fix(db): limit signal snapshot dedupe to cache signals

[JSONbored]

Motivation

• A recent global dedupe removed all but the latest signal_snapshots row per (signal_type, target_key), which deleted bounded historical series that callers rely on (e.g. contributor-decision-pack and queue-health).
• The retention job must remove superseded cache snapshots while preserving historical series used for dashboards and trend computations.

Description

• Add a LATEST_ONLY_SIGNAL_SNAPSHOT_TYPES allowlist containing latest-only cache signal types (repo-culture-profile, repo-doc-refresh-attempt) and document the rationale in the function comment.
• Restrict dedupeSignalSnapshots to only consider signal types present in the allowlist by selecting DISTINCT signal_type WHERE signal_type IN (...) and processing those types only.
• Preserve existing batching/cap logic and dry-run behavior while changing the query shape to bind allowlist placeholders for safe SQL parameters.
• Update test/unit/retention.test.ts to use the canonical latest-only signal names, add a regression test that verifies historical-series signal types (e.g. contributor-decision-pack, queue-health) are preserved, and adjust test mocks accordingly.

Testing

• Ran npx vitest run test/unit/retention.test.ts and all tests in that file passed.
• Ran npm run typecheck which completed successfully.
• Attempted a coverage run with npm run test:coverage scoped to the retention tests; the tests executed successfully but V8 coverage remapping failed with TypeError: jsTokens is not a function before producing a coverage report (test assertions themselves passed).

---

Codex Task

[superagent-security]

Superagent didn't find any vulnerabilities or security issues in this PR.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 93.68%. Comparing base (b8a1186) to head (a5d6c22).
⚠️ Report is 18 commits behind head on main.
✅ All tests successful. No failed tests found.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #3971   +/-   ##
=======================================
  Coverage   93.68%   93.68%           
=======================================
  Files         373      373           
  Lines       34895    34897    +2     
  Branches    12768    12768           
=======================================
+ Hits        32692    32694    +2     
  Misses       1584     1584           
  Partials      619      619           

Files with missing lines	Coverage Δ
src/db/retention.ts	96.49% <100.00%> (+0.12%)	⬆️

🚀 New features to boost your workflow:

• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[loopover-orb]

Warning

🟨🟨🟨🟨🟨🟨🟨🟨🟨🟨🟨🟨

⏸️ Gittensory review result - manual review recommended

_{Review updated: 2026-07-07 08:21:51 UTC}

2 files · 1 AI reviewer · 2 blockers · readiness 93/100 · CI green · clean

⏸️ Suggested Action - Manual Review

• No linked issue detected — If this PR is intended to solve an issue, link it explicitly in the PR body.
• Maintainer requires a linked issue — Link the relevant issue (for example Closes #123) before opening the PR.

Review summary
This narrows dedupeSignalSnapshots to only touch a hardcoded allowlist (repo-culture-profile, repo-doc-refresh-attempt) instead of every distinct signal_type, correctly fixing the prior regression that deleted bounded historical series like queue-health and contributor-decision-pack. The IN-clause/placeholder binding is correctly wired (?1..?N matched to LATEST_ONLY_SIGNAL_SNAPSHOT_TYPES spread in bind), and the per-type stale-row deletion logic inside the loop is unchanged and still correct. Tests were updated to use the canonical signal names and a new regression test explicitly asserts historical types are left untouched, closing the gap that let the original incident happen.

Nits — 4 non-blocking

• The latest-only-vs-historical distinction is now encoded as a standalone string array in retention.ts; if there's a central signal-type registry elsewhere in the codebase, consider deriving this list from it instead of a second hardcoded source of truth that can drift.
• The PR description doesn't cite a tracking issue number for this regression fix beyond referencing obs(selfhost): monitor Cloudflare D1 database size and add bounded dedup retention for signal_snapshots #3810 in a comment — worth confirming this closes/links an actual open issue per repo convention.
• Consider adding a comment near RETENTION_POLICY or the signal-snapshot writer noting that any new latest-only cache signal type must be added to LATEST_ONLY_SIGNAL_SNAPSHOT_TYPES, so future signal types default to 'preserved' rather than silently needing an opt-in.
• The dry-run/no-meta defensive tests in retention.test.ts were adjusted to branch on `sql.includes('SELECT DISTINCT')`; this is a bit brittle (string-matching SQL) but acceptable as an existing test pattern.

Concerns raised — review before merging

• No linked issue detected — If this PR is intended to solve an issue, link it explicitly in the PR body.
• Maintainer requires a linked issue — Link the relevant issue (for example Closes #123) before opening the PR.

Signal	Result	Evidence
Code review	❌ 2 blockers	1 reviewer
Linked issue	⚠️ Missing	No linked issue or no-issue rationale found.
Related work	✅ No active overlap found	No same-issue or scoped active PR overlap found.
Change scope	✅ 20/20	Low review scope from cached public metadata (no linked issue context).
Validation posture	✅ 25/25	PR body includes validation/test evidence.
Contributor workload	✅ 10/10	Author activity: 51 registered-repo PR(s), 43 merged, 343 issue(s).
Contributor context	✅ Confirmed Gittensor contributor	JSONbored; Gittensor profile; 51 PR(s), 343 issue(s).
Gate result	❌ Blocking	Repo-configured hard blocker found.

Review context

• Author: JSONbored
• Role context: owner (maintainer lane)
• Public audience mode: oss maintainer
• Lane context: Repository is configured for direct PR review.
• Public profile languages: Python, TypeScript, JavaScript, Ruby, Go, Kotlin, MDX, Shell
• Official Gittensor activity: 51 PR(s), 343 issue(s).
• PR-specific overlap: none found.

Contributor next steps

• Treat this as maintainer-lane context rather than normal contributor-lane activity.
• Explain no-issue PR.
• Link the issue being solved, or explicitly explain why this is a no-issue PR.

Signal definitions

• Related work = same linked issue, overlapping active PRs, or title/path similarity.
• Change scope = cached public metadata such as size labels, draft state, and review-burden hints.
• Validation posture = whether the PR provides enough public validation/test evidence for maintainer review.
• Contributor workload = public contributor activity and cleanup pressure, not a repo-wide quality failure.
• Contributor context = public GitHub/Gittensor identity context; non-Gittensor status is not a blocker.

_{🟩 Safe / merged · 🟦 Advisory · 🟨 Held for review · 🟥 Blocked / closed}

---

💰 Earn for open-source contributions like this. Gittensor lets GitHub contributors earn for the work they already do — register to start earning →.

Checked by Gittensory, a quiet PR intelligence layer for OSS maintainers.

• Re-run Gittensory review