refactor(store): install_meta_belief → INSERT OR IGNORE for atomic idempotence (#772)

[robotrocketscience]

Closes #772. Follow-up to #755 (PR #771) per the disposition in #772.

What this PR does

Refactors MemoryStore.install_meta_belief from a SELECT-then-INSERT
idempotency check to the project's established INSERT OR IGNORE +
rowcount idiom (10 existing call sites in store.py). Eliminates the
TOCTOU window between SELECT and INSERT.

Surface

• src/aelfrice/store.py — install_meta_belief body. Validation
  (ValueError raises for invalid static_default / half_life_seconds
  / unknown signal classes) runs unchanged before the INSERT. Net
  reduction: 11 lines → 8 lines, one fewer SQLite round-trip.

• tests/test_meta_beliefs.py — new test
  test_install_concurrent_race_returns_false_not_raises exercises the
  rowcount == 0 return path: second install on the same key returns
  False, does not raise sqlite3.IntegrityError. Single-threaded
  (deterministic); the existing
  test_install_meta_belief_is_idempotent_returns_false_on_reinstall
  covers the no-race idempotency case.

Behaviour deltas

Case	Before	After
First install	True	True
Re-install (same key, sequential)	False	False
Concurrent install race (would-have-raced)	IntegrityError raised on loser	False returned on loser
Invalid config (out-of-range / unknown class)	ValueError	ValueError (unchanged)

Acceptance mapping (#772)

Acceptance item	Where covered
install_meta_belief switches to INSERT OR IGNORE + rowcount check	store.py body (commit 55e5764)
Existing test_install_meta_belief_is_idempotent_returns_false_on_reinstall still passes	pytest tests/test_meta_beliefs.py → 20 passed
Existing test_install_rejects_invalid_config still passes (validation order unchanged)	Same
New test exercises the would-have-raced path	test_install_concurrent_race_returns_false_not_raises (commit b9f7c0d)

Verification

• uv run pytest tests/test_meta_beliefs.py — 20 passed.
• git log --format='%h %G? %s' github/main..HEAD — both commits signed (G).
• git merge-base --is-ancestor github/main HEAD — FF on github/main.
• Discretion grep on the diff — clean.

Out of scope

• BEGIN IMMEDIATE for update_meta_belief's read-decay-write cycle.
  Per refactor(store): install_meta_belief → INSERT OR IGNORE for atomic idempotence (post-#755) #772 disposition: single-writer federation rule (v3.0 design decision: federation write model — multi-writer CRDT vs read-only #661,
  d0c5ecdebb3f0f4d) makes the race not manifest, and no other
  store.py method uses BEGIN IMMEDIATE. Consistency over local
  fix.
• encode_signal_weights canonicality (int vs float bytes). Separate
  follow-up if surfaced.

[sourcery-ai]

Sorry @robotrocketscience, you have reached your weekly rate limit of 500000 diff characters.

Please try again later or upgrade to continue using Sourcery

[coderabbitai]

Warning

Rate limit exceeded

@robotrocketscience has exceeded the limit for the number of commits that can be reviewed per hour. Please wait 22 minutes and 21 seconds before requesting another review.

You’ve run out of usage credits. Purchase more in the billing tab.

⌛ 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: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: a606aac0-22f4-4b6d-8ca1-57ea857d46f3

📥 Commits

Reviewing files that changed from the base of the PR and between 0f0d138 and b9f7c0d.

📒 Files selected for processing (2)

• src/aelfrice/store.py
• tests/test_meta_beliefs.py

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch feat/issue-772-install-or-ignore

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

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

[robotrocketscience]

[claim:review:curie:2026-05-14T01:41:31Z]

[robotrocketscience]

Review (LGTM, adding ready-to-merge).

Diff matches issue #772 exactly: SELECT-then-check → INSERT OR IGNORE + cur.rowcount > 0. Validation order preserved (ValueError raises run before the insert). Net -3 lines, one fewer SQLite round-trip.

Idiom alignment: 10 existing INSERT OR IGNORE call sites in store.py, multiple rowcount > 0 / or 0 returns — this fits.

Test: test_install_concurrent_race_returns_false_not_raises exercises the rowcount==0 path single-threaded as the acceptance item specifies. The try/except sqlite3.IntegrityError → AssertionError framing makes the regression-mode obvious if the idiom drifts back.

Verification:

• FF on github/main ✓
• Both commits SSH-signed (G) ✓
• pytest (3.12) + pytest (3.13) SUCCESS ✓
• Staging gate (secrets / pattern / history / commit-msg / pr-title / pr-body) SUCCESS ✓
• CodeQL (python + actions) SUCCESS ✓
• deptry / vulture / typos SUCCESS ✓
• Discretion grep on diff vs main: clean ✓

Minor observation, non-blocking: encode_signal_weights(signal_weights) now runs on the no-op re-install path (it short-circuited before). It's pure / cheap so no functional concern.

[robotrocketscience]

[release:review:curie:2026-05-14T01:43:33Z]

[github-actions]

merge-train: merged b9f7c0d → main via FF push.