diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml deleted file mode 100644 index cdb2ab14..00000000 --- a/.github/workflows/ci.yml +++ /dev/null @@ -1,40 +0,0 @@ -name: CI - -on: - push: - pull_request: - -jobs: - node: - name: Node ${{ matrix.node-version }} - runs-on: ubuntu-latest - strategy: - fail-fast: false - matrix: - node-version: ['22', '24'] - steps: - - uses: actions/checkout@v4 - - uses: actions/setup-node@v4 - with: - node-version: ${{ matrix.node-version }} - cache: npm - - run: npm ci - - run: npm run test:unit - - bun: - name: Bun (latest) - runs-on: ubuntu-latest - steps: - - uses: actions/checkout@v4 - - uses: actions/setup-node@v4 - with: - node-version: '22' - cache: npm - - uses: oven-sh/setup-bun@v2 - with: - bun-version: latest - - run: npm ci - # test:bun imports the built dist/, so build first. - - run: npm run build - # Bun as a runtime is the supported Bun story (`bun add` / `bun run`). - - run: npm run test:bun diff --git a/CHANGELOG.md b/CHANGELOG.md index b6dd91fd..a5f344cc 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -2,6 +2,14 @@ All notable changes to this project will be documented in this file. See [standard-version](https://github.com/conventional-changelog/standard-version) for commit guidelines. +### [8.10.1](https://source.soulcraft.com/soulcraft/brainy/compare/v8.10.0...v8.10.1) (2026-07-24) + +- refactor: remove the orphaned transaction-result type left behind by the dead-path removal (edf123a5) +- fix: warm() metadata surface routes through the active provider (warm hook added to the metadata contract); add maintenanceDebt() observability surface (5b2cbf74) +- fix: transaction timeouts are a typed no-hot-retry contract; engine-side non-retry pinned; dead transaction path removed (003e2a74) +- chore: the forge is the address — retire the archived mirror from every live surface (22702b81) + + ### [8.10.0](https://github.com/soulcraftlabs/brainy/compare/v8.9.0...v8.10.0) (2026-07-23) - docs: adoption storefront — contributing guide, security policy, README support + cor section (9a99a7b) diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md index ef9c4a51..d277091d 100644 --- a/CONTRIBUTING.md +++ b/CONTRIBUTING.md @@ -10,10 +10,6 @@ The source of truth is a self-hosted forge: **source.soulcraft.com/soulcraft/bra It's anonymously readable and cloneable — no account needed to browse, clone, or build. -**github.com/soulcraftlabs/brainy** is a public read-only mirror. It's a fine -place to read code or star the project, but issues and pull requests opened -there won't be picked up — please use one of the paths below instead. - ## How to contribute **Found a bug, or have an idea?** Email **brainy@soulcraft.com**. No account, diff --git a/README.md b/README.md index 2fc42060..2caf6493 100644 --- a/README.md +++ b/README.md @@ -13,7 +13,7 @@

npm version npm downloads - CI + CI Documentation MIT License TypeScript diff --git a/RELEASES.md b/RELEASES.md index 89bf38e7..03d7283a 100644 --- a/RELEASES.md +++ b/RELEASES.md @@ -31,6 +31,68 @@ is sometimes cited as a 7.x removal — those methods never existed on 7.x; the --- +## v8.10.1 — 2026-07-24 (the no-hot-retry contract + warm()'s metadata surface under native providers) + +From a production incident: a native-provider op ground 38-40s inside a transaction, +blew the ~32s apply-phase budget, was rolled back (zero loss, by design), and a +downstream pipeline hot-retried the identical operation into a 6-minute, 100%-CPU +storm. Investigation confirmed Brainy itself never auto-retries a timed-out +transaction — the storm was entirely the consumer's own retry loop, driven by a +"retryable" doc-prose claim with no machine-readable contract to branch on. This +release closes that contract gap and, separately, fixes a real `warm()` reporting gap +surfaced by the same investigation. + +- **`TransactionTimeoutError` is now a machine-readable no-hot-retry contract.** Two + new typed, always-`true` fields replace prose-only guidance: + - `retryable: true` — the operation MAY succeed on a later attempt, once the + underlying slowness resolves or the budget is deliberately raised + (`transactionBudgetFloorMs`, or a batch's own `timeoutMs` override). + - `hotRetryUnsafe: true` — an immediate, identical retry re-pays the FULL cost of + the work that just timed out (it does not resume partway) and can cascade into + exactly the CPU storm above. **Never loop on this error.** The documented pattern + is a latch, not a retry loop: + ``` + on TransactionTimeoutError: + record { at: Date.now(), error } + rethrow loudly to your own caller + hold a cooldown window before any re-attempt + clear the latch only on a subsequent success + ``` + - `context` (unchanged, now fully documented) carries the backoff inputs: + `timeoutMs`, `operationIndex`, `elapsedMs`, `totalOperations`, `operationName`. + - Every "retryable" doc-prose site referencing this error (`transact()`'s + `timeoutMs` option, `transactionBudgetFloorMs`, `Transaction.execute()`) now + points at these fields instead of bare prose. + - Regression-pinned: the engine never internally re-drives a timed-out operation + (verified via an execution counter through both the single-op write path and + `add()`'s upsert-race retry loop), so this has always been true — it is now + provable and typed. +- **Dead code removed**: `TransactionManager.executeTransactionWithResult()` had zero + callers in this codebase and is deleted. +- **`brain.warm()`'s metadata surface now routes through the ACTIVE provider.** A + production deployment's warm report showed `metadata: 'unavailable'` under a native + metadata provider — the previous logic only duck-typed the built-in JS manager's + `hydrateAll()` method, which a native provider has no reason to implement. The + metadata provider contract (`MetadataIndexProvider`, `src/plugin.ts`) gains an + optional `warm?(): Promise` hook, mirroring the existing vector and graph + provider hooks. `brain.warm()` now checks the active provider's own `warm()` FIRST, + falls back to the JS manager's `hydrateAll()` when absent, and only reports + `'unavailable'` when neither exists — never `init()` as a stand-in, since a native + provider's `init()` may be a cheap verify rather than a real warm. A native + provider lights this surface up the same way `@soulcraft/cor` already lights the + vector and graph surfaces: implement `warm()` on its metadata provider. +- **New: `brain.maintenanceDebt()`** — the observability seam so an operator sees a + provider's outstanding background maintenance work (pending bytes/items, last pass + outcome, whether it's converging) BEFORE it grinds into the kind of budget-busting + op this release's timeout contract exists for, instead of discovering it as a CPU + storm. It is a pure passthrough: brainy applies no thresholds, no polling, and no + estimation — it calls each active provider's own optional `maintenanceDebt?()` hook + (vector, metadata, graph — the same three contracts `warm?()` lives on) and reports + the payload verbatim, or `'unavailable'` when a surface's provider doesn't track + debt. Useful as a pre-warm/post-warm check or a boot gate. `@soulcraft/cor` does not + yet implement the hook as of this release — expect it on cor's next release; until + then all three surfaces honestly report `'unavailable'`. + ## Unreleased (the warm contract: cold-restart writes stop paying demand-load latency) From a production deployment's cold-restart incident: the FIRST writes after every diff --git a/package-lock.json b/package-lock.json index 37aeb81d..d0c7b9d9 100644 --- a/package-lock.json +++ b/package-lock.json @@ -1,12 +1,12 @@ { "name": "@soulcraft/brainy", - "version": "8.10.0", + "version": "8.10.1", "lockfileVersion": 3, "requires": true, "packages": { "": { "name": "@soulcraft/brainy", - "version": "8.10.0", + "version": "8.10.1", "license": "MIT", "dependencies": { "@msgpack/msgpack": "^3.1.2", diff --git a/package.json b/package.json index e4bc8144..ce670369 100644 --- a/package.json +++ b/package.json @@ -1,6 +1,6 @@ { "name": "@soulcraft/brainy", - "version": "8.10.0", + "version": "8.10.1", "description": "Universal Knowledge Protocol™ - World's first Triple Intelligence database unifying vector, graph, and document search in one API. Stage 3 CANONICAL: 42 nouns × 127 verbs covering 96-97% of all human knowledge.", "main": "dist/index.js", "module": "dist/index.js", @@ -128,13 +128,13 @@ "publishConfig": { "access": "public" }, - "homepage": "https://github.com/soulcraftlabs/brainy", + "homepage": "https://source.soulcraft.com/soulcraft/brainy", "bugs": { - "url": "https://github.com/soulcraftlabs/brainy/issues" + "url": "https://source.soulcraft.com/soulcraft/brainy/issues" }, "repository": { "type": "git", - "url": "git+https://github.com/soulcraftlabs/brainy.git" + "url": "git+https://source.soulcraft.com/soulcraft/brainy.git" }, "files": [ "dist/**/*.js", diff --git a/scripts/release.sh b/scripts/release.sh index 42f5b345..43fa50bd 100755 --- a/scripts/release.sh +++ b/scripts/release.sh @@ -142,7 +142,7 @@ else fi # Create new changelog entry -CHANGELOG_ENTRY="### [${NEW_VERSION}](https://github.com/soulcraftlabs/brainy/compare/v${CURRENT_VERSION}...v${NEW_VERSION}) ($(date +%Y-%m-%d)) +CHANGELOG_ENTRY="### [${NEW_VERSION}](https://source.soulcraft.com/soulcraft/brainy/compare/v${CURRENT_VERSION}...v${NEW_VERSION}) ($(date +%Y-%m-%d)) ${COMMITS} " @@ -175,42 +175,59 @@ echo -e "${BLUE}7️⃣ Creating git tag v${NEW_VERSION}...${NC}" git tag -a "v${NEW_VERSION}" -m "Release v${NEW_VERSION}" echo -e "${GREEN}✅ Tag created${NC}\n" -# Step 9: Push to origin (source of truth) and the public GitHub mirror +# Step 9: Push to origin — the forge is the one home (ruled 2026-07-23; the +# old public GitHub repo is archived history, no longer part of any release). echo -e "${BLUE}8️⃣ Pushing to origin...${NC}" git push --follow-tags origin "$CURRENT_BRANCH" echo -e "${GREEN}✅ Pushed to origin${NC}\n" -# The public GitHub repo is a mirror of origin with an unknown sync cadence. -# `gh release create` below targets GitHub directly: if the new tag hasn't -# reached GitHub yet, gh would CREATE it — pointed at GitHub's default-branch -# head, i.e. the wrong commit. Push branch+tag to GitHub explicitly, then -# verify the tag resolves there to the same commit before any release is cut. -GITHUB_URL="https://github.com/soulcraftlabs/brainy.git" -echo -e "${BLUE}8️⃣½ Pushing to the public GitHub mirror...${NC}" -git push --follow-tags "$GITHUB_URL" "$CURRENT_BRANCH" -LOCAL_TAG_SHA="$(git rev-parse "v${NEW_VERSION}^{}")" -GITHUB_TAG_SHA="$(git ls-remote --tags "$GITHUB_URL" "v${NEW_VERSION}^{}" | cut -f1)" -if [ "$LOCAL_TAG_SHA" != "$GITHUB_TAG_SHA" ]; then - echo -e "${RED}❌ Tag v${NEW_VERSION} on GitHub (${GITHUB_TAG_SHA:-absent}) does not match local (${LOCAL_TAG_SHA}) — aborting before npm publish. Fix the mirror, then re-run.${NC}" +# Step 10: Publish — forge FIRST (home), npmjs second (the world's storefront). +# The fleet-wide ~/.npmrc maps the @soulcraft scope to the forge registry, and +# a scope mapping BEATS `--registry` on the command line — so each publish +# names its registry via the scope override explicitly. Nothing implicit. +FORGE_NPM_REG="https://source.soulcraft.com/api/packages/soulcraft/npm/" +FORGE_NPM_TOKEN_FILE="$HOME/.config/soulcraft/npm-publish-brainy.token" +echo -e "${BLUE}9️⃣ Publishing to the forge registry (home)...${NC}" +if [ -f "$FORGE_NPM_TOKEN_FILE" ]; then + TMPRC="$(mktemp)" + chmod 600 "$TMPRC" + { + echo "@soulcraft:registry=${FORGE_NPM_REG}" + echo "//source.soulcraft.com/api/packages/soulcraft/npm/:_authToken=$(cat "$FORGE_NPM_TOKEN_FILE")" + } > "$TMPRC" + if npm publish --tag "$NPM_TAG" --userconfig "$TMPRC"; then + echo -e "${GREEN}✅ Published to the forge${NC}\n" + else + rm -f "$TMPRC" + echo -e "${RED}❌ Forge publish FAILED — aborting before npmjs so the pair never diverges. Fix and re-run.${NC}" + exit 1 + fi + rm -f "$TMPRC" +else + echo -e "${RED}❌ Forge publish token missing (${FORGE_NPM_TOKEN_FILE}) — aborting. The forge is home; publish it first or restage the token.${NC}" exit 1 fi -echo -e "${GREEN}✅ GitHub mirror has the tag at the right commit${NC}\n" -# Step 10: Publish to npm -echo -e "${BLUE}9️⃣ Publishing to npm (dist-tag: ${NPM_TAG})...${NC}" -npm publish --tag "$NPM_TAG" +echo -e "${BLUE}9️⃣½ Publishing to npmjs (storefront, dist-tag: ${NPM_TAG})...${NC}" +npm publish --tag "$NPM_TAG" "--@soulcraft:registry=https://registry.npmjs.org/" # Brainy is the only PUBLIC @soulcraft package — verify visibility after every publish. -npm access get status @soulcraft/brainy || true -echo -e "${GREEN}✅ Published to npm${NC}\n" +npm access get status @soulcraft/brainy "--@soulcraft:registry=https://registry.npmjs.org/" || true +echo -e "${GREEN}✅ Published to npmjs${NC}\n" -# Step 11: Create GitHub release -echo -e "${BLUE}🔟 Creating GitHub release...${NC}" -if [ "$PRERELEASE" = true ]; then - gh release create "v${NEW_VERSION}" --generate-notes --prerelease +# Step 11: Release object on the forge (presentational — the tag, CHANGELOG, +# and RELEASES.md are the record; this just gives the forge UI a release page). +echo -e "${BLUE}🔟 Creating forge release...${NC}" +if [ -n "${FORGEJO_RELEASE_TOKEN:-}" ]; then + if curl -sf -X POST "https://source.soulcraft.com/api/v1/repos/soulcraft/brainy/releases" \ + -H "Authorization: token ${FORGEJO_RELEASE_TOKEN}" -H "Content-Type: application/json" \ + -d "{\"tag_name\":\"v${NEW_VERSION}\",\"name\":\"v${NEW_VERSION}\",\"prerelease\":${PRERELEASE}}" >/dev/null; then + echo -e "${GREEN}✅ Forge release created${NC}\n" + else + echo -e "${RED}⚠️ Forge release API call failed — tag + CHANGELOG remain the record; create the release page via the forge UI if wanted${NC}\n" + fi else - gh release create "v${NEW_VERSION}" --generate-notes + echo -e "${RED}⚠️ FORGEJO_RELEASE_TOKEN unset — no release page created; tag + CHANGELOG remain the record${NC}\n" fi -echo -e "${GREEN}✅ GitHub release created${NC}\n" # Step 12: Push public docs to the soulcraft.com docs ingest door # (VENUE-DOCS-RELEASE-PUSH). Skips with a loud warning when @@ -229,4 +246,4 @@ echo -e "${GREEN}🎉 Release ${NEW_VERSION} complete!${NC}" echo -e "${GREEN}━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━${NC}" echo "" echo -e "📦 npm: ${BLUE}https://www.npmjs.com/package/@soulcraft/brainy/v/${NEW_VERSION}${NC}" -echo -e "🐙 GitHub: ${BLUE}https://github.com/soulcraftlabs/brainy/releases/tag/v${NEW_VERSION}${NC}" +echo -e "🏠 Forge: ${BLUE}https://source.soulcraft.com/soulcraft/brainy/releases/tag/v${NEW_VERSION}${NC}" diff --git a/src/brainy.ts b/src/brainy.ts index 37b6e491..007cdb32 100644 --- a/src/brainy.ts +++ b/src/brainy.ts @@ -65,7 +65,8 @@ import type { OpaqueIdSet, AtGenerationVectors, VectorIndexProvider, - GraphIndexProvider + GraphIndexProvider, + ProviderMaintenanceDebt } from './plugin.js' import type { BrainyPlugin, @@ -424,6 +425,30 @@ export interface WarmReport { totalDurationMs: number } +/** + * @description Result of {@link Brainy.maintenanceDebt}: one outcome per + * index surface, mirroring {@link WarmReport}'s shape. + * - `'reported'` — the active provider for this surface implements + * `maintenanceDebt?()` and its {@link ProviderMaintenanceDebt} payload is + * attached verbatim under `debt`. + * - `'unavailable'` — the active provider does not implement the hook, so + * nothing is known; brainy never estimates or infers a payload on its + * behalf. + */ +export type MaintenanceDebtOutcome = 'reported' | 'unavailable' + +/** + * @description Per-surface result of {@link Brainy.maintenanceDebt}. Brainy + * performs no thresholding, polling, or estimation over this data — it is a + * pure passthrough of each active provider's own self-report (the provider + * owns the numbers; the operator owns the policy). + */ +export interface MaintenanceDebtReport { + vector: { outcome: MaintenanceDebtOutcome; debt?: ProviderMaintenanceDebt } + metadata: { outcome: MaintenanceDebtOutcome; debt?: ProviderMaintenanceDebt } + graph: { outcome: MaintenanceDebtOutcome; debt?: ProviderMaintenanceDebt } +} + /** * How long a failed aggregation-backfill walk suppresses fresh walk attempts. * Within the window, queries rethrow the recorded failure instantly (loud, @@ -14313,9 +14338,16 @@ export class Brainy implements BrainyInterface { * *some* backing storage as a side effect but is reported honestly as * `'probed'`, never `'warmed'`. An empty index or unknown dimension has * nothing to probe (`'unavailable'`). - * - **Metadata**: full hydration — every persisted field's sparse index is - * loaded from storage (`MetadataIndexManager.hydrateAll()`), not just the - * heuristic common-fields subset `init()` warms. + * - **Metadata**: calls the provider's own `warm?()` when the active + * `'metadataIndex'` provider implements it (`'warmed'`) — the seam a + * native metadata provider lights up so it is not duck-typed against the + * JS manager's method. Otherwise falls back to full hydration on the + * built-in JS manager — every persisted field's sparse index is loaded + * from storage (`MetadataIndexManager.hydrateAll()`), not just the + * heuristic common-fields subset `init()` warms — and reports `'warmed'`. + * Neither seam present → `'unavailable'` (honest: `init()` is never used + * as a substitute here, since a native provider's `init()` may be a + * cheap verify rather than a real warm). * - **Graph**: calls the provider's own `warm?()` when the active graph * provider implements it; otherwise re-runs its existing eager cold-load * `init()` seam (idempotent — the JS adjacency index's `init()` already @@ -14371,12 +14403,23 @@ export class Brainy implements BrainyInterface { // --- Metadata -------------------------------------------------------- const metadataStart = Date.now() let metadataOutcome: WarmOutcome + const metadataProvider = this.metadataIndex as unknown as MetadataIndexProvider const metadataWithHydrate = this.metadataIndex as unknown as { hydrateAll?: () => Promise } - if (typeof metadataWithHydrate.hydrateAll === 'function') { + if (typeof metadataProvider.warm === 'function') { + // Active provider (e.g. a native metadata index) declares its own warm + // seam — route through it FIRST so a native provider's warmth is + // reported honestly instead of being duck-typed against the JS + // manager's hydrateAll(), which a native provider does not implement. + await metadataProvider.warm() + metadataOutcome = 'warmed' + } else if (typeof metadataWithHydrate.hydrateAll === 'function') { + // Built-in JS manager path — full sparse-index hydration. await metadataWithHydrate.hydrateAll() metadataOutcome = 'warmed' } else { - // No hydration seam on this metadata provider — nothing to run. + // No hydration seam on this metadata provider — nothing to run. (No + // init() fallback here: init() on a native provider may be a cheap + // verify, and reporting that as warmth would lie.) metadataOutcome = 'unavailable' } const metadataDurationMs = Date.now() - metadataStart @@ -14410,6 +14453,63 @@ export class Brainy implements BrainyInterface { } } + /** + * Read each index surface's self-reported outstanding maintenance work — + * the observability seam so an operator sees a grind coming (rising + * pending bytes/items, a stalled background pass) instead of discovering + * it as a CPU storm or a transaction blowing its budget mid-flight (see + * {@link TransactionTimeoutError}). + * + * PURE PASSTHROUGH: for each of vector/metadata/graph, this calls ONLY the + * ACTIVE provider's own `maintenanceDebt?()` hook (the same per-surface + * provider resolution {@link Brainy.warm} uses) and reports its + * {@link ProviderMaintenanceDebt} payload verbatim. There is no JS-side + * fallback computation, no threshold evaluation, and no polling — brainy + * surfaces the truth the provider measured; the provider owns the numbers + * and the operator owns the policy (what threshold matters, what action to + * take). A surface whose active provider does not implement the hook + * reports `'unavailable'` — never a guessed or zeroed payload. + * + * @returns A {@link MaintenanceDebtReport}: per-surface outcome + payload. + * @example + * ```typescript + * const debt = await brain.maintenanceDebt() + * if (debt.metadata.outcome === 'reported' && debt.metadata.debt?.pendingBytes) { + * console.log('metadata pending bytes:', debt.metadata.debt.pendingBytes) + * } + * ``` + */ + async maintenanceDebt(): Promise { + await this.ensureInitialized({ needs: ['vector', 'metadata', 'graph'] }) + + // --- Vector --------------------------------------------------------- + const vectorProvider = this.index as VectorIndexProvider & { + maintenanceDebt?: () => Promise + } + const vector = + typeof vectorProvider.maintenanceDebt === 'function' + ? { outcome: 'reported' as const, debt: await vectorProvider.maintenanceDebt() } + : { outcome: 'unavailable' as const } + + // --- Metadata -------------------------------------------------------- + const metadataProvider = this.metadataIndex as unknown as MetadataIndexProvider + const metadata = + typeof metadataProvider.maintenanceDebt === 'function' + ? { outcome: 'reported' as const, debt: await metadataProvider.maintenanceDebt() } + : { outcome: 'unavailable' as const } + + // --- Graph ------------------------------------------------------------- + const graphProvider = this.graphIndex as GraphIndexProvider & { + maintenanceDebt?: () => Promise + } + const graph = + typeof graphProvider.maintenanceDebt === 'function' + ? { outcome: 'reported' as const, debt: await graphProvider.maintenanceDebt() } + : { outcome: 'unavailable' as const } + + return { vector, metadata, graph } + } + /** * Explicitly warm up the embedding engine * diff --git a/src/db/types.ts b/src/db/types.ts index 4c8a4957..a7f86a89 100644 --- a/src/db/types.ts +++ b/src/db/types.ts @@ -121,9 +121,13 @@ export interface TransactOptions { * with the batch: `max(30 000, opCount × 2 000)` — production imports on * network-attached disks measure ~2 s per operation, so a flat 30 s budget * silently capped honest bulk work at ~15 operations. A tripped budget - * rolls the whole batch back and throws a retryable - * `TransactionTimeoutError` naming the operation it stopped at, the batch - * size, and the elapsed/budget times. + * rolls the whole batch back and throws a `TransactionTimeoutError` naming + * the operation it stopped at, the batch size, and the elapsed/budget + * times. That error is retryable-with-latch, never hot-retry: its + * `retryable` field says a later attempt may succeed, its + * `hotRetryUnsafe` field says an immediate identical retry re-pays the + * full cost that just timed out — callers must latch and back off, never + * loop. */ timeoutMs?: number } diff --git a/src/index.ts b/src/index.ts index 00adc191..e60739f7 100644 --- a/src/index.ts +++ b/src/index.ts @@ -31,6 +31,11 @@ export type { DiagnosticsResult } from './brainy.js' // brain.warm() — eager index/storage readiness report (per-surface honest // outcome + timing). See the WarmReport JSDoc in brainy.ts. export type { WarmReport, WarmOutcome } from './brainy.js' +// brain.maintenanceDebt() — per-surface passthrough of each active +// provider's self-reported background maintenance debt. See the +// MaintenanceDebtReport JSDoc in brainy.ts and ProviderMaintenanceDebt in +// plugin.ts for the measure-only-what-you-track contract. +export type { MaintenanceDebtReport, MaintenanceDebtOutcome } from './brainy.js' export type { GraphAuditReport, GraphAuditDiscrepancy @@ -227,6 +232,11 @@ export type { FamilyStamp, StampMembers, StampVerdict } from './db/familyStamp.j export { isVersionedIndexProvider } from './plugin.js' export type { VersionedIndexProvider } from './plugin.js' export type { ProviderInvariantReport, InvariantResult, InvariantHeal } from './plugin.js' +// Optional provider self-report of outstanding background maintenance work +// (compaction, deferred writes, etc.) — the payload type for +// brain.maintenanceDebt(). See the measure-only-what-you-track contract on +// ProviderMaintenanceDebt in plugin.ts. +export type { ProviderMaintenanceDebt } from './plugin.js' // Optional native graph-acceleration engine (cor 3.0) — the published provider // contract + its columnar wire types. Brainy feature-detects an implementation // and falls back to its pure-TS adjacency when absent. diff --git a/src/plugin.ts b/src/plugin.ts index 02639955..ce973386 100644 --- a/src/plugin.ts +++ b/src/plugin.ts @@ -171,6 +171,37 @@ export interface ProviderInvariantReport { durationMs: number } +/** + * @description A provider's self-report of its own outstanding background + * maintenance work (compaction, deferred writes, a build-new→verify→swap in + * flight, etc.) — the observability seam so an operator sees a grind coming + * (rising pending bytes/items, a stalled pass) instead of discovering it as a + * CPU storm or a timeout under transaction budget pressure. Every field is + * OPTIONAL and every field is a MEASUREMENT: a provider reports ONLY what it + * actually tracks, never an estimate dressed up as a fact. Absence of the + * {@link VectorIndexProvider.maintenanceDebt} / + * {@link GraphIndexProvider.maintenanceDebt} / + * {@link MetadataIndexProvider.maintenanceDebt} hook itself means the + * provider does not track debt at all — brainy reports that surface + * `'unavailable'` rather than inventing zeros. Brainy performs NO threshold + * checks, NO polling, and NO JS-side estimation over this payload — it is a + * pure passthrough via {@link Brainy.maintenanceDebt}; the provider owns the + * numbers and the operator owns the policy (what threshold matters, what to + * do about it). + */ +export interface ProviderMaintenanceDebt { + /** Bytes of outstanding/unmerged work, if the provider measures it (e.g. unflushed writes, unmerged segments). */ + pendingBytes?: number + /** Count of outstanding items (records, segments, nodes) awaiting the provider's background pass. */ + pendingItems?: number + /** Epoch millis when the provider's last maintenance pass finished, if it tracks one. */ + lastPassCompletedAt?: number + /** How the last pass ended, if the provider tracks pass outcomes. */ + lastPassOutcome?: 'completed' | 'partial' | 'failed' + /** `true` if the provider's own measurements show debt trending down (making progress); `false` if flat or growing; omitted if the provider can't tell. */ + converging?: boolean +} + /** * The `'metadataIndex'` provider — a drop-in for `MetadataIndexManager`. * Brainy calls this surface via `this.metadataIndex.*` (see `brainy.ts`) and @@ -181,6 +212,34 @@ export interface MetadataIndexProvider { flush(): Promise rebuild(): Promise + /** + * @description OPTIONAL. Eagerly load/fault-in backing storage (e.g. mmap + * pretouch, full sparse-index hydration) so first queries run at + * steady-state cost. Optional; absence means the provider demand-loads. + * Mirrors {@link GraphIndexProvider.warm} / the vector provider's `warm?()` + * (`src/plugin.ts` VectorIndexProvider). Distinct from `init()`: `init` is + * required and runs once automatically during brain startup; `warm` is a + * separate, explicit step a caller opts into via `brain.warm()` (or + * `warmOnOpen`) to pre-pay demand-load cost `init` left lazy. Idempotent — + * calling it more than once must be safe and cheap on a brain that is + * already warm. A provider that already loads everything eagerly in + * `init()` may implement `warm` as a no-op or omit it — `brain.warm()` + * falls back to the built-in JS manager's `hydrateAll()` duck-type when + * absent, and to an honest `'unavailable'` when neither exists. + */ + warm?(): Promise + + /** + * @description OPTIONAL self-reported {@link ProviderMaintenanceDebt} — + * the observability seam so an operator sees outstanding background + * maintenance work (e.g. unmerged postings) BEFORE it grinds a transaction + * into a budget-busting op. Absence means this provider does not track + * debt; `brain.maintenanceDebt()` reports this surface `'unavailable'` + * rather than guessing. See {@link ProviderMaintenanceDebt} for the + * measure-only-what-you-track contract. + */ + maintenanceDebt?(): Promise + /** * @description OPTIONAL honest durability signal (readiness contract, * mirrors `isReady?()` on the graph and vector providers). `true` ⇔ the @@ -395,6 +454,18 @@ export interface GraphIndexProvider { */ warm?(): Promise + /** + * @description OPTIONAL self-reported {@link ProviderMaintenanceDebt} — + * the observability seam so an operator sees outstanding background + * maintenance work (e.g. a build-new→verify→swap in flight, unmerged + * adjacency segments) BEFORE it grinds a transaction into a + * budget-busting op. Absence means this provider does not track debt; + * `brain.maintenanceDebt()` reports this surface `'unavailable'` rather + * than guessing. See {@link ProviderMaintenanceDebt} for the + * measure-only-what-you-track contract. + */ + maintenanceDebt?(): Promise + /** * @description OPTIONAL. A native provider returns true from the moment its * `init()` detects a large epoch-drift until its background @@ -1057,6 +1128,18 @@ export interface VectorIndexProvider { */ warm?(): Promise + /** + * @description OPTIONAL self-reported {@link ProviderMaintenanceDebt} — + * the observability seam so an operator sees outstanding background + * maintenance work (e.g. unflushed writes, a pending rebuild) BEFORE it + * grinds a transaction into a budget-busting op. Absence means this + * provider does not track debt; `brain.maintenanceDebt()` reports this + * surface `'unavailable'` rather than guessing. See + * {@link ProviderMaintenanceDebt} for the measure-only-what-you-track + * contract. + */ + maintenanceDebt?(): Promise + /** * @description OPTIONAL honest durability signal (readiness contract, * mirrors {@link GraphIndexProvider.isReady}). `true` ⇔ the persisted diff --git a/src/transaction/Transaction.ts b/src/transaction/Transaction.ts index 79b53006..ede65e83 100644 --- a/src/transaction/Transaction.ts +++ b/src/transaction/Transaction.ts @@ -62,9 +62,10 @@ const DEFAULT_BUDGET_FLOOR_MS = 30_000 * NEXT operation may start (see {@link Transaction.execute}), never whether * already-completed work is rolled back after the fact. A trip mid-batch * still rolls back every operation applied so far, atomically, and throws a - * retryable, fully-labeled TransactionTimeoutError — that zero-loss guarantee - * doesn't change; only the point at which the clock stops mattering does (at - * the last operation, not one check later). + * fully-labeled `TransactionTimeoutError` — retryable-with-latch, never + * hot-retry (see its `retryable` and `hotRetryUnsafe` fields) — that + * zero-loss guarantee doesn't change; only the point at which the clock + * stops mattering does (at the last operation, not one check later). * * @param opCount - Number of operations in the batch. * @param override - A full override for this call; wins over everything else. diff --git a/src/transaction/TransactionManager.ts b/src/transaction/TransactionManager.ts index 0f13b6a3..5abf48ba 100644 --- a/src/transaction/TransactionManager.ts +++ b/src/transaction/TransactionManager.ts @@ -19,7 +19,6 @@ import { Transaction } from './Transaction.js' import { TransactionFunction, - TransactionResult, TransactionOptions } from './types.js' import { TransactionError } from './errors.js' @@ -105,34 +104,6 @@ export class TransactionManager { } } - /** - * Execute a transaction and return detailed result - */ - async executeTransactionWithResult( - fn: TransactionFunction, - options?: TransactionOptions - ): Promise> { - const startTime = Date.now() - const transaction = new Transaction(options) - - try { - const value = await fn(transaction) - await transaction.execute() - - const executionTimeMs = Date.now() - startTime - - return { - value, - operationCount: transaction.getOperationCount(), - executionTimeMs - } - - } catch (error) { - // Transaction failed - throw error - } - } - /** * Get transaction statistics */ diff --git a/src/transaction/errors.ts b/src/transaction/errors.ts index c270d0ed..d8382a4b 100644 --- a/src/transaction/errors.ts +++ b/src/transaction/errors.ts @@ -73,14 +73,47 @@ export class InvalidTransactionStateError extends TransactionError { /** * Error for transaction timeout + * + * Machine-readable no-hot-retry contract: {@link retryable} and + * {@link hotRetryUnsafe} are both always `true` on this class — they exist + * so a caller can branch on the *shape* of the error instead of parsing + * message text. Read them together: the operation may eventually succeed, + * but never by looping on it immediately. + * + * `context` (inherited from {@link TransactionError}) carries the caller's + * backoff inputs — see the field docs below. */ export class TransactionTimeoutError extends TransactionError { + /** + * The failed operation MAY succeed on a later attempt — once the + * underlying slowness resolves (e.g. a cold page cache warms up) or the + * budget is deliberately raised (`transactionBudgetFloorMs`, or a larger + * `timeoutMs` override on the batch). This is a statement about eventual + * retryability, not a license to retry now — see {@link hotRetryUnsafe}. + */ + public readonly retryable = true + + /** + * An immediate, identical retry re-pays the FULL cost of the work that + * just timed out — it does not resume partway. Looping on this error + * (hot-retrying) repeats that full cost every attempt and can cascade + * into a CPU/resource storm on the caller's side. Callers MUST latch: on + * this error, record `{ at: Date.now(), error }`, surface one loud + * failure to their own caller, and hold a cooldown window before any + * re-attempt (clearing the latch only on success). Never retry this error + * in a tight loop. + */ + public readonly hotRetryUnsafe = true + constructor( timeoutMs: number, operationIndex: number, telemetry?: { + /** Milliseconds elapsed in the transaction when the budget tripped. */ elapsedMs?: number + /** Total number of operations in the batch that timed out. */ totalOperations?: number + /** Name of the operation the batch was about to start when it tripped, if named. */ operationName?: string } ) { @@ -93,8 +126,16 @@ export class TransactionTimeoutError extends TransactionError { telemetry?.elapsedMs !== undefined ? `${telemetry.elapsedMs}ms elapsed, ` : '' super( `Transaction timed out at operation ${progress}${name} — ${elapsed}budget ${timeoutMs}ms. ` + - `The batch rolled back atomically; retry with a higher timeoutMs or a smaller batch.`, - { timeoutMs, operationIndex, ...telemetry } + `The batch rolled back atomically; retryable after the underlying slowness resolves or ` + + `the budget is raised, but hot-retry-unsafe — latch and back off, never loop.`, + { + // Caller backoff inputs — all present on every instance: + /** Configured budget (ms) that was exceeded. */ + timeoutMs, + /** Index of the operation the batch was about to start when it tripped. */ + operationIndex, + ...telemetry + } ) this.name = 'TransactionTimeoutError' } diff --git a/src/transaction/types.ts b/src/transaction/types.ts index 9a3a2eaa..6cbc56ca 100644 --- a/src/transaction/types.ts +++ b/src/transaction/types.ts @@ -66,26 +66,6 @@ export interface TransactionContext { */ export type TransactionFunction = (ctx: TransactionContext) => Promise -/** - * Transaction execution result - */ -export interface TransactionResult { - /** - * Result value from user function - */ - value: T - - /** - * Number of operations executed - */ - operationCount: number - - /** - * Execution time in milliseconds - */ - executionTimeMs: number -} - /** * Transaction execution options */ diff --git a/src/types/brainy.types.ts b/src/types/brainy.types.ts index 8bede8d3..b5f286ed 100644 --- a/src/types/brainy.types.ts +++ b/src/types/brainy.types.ts @@ -1658,8 +1658,10 @@ export interface BrainyConfig { * **start** — never whether already-completed work gets rolled back after * the fact (a single-op write can never time out post-hoc: it either runs * or it commits). A trip mid-batch still rolls back every applied operation - * atomically and throws a retryable `TransactionTimeoutError`; only the - * floor of the formula is configurable here. + * atomically and throws a `TransactionTimeoutError` that is + * retryable-with-latch, never hot-retry (see its `retryable` and + * `hotRetryUnsafe` fields); only the floor of the formula is configurable + * here. * * Raise this when a cold store's first writes after a restart legitimately * take longer than 30s per operation (e.g. page-cache-cold canonical writes diff --git a/tests/transaction/TransactionManager.unit.test.ts b/tests/transaction/TransactionManager.unit.test.ts index 86b1692c..29e7f7ae 100644 --- a/tests/transaction/TransactionManager.unit.test.ts +++ b/tests/transaction/TransactionManager.unit.test.ts @@ -5,7 +5,6 @@ * - High-level transaction API * - Statistics tracking * - Error handling - * - Result wrapping */ import { describe, it, expect, beforeEach } from 'vitest' @@ -84,42 +83,6 @@ describe('TransactionManager', () => { }) }) - describe('executeTransactionWithResult', () => { - it('should return detailed result', async () => { - const result = await manager.executeTransactionWithResult(async (tx) => { - tx.addOperation({ - execute: async () => { - await new Promise(resolve => setTimeout(resolve, 1)) - return async () => {} - } - }) - tx.addOperation({ execute: async () => undefined }) - return 'success' - }) - - expect(result.value).toBe('success') - expect(result.operationCount).toBe(2) - expect(result.executionTimeMs).toBeGreaterThanOrEqual(0) - }) - - it('should measure execution time', async () => { - const result = await manager.executeTransactionWithResult(async (tx) => { - tx.addOperation({ - execute: async () => { - await new Promise(resolve => setTimeout(resolve, 25)) - return async () => {} - } - }) - return 'done' - }) - - // Timer coalescing can fire a setTimeout up to a few ms EARLY under - // load, so assert well below the sleep — this tests that time is - // MEASURED, not the OS timer's precision. - expect(result.executionTimeMs).toBeGreaterThanOrEqual(20) - }) - }) - describe('Statistics Tracking', () => { it('should track total transactions', async () => { await manager.executeTransaction(async (tx) => { diff --git a/tests/unit/brainy/maintenance-debt.test.ts b/tests/unit/brainy/maintenance-debt.test.ts new file mode 100644 index 00000000..4026d655 --- /dev/null +++ b/tests/unit/brainy/maintenance-debt.test.ts @@ -0,0 +1,122 @@ +/** + * @module tests/unit/brainy/maintenance-debt + * @description Coverage for `brain.maintenanceDebt()` (8.10.1) — the + * observability seam so an operator sees a provider's outstanding background + * maintenance work (compaction, deferred writes, a build-new→verify→swap in + * flight, ...) BEFORE it grinds a transaction into a budget-busting op, the + * same failure class documented on `TransactionTimeoutError` + * (src/transaction/errors.ts). Sibling to tests/unit/brainy/warm.test.ts, + * which establishes this file's technique: shape the probe points brain.ts + * reads (`typeof provider.maintenanceDebt === 'function'`) directly on the + * REAL, live provider instances rather than hand-rolling full fakes for the + * larger `MetadataIndexProvider` / `GraphIndexProvider` interfaces. + * + * `brain.maintenanceDebt()` is a PURE PASSTHROUGH: no thresholds, no + * polling, no JS-side estimation — these tests pin exactly that by asserting + * the returned payload is the provider's object, verbatim. + */ +import { describe, it, expect } from 'vitest' +import { Brainy } from '../../../src/brainy.js' +import { NounType } from '../../../src/types/graphTypes.js' +import type { ProviderMaintenanceDebt } from '../../../src/plugin.js' + +// Brainy's ValidationConfig fixes vectors at exactly 384 dimensions +// (src/utils/paramValidation.ts) — match it so add() doesn't reject test data. +const DIM = 384 +const V = (seed = 1): number[] => Array.from({ length: DIM }, (_, i) => Math.sin(seed + i)) + +async function freshBrain(): Promise> { + const brain = new Brainy({ + requireSubtype: false, + storage: { type: 'memory' }, + silent: true + }) + await brain.init() + await brain.add({ data: 'a', type: NounType.Thing, vector: V(1) }) + return brain +} + +describe('brain.maintenanceDebt()', () => { + it('reports "unavailable" for every surface when no active provider implements maintenanceDebt() (the built-in JS stack today)', async () => { + const brain = await freshBrain() + + const report = await brain.maintenanceDebt() + + expect(report.vector.outcome).toBe('unavailable') + expect(report.vector.debt).toBeUndefined() + expect(report.metadata.outcome).toBe('unavailable') + expect(report.metadata.debt).toBeUndefined() + expect(report.graph.outcome).toBe('unavailable') + expect(report.graph.debt).toBeUndefined() + + await brain.close() + }) + + it('reports "reported" + the exact payload when the active provider implements maintenanceDebt() (verbatim passthrough, no thresholding)', async () => { + const brain = await freshBrain() + + const vectorDebt: ProviderMaintenanceDebt = { + pendingBytes: 4_096, + pendingItems: 12, + lastPassCompletedAt: 1_700_000_000_000, + lastPassOutcome: 'completed', + converging: true + } + ;(brain as any).index.maintenanceDebt = async () => vectorDebt + + const report = await brain.maintenanceDebt() + + expect(report.vector.outcome).toBe('reported') + // Verbatim passthrough — the exact object, not a re-derived copy. + expect(report.vector.debt).toBe(vectorDebt) + // Untouched surfaces stay honestly 'unavailable'. + expect(report.metadata.outcome).toBe('unavailable') + expect(report.graph.outcome).toBe('unavailable') + + await brain.close() + }) + + it('mixed surfaces: each surface\'s outcome depends ONLY on its OWN active provider — one surface reporting never leaks into another', async () => { + const brain = await freshBrain() + + const metadataDebt: ProviderMaintenanceDebt = { + pendingItems: 3, + lastPassOutcome: 'partial', + converging: false + } + const graphDebt: ProviderMaintenanceDebt = { + pendingBytes: 0, + converging: true + } + ;(brain as any).metadataIndex.maintenanceDebt = async () => metadataDebt + ;(brain as any).graphIndex.maintenanceDebt = async () => graphDebt + // Vector is deliberately left unpatched. + + const report = await brain.maintenanceDebt() + + expect(report.vector.outcome).toBe('unavailable') + expect(report.vector.debt).toBeUndefined() + + expect(report.metadata.outcome).toBe('reported') + expect(report.metadata.debt).toBe(metadataDebt) + + expect(report.graph.outcome).toBe('reported') + expect(report.graph.debt).toBe(graphDebt) + + await brain.close() + }) + + it('an empty ProviderMaintenanceDebt object (every field omitted) is still honestly "reported" — presence of the hook, not the payload\'s richness, drives the outcome', async () => { + const brain = await freshBrain() + + const emptyDebt: ProviderMaintenanceDebt = {} + ;(brain as any).graphIndex.maintenanceDebt = async () => emptyDebt + + const report = await brain.maintenanceDebt() + + expect(report.graph.outcome).toBe('reported') + expect(report.graph.debt).toEqual({}) + + await brain.close() + }) +}) diff --git a/tests/unit/brainy/warm.test.ts b/tests/unit/brainy/warm.test.ts index ce213696..8a0bb4da 100644 --- a/tests/unit/brainy/warm.test.ts +++ b/tests/unit/brainy/warm.test.ts @@ -307,4 +307,92 @@ describe('brain.warm()', () => { expect(report.totalDurationMs).toBeGreaterThanOrEqual(0) await brain.close() }) + + // --- Metadata leg routes through the ACTIVE provider (8.10.1) ----------- + // + // `MetadataIndexProvider` is a ~50-method interface (src/plugin.ts) — far + // too large to hand-write a compliant fake class the way `FakeVectorProvider` + // fakes the ~8-method `VectorIndexProvider` above. Test (c) already + // establishes this file's pattern for the metadata leg: exercise the REAL + // `MetadataIndexManager` instance and shape just the probe points brain.ts + // reads (`typeof provider.warm === 'function'` / + // `typeof provider.hydrateAll === 'function'`) directly on that instance. + // Shadowing an own property on the live object stands in for "a different + // provider implementation" without needing a hand-rolled full fake — the + // rest of the real manager (used by add()/init() above) is untouched. + describe('metadata leg — warm() routes through the active provider', () => { + it('(f) calls the ACTIVE metadata provider\'s warm() when present and reports "warmed", never falling back to hydrateAll', async () => { + const brain = new Brainy({ + requireSubtype: false, + storage: { type: 'memory' }, + silent: true + }) + await brain.init() + await brain.add({ data: 'a', type: NounType.Thing, vector: V(1) }) + + const metadataIndex = (brain as any).metadataIndex + let warmCalls = 0 + let hydrateAllCalls = 0 + const origHydrateAll = metadataIndex.hydrateAll.bind(metadataIndex) + metadataIndex.hydrateAll = async (...args: unknown[]) => { + hydrateAllCalls++ + return origHydrateAll(...args) + } + // Simulates a native metadata provider declaring the optional `warm()` + // hook added to `MetadataIndexProvider` (src/plugin.ts) in 8.10.1. + metadataIndex.warm = async () => { + warmCalls++ + } + + const report = await brain.warm() + + expect(warmCalls).toBe(1) + expect(hydrateAllCalls).toBe(0) // warm() ran — no hydrateAll fallback + expect(report.metadata.outcome).toBe('warmed') + await brain.close() + }) + + it('reports "unavailable" when the active metadata provider implements neither warm() nor hydrateAll() (the honest branch a native provider without either hook must hit)', async () => { + const brain = new Brainy({ + requireSubtype: false, + storage: { type: 'memory' }, + silent: true + }) + await brain.init() + await brain.add({ data: 'a', type: NounType.Thing, vector: V(1) }) + + const metadataIndex = (brain as any).metadataIndex + // Shadow away BOTH optional hooks — models a genuinely native provider + // that (unlike the built-in JS manager) offers neither seam. This must + // never fall back to calling init() as a stand-in for warmth. + metadataIndex.warm = undefined + metadataIndex.hydrateAll = undefined + + const report = await brain.warm() + + expect(report.metadata.outcome).toBe('unavailable') + await brain.close() + }) + + it('the built-in JS manager (no warm()) still reports "warmed" via its existing hydrateAll() duck-type — unchanged by the new provider hook', async () => { + const brain = new Brainy({ + requireSubtype: false, + storage: { type: 'memory' }, + silent: true + }) + await brain.init() + await brain.add({ data: 'a', type: NounType.Thing, vector: V(1) }) + + // No patching at all — the default built-in MetadataIndexManager has + // hydrateAll() but no warm(), exactly as it did before this change. + const metadataIndex = (brain as any).metadataIndex + expect(typeof metadataIndex.warm).not.toBe('function') + expect(typeof metadataIndex.hydrateAll).toBe('function') + + const report = await brain.warm() + + expect(report.metadata.outcome).toBe('warmed') + await brain.close() + }) + }) }) diff --git a/tests/unit/transaction/timeout-never-internally-retried.test.ts b/tests/unit/transaction/timeout-never-internally-retried.test.ts new file mode 100644 index 00000000..a43a81a8 --- /dev/null +++ b/tests/unit/transaction/timeout-never-internally-retried.test.ts @@ -0,0 +1,141 @@ +/** + * @module tests/unit/transaction/timeout-never-internally-retried + * @description Regression pin for the no-hot-retry contract (8.10.1). + * + * A production incident: a native-provider op ground 38-40s inside a + * transaction, blew the ~32s budget, was rolled back, and a CONSUMER pipeline + * hot-retried the identical operation into a 6-minute 100%-CPU storm. + * Investigation established brainy itself never auto-retries a + * `TransactionTimeoutError` — the storm was entirely the consumer's hot-retry + * loop, driven by a "retryable" doc-prose claim with no machine-readable + * contract. This file pins the brainy-side half of that story so it can never + * regress silently: + * + * (i) the underlying engine (`TransactionManager.executeTransaction()` → + * `Transaction.execute()`) — the exact machinery every single-record + * write (`add`/`update`/`remove`/...) drives via + * `Brainy.persistSingleOp()` — never internally re-executes a timed-out + * operation, and the error it surfaces carries `retryable === true` and + * `hotRetryUnsafe === true` (see `src/transaction/errors.ts`). + * + * Constructed directly (mirrors the existing + * `tests/unit/transaction/timeout-rollback.test.ts` pattern) rather than + * through a real `brain.add()` call: `transactTimeoutBudget()` floors + * every single-op write's budget at `opCount * 2000`ms with NO override + * seam (`transactionBudgetFloorMs` only RAISES that floor — it cannot + * lower it below the per-op-count term), so getting a real `add()` to + * time out requires a multi-second sleep. The engine-level + * `options.timeout` override used here is the exact same + * `TransactionManager`/`Transaction` code `persistSingleOp` calls — + * pinning it here pins add()'s guarantee without paying that wall-clock + * cost. + * + * (ii) `Brainy.add()`'s upsert-race retry loop (src/brainy.ts, + * `MAX_UPSERT_ATTEMPTS = 10`) — proving the loop's `catch` treats a + * `TransactionTimeoutError` as terminal (immediate rethrow) rather than + * the `InsertPreconditionExistsSignal` it retries on, so a mid-flight + * timeout can never be silently swallowed and re-attempted up to 10 + * times. + */ +import { describe, it, expect } from 'vitest' +import { TransactionManager } from '../../../src/transaction/TransactionManager.js' +import type { Operation, RollbackAction } from '../../../src/transaction/types.js' +import { TransactionTimeoutError } from '../../../src/transaction/errors.js' +import { Brainy } from '../../../src/brainy.js' +import { NounType } from '../../../src/types/graphTypes.js' + +const sleep = (ms: number) => new Promise((r) => setTimeout(r, ms)) + +// Brainy's ValidationConfig fixes vectors at exactly 384 dimensions +// (src/utils/paramValidation.ts) — match it so add() doesn't reject test data. +const DIM = 384 +const V = (): number[] => Array(DIM).fill(0.1) + +/** An operation that counts every `execute()` invocation — the re-drive detector. */ +function countingOp(opts: { delayMs?: number; name: string }): Operation & { calls: number } { + const op = { + name: opts.name, + calls: 0, + async execute(): Promise { + op.calls++ + if (opts.delayMs) await sleep(opts.delayMs) + return async () => {} + } + } + return op +} + +describe('transaction timeouts are never internally re-driven (8.10.1 no-hot-retry contract)', () => { + it('(i) the single-op engine (TransactionManager.executeTransaction / Transaction.execute — what add() drives via persistSingleOp) runs the overrun operation EXACTLY once and surfaces ONE retryable+hotRetryUnsafe error', async () => { + const manager = new TransactionManager() + const op0 = countingOp({ name: 'op0-overruns-budget', delayMs: 30 }) + const op1 = countingOp({ name: 'op1-must-never-start' }) + + let caught: unknown + try { + await manager.executeTransaction( + async (tx) => { + tx.addOperation(op0) + tx.addOperation(op1) + }, + // Tiny explicit override — the same override seam `transact()` + // exposes as `options.timeoutMs`; wins outright over the + // opCount*2000 floor that gates every real single-op write + // (transactTimeoutBudget()'s override semantics). + { timeout: 5 } + ) + } catch (err) { + caught = err + } + + expect(caught).toBeInstanceOf(TransactionTimeoutError) + const err = caught as TransactionTimeoutError + // The machine-readable contract callers branch on instead of parsing + // message text (src/transaction/errors.ts). + expect(err.retryable).toBe(true) + expect(err.hotRetryUnsafe).toBe(true) + + // The re-drive assertion: op0 (the one that overran) executed EXACTLY + // once — nothing inside TransactionManager/Transaction looped back and + // re-ran it — and op1 never started at all (the budget gate stopped it + // before it began, per Transaction.execute()'s per-operation loop). + expect(op0.calls).toBe(1) + expect(op1.calls).toBe(0) + }) + + it('(ii) add()\'s upsert-race retry loop (MAX_UPSERT_ATTEMPTS=10) exits on the FIRST TransactionTimeoutError — attempt counter stays at 1, never mistaken for the lost-insert-race signal it retries on', async () => { + const brain = new Brainy({ + requireSubtype: false, + storage: { type: 'memory' }, + silent: true + }) + await brain.init() + + let persistSingleOpCalls = 0 + const timeoutError = new TransactionTimeoutError(5, 1, { + elapsedMs: 6, + totalOperations: 2, + operationName: 'SaveNounMetadata' + }) + // Stub the private commit seam add() drives (persistSingleOp) to throw + // the exact error type the real engine surfaces on a mid-flight timeout. + // This test pins the upsert loop's EXCEPTION-HANDLING contract (does it + // retry a TransactionTimeoutError like it retries + // InsertPreconditionExistsSignal?), not the timing mechanics of a real + // timeout — those are pinned by test (i) and by + // tests/unit/transaction/timeout-rollback.test.ts. + ;(brain as any).persistSingleOp = async (): Promise => { + persistSingleOpCalls++ + throw timeoutError + } + + await expect( + brain.add({ data: 'a', type: NounType.Thing, vector: V() }) + ).rejects.toBe(timeoutError) + + // The loop's attempt counter: exactly one call, never retried up to + // MAX_UPSERT_ATTEMPTS. + expect(persistSingleOpCalls).toBe(1) + await brain.close() + }) +})