already done" responses count as before; expired notifications with response IS NULL count if a same-date completion exists; all other explicit responses are untouched. GraduationResult shape unchanged.
• tests/test_graduation.py — new TestEvaluateGraduationHybridCredit class covers the five acceptance tests from the issue (routine-cascade credit, individual credit, no-credit-without-row, no-credit-on-date-mismatch, explicit-"Already done" back-compat) plus an explicit-negative-override guard.

How to Verify

1. git checkout feature/2G-Bug-02-evaluate-graduation-hybrid-credit
2. ruff check . — passes clean
3. pytest -v tests/test_graduation.py — all 160 graduation tests pass, including the 6 new hybrid-credit tests
4. pytest -v — full suite: 1292 passed, 4 skipped

Deviations

None. Implementation matches the GitHub issue acceptance criteria and [2G-01] v2 Amendment 12.

Per Wave 2 brief guidance (D7, Wave 3 #186): the hybrid-credit logic here is deliberately inline. Duplication with the companion step-down fix (#185) is expected and retained — the shared read_completion_history helper refactor lands in Wave 3 as #186.

Test Results

```
$ ruff check .
All checks passed!

$ pytest -v
====================== 1292 passed, 4 skipped in 21.73s =======================
```

New tests in TestEvaluateGraduationHybridCredit:

• test_routine_cascade_completion_credited_toward_graduation
• test_individual_completion_credited_toward_graduation
• test_no_credit_when_no_completion_row
• test_no_credit_when_date_mismatch
• test_response_already_done_still_counts_without_completion_row
• test_explicit_negative_response_not_overridden_by_completion

All existing graduation tests continue to pass (AC #6).

Acceptance Checklist

• evaluate_graduation reads both notification_queue.response='Already done' and habit_completions rows dated to expired-nudge dates
• Expired notification with response IS NULL + matching same-date HabitCompletion → counted toward already_done_count
• Explicit "Already done" continues to count regardless of completion-row presence (back-compat)
• Non-"Already done" responses (e.g., "Skip today") are NOT overridden by a completion row
• GraduationResult response shape unchanged; only already_done_count / derived current_rate differ
• All existing graduation tests continue to pass
• New integration tests added (see above)

Closes #184