fix(service-analytics): narrow the executeAggregate config mirrors' aggregations[].method to AggregationFunction

[os-elon]

Fixes#12940

Type-only. #12776 narrowed StrategyContext.executeAggregate's aggregations[].method from string to the six-value AggregationFunction; this brings @objectstack/service-analytics' own consumer-local mirrors of that same slot into line, so the compile-time vocabulary reaches the people who write a custom bridge instead of stopping at the package boundary.

Sites: three, not two

The card enumerated two. Deriving the population from the tree (grep -rn 'method:' packages/services/service-analytics/src) found three declarations of the slot, all narrowed here:

site	what it is
plugin.ts — AnalyticsServicePluginOptions.executeAggregate	the declaration an app author's own bridge is typed against
analytics-service.ts — AnalyticsServiceConfig.executeAggregate	the config twin whose comment claims lockstep with the contract
plugin.ts — parseEngineAggregateFunction(method, alias)	the auto-bridge's runtime parse, the third place a reader was told the vocabulary is open

Producers were already narrow before this PR (ObjectQLStrategy's aggregation locals carry AggregationFunction end to end since #12776), so this closes the split rather than opening a new one.

Prose repaired

• The lockstep comment (analytics-service.ts) now states the lockstep member by member and names method as the enum, so the next drift is visible in review rather than silently true-then-false.
• The historical justification in the auto-bridge and in parseEngineAggregateFunction's docblock no longer explains the parse by pointing at a method: string declaration that no longer exists. Both now use the wording the finding: StrategyContext.executeAggregate declares aggregations[].method as string while the engine contract declares the six-value AggregationFunction #12776 changeset already settled on: with both ends of the method to function rename declaring the same enum, the parse is defence in depth behind a compile-time check rather than the only check.
• The runtime parse is unchanged and stays — its docblock now says why erased types keep it load-bearing (a JavaScript app supplying its own bridge; host drift through a cube that never met CubeSchema, which is the path aggregate-bridge-function-vocabulary.test.ts drives end to end). finding: two consumer-local structural DataEngineLike re-declarations remain after the #11493 deletion (datasource-admin-plugin, service-analytics) #11833 is why it exists.

Verification (all at 978a4a9)

• pnpm --filter '@objectstack/service-analytics^...' build then pnpm --filter @objectstack/service-analytics typecheck — tsc --noEmit, exit 0.
• pnpm --filter @objectstack/service-analytics exec vitest run --maxWorkers=2 — 83 files, 1805 tests passed.
• The typecheck really covers the test mirrors: tsc --noEmit --listFiles puts 81 test files from src/__tests__ in the program, including all nine that carry their own { field: string; method: string; alias: string } doubles. Those nine needed no edit — a handler accepting method: string accepts a superset and stays assignable (parameter contravariance), exactly as the finding: StrategyContext.executeAggregate declares aggregations[].method as string while the engine contract declares the six-value AggregationFunction #12776 changeset predicted for implementors. No any, no @ts-expect-error, no widening, and nothing excluded from the tsc program.
• Reverse verification A (caller position) — a temporary probe filling method: 'median' against both config mirrors: tsc exit 2 with TS2322: Type '"median"' is not assignable to type '"min" | "max" | "count" | "sum" | "avg" | "count_distinct"' at both sites, while a sibling method: 'sum' call stayed clean. Probe removed; git diff HEAD empty; tsc exit 0.
• Reverse verification B (ablation of the narrowing) — widening the plugin mirror back to string turned tsc red at plugin.ts(346,52), TS2345: Argument of type 'string' is not assignable to parameter of type ..., i.e. at the parseEngineAggregateFunction call. That is what makes the third site load-bearing rather than decorative. Mutation confirmed on disk by anchored grep -c (1 widened / 0 narrowed) and by git hash-object; restored with git checkout HEAD -- and proven byte-identical to the HEAD blob, git diff HEAD empty, tsc exit 0 again.
• Gates derived at edit time from the actual changed set with node scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack (26 matched families plus the convention-triggered check:dispatcher-error-vocabulary) — all green, including check:type-check-coverage (OK - 66/78 workspace packages type-checked), check:test-source-alias, check:type-source-resolution, check:cross-package-test-inputs, check-adr-0087-registration (this PR adds no declared-breaking changeset), check-empty-changeset, check-changeset-no-major.
• Not measured locally, named rather than counted as passes: check-test-completeness.mjs and scripts/pm/check-half-states.mjs both exit 3 = PREREQUISITE NOT MET in this container (no test log to hand the first; no valid GitHub credential for the second). check:type-check-debt exits 1 only in its --re-measure leg, which refuses to measure without a built closure for 48 workspace dependencies; its --self-test and coverage legs passed. That refusal cannot hide a ratchet red from this diff: no DEBT entry depends on service-analytics, and of the three TEST_DEBT entries that do (runtime, cli, verify), only runtime references the narrowed surface at all — one test double typed options: any. CI builds the closure before the step, as lint.yml does.
• Lint, narrowed and declared: eslint --no-inline-config --format json over the 2 changed files, 0 errors / 0 warnings. The narrowing is a measurement, not a skip: this repo runs one eslint.config.mjs which, by its own documented and positively-controlled statement, "never enables type-aware linting (no parserOptions.project, no typed @typescript-eslint rules) for ANY file", so a declaration change here cannot move a verdict on a file it does not edit.

Changeset: patch for @objectstack/service-analytics.

---

Generated by Claude Code

[github-actions]

📓 Docs Drift Check

This PR changes 1 package(s): @objectstack/service-analytics, touching 3 documentable anchor(s).

1 hand-written doc(s) NAME something this change touched and may need an implementation-accuracy re-verification:

• content/docs/plugins/packages.mdx(via AnalyticsServicePlugin (symbol))

What this run could not see

• the SDK route bridge reached 47 of 219 client-bound route-ledger rows — the other 172 have no registrar path: tail to select them, so pages documenting THEIR client methods cannot appear above, on this or any run. Of those 172: 14 are remediable by widening that discovery convention (an in-repo file declares the path; the convention did not scan it); 56 are structural — on a ledger where NOT ONE row is declared in-repo, so no discovery change reaches them at any price; 102 are undecided (no in-repo declaration, on a ledger that has other in-repo registrars — absence and an unreadable spelling are not distinguishable here). The rows themselves: node scripts/docs-audit/affected-docs.mjs --bridge-coverage
• a page that states a rule by its inputs shares no identifier with the emitter that implements the rule, so an emitter-only diff cannot list it — not on this run and not on any run. Measured on fix(driver-sql): emit varchar(maxLength) for a text field a declared index keys on #11430: content/docs/protocol/objectql/types.mdx documents the text-family column mapping by the ObjectQL type names it maps FROM (text / textarea / html) while the diff changed createColumn; it went unlisted, and it was the page that diff falsified, in four places. No shared token exists to detect this on, so a rule your change carries has to be re-read by hand in the pages that restate it.

Coarse fallback — 8 page(s) merely mention a changed package (the pre-#9192 predicate, kept for the deliberately-wide backstop): node scripts/docs-audit/affected-docs.mjs --json e5ed9434365e6a7520b5070da441f0552030b37d → packageMentionDocs.

Which tree this was computed on

This run read content/docs from 864525b3b12f52c693b27c0b3b8e01aec2e3d96e — the merge of head 978a4a9abb79e731d905ba468aea0a2a37d4f01d into base e5ed9434365e6a7520b5070da441f0552030b37d, which is what actions/checkout gives a pull_request run. Not the PR head.

A worktree cut from an older main holds a different content/docs, so re-deriving there can legitimately return a different list — that is a different tree, not a wrong row. To answer on the same tree:

# while this PR is open — GitHub drops the merge commit once it closes
git fetch origin 864525b3b12f52c693b27c0b3b8e01aec2e3d96e && git checkout 864525b3b12f52c693b27c0b3b8e01aec2e3d96e
# afterwards, rebuild it from the two parents, which stay fetchable
git fetch origin e5ed9434365e6a7520b5070da441f0552030b37d 978a4a9abb79e731d905ba468aea0a2a37d4f01d && git checkout -B drift-repro e5ed9434365e6a7520b5070da441f0552030b37d && git merge --no-ff 978a4a9abb79e731d905ba468aea0a2a37d4f01d
node scripts/docs-audit/affected-docs.mjs --json e5ed9434365e6a7520b5070da441f0552030b37d

⚠️ That checkout carried uncommitted changes, so the commit above does not fully identify what was read.

Advisory only, and a precision-first one (#9192): a page is listed because it names a
symbol, wire route or SDK method this diff touched — not because it mentions a changed
package. Each row says which anchor put it there, so a wrong row is reportable rather than
merely annoying. To re-verify, run the docs-accuracy-audit workflow scoped to these files:
node scripts/docs-audit/affected-docs.mjs e5ed9434365e6a7520b5070da441f0552030b37d → pass the list as
args.docs, on the commit named under Which tree this was computed on.