feat(ai-partner): Phase 1 E2E smoke test behind feature flag (#1453)#1481
Merged
Conversation
Implements the Amicus retrieval service layer that embeds a user query,
runs sqlite-vec MATCH against scripture.db, and re-ranks with profile +
context boosts.
New modules under app/src/services/amicus/:
- types.ts — shared types (RetrievedChunk, RetrievalContext,
CompressedProfile, AmicusError + AmicusErrorCode)
- embed.ts — calls proxy /ai/embed; retries once on 5xx/network;
translates proxy 401/402 to PROXY_UNAUTHORIZED,
network failure to OFFLINE
- vectorSearch.ts — packVector (little-endian float32), distance→similarity,
searchByVector executing the MATCH join on
embeddings/chunk_text/chunk_metadata. Throws
EXTENSION_NOT_LOADED if sqlite-vec isn't loaded.
- rerank.ts — pure functions: current-chapter ×1.5, preferred
scholars ×1.1, traditions ×1.05, 2-per-scholar
diversity cap, top-10
- retrieval.ts — top-level orchestrator with latency timing
- index.ts — tiny public barrel
30 Jest tests (4 suites) across the pipeline:
- rerank: boosts, diversity, ordering
- vectorSearch: packing, similarity, error on extension missing
- embed: shape validation, retry, 401/402, offline, timeout
- retrieval: end-to-end with fake fetch + mock DB
All 3213 app tests still pass; tsc --noEmit + eslint 0 warnings.
https://claude.ai/code/session_01Pht3kzgdvkn81DDfL9SnFe
Deterministic, on-device generator that reads engagement signals from user.db and produces a 100-200-token prose summary. Sent to Amicus on every request; user-inspectable in Settings (#1459). Migration v15 adds partner_profile_cache (singleton, id=1) to user.db. The generator SHA-256-hashes the raw signals and skips regeneration when the hash matches a <7-day-old cached row. New modules under app/src/services/amicus/profile/: - types.ts — RawSignals, CompressedProfile, ProfileForInspection - signals.ts — queries for total_chapters_read, last_30_day, top scholars, tradition + genre distribution, journey state, recent chapters, current focus. Tables that don't exist yet (scholar_opens, journey_progress) degrade gracefully to empty. - templates.ts — prose assembly with six optional sections; thin profiles get a deliberately minimal "new to study" phrasing so Amicus doesn't over-personalize. - generator.ts — generateProfile(force), getProfileForInspection(), clearProfile(), hashRawSignals(). Stable sorted-key canonicalization so identical signals always hash identically. - index.ts — public barrel 16 Jest tests: template section inclusion rules, thick/thin profiles, truncation, 24-hour reading window, cache hit/miss/expire/force, hash determinism and key-order independence, clearProfile DELETE. Also bumps __tests__/{db,unit}/userDatabase.test.ts migration counts from 14 → 15 to match. All 3,199 app tests pass; tsc clean. https://claude.ai/code/session_01Pht3kzgdvkn81DDfL9SnFe
…tor' into claude/issue-1453-e2e-smoke-test
Depends on #1447, #1448, #1450, #1451, #1452 (merge commits stacked). New modules: - app/src/constants/featureFlags.ts — new registry; AMICUS_SMOKE_TEST flag (__DEV__ && EXPO_PUBLIC_AMICUS_SMOKE=true) - app/src/services/amicus/__smoke__/canned_queries.json — 10 queries covering synthesis, lexicon, debate, cross-testament typology, journey stops, deliberate corpus gap (q06), meta-FAQ lookup, character study, archaeology, textual transmission - app/src/services/amicus/__smoke__/runSmokeTest.ts — harness driving retrieve() + /ai/chat streaming; extracts citations, parses the trailing gap_signal, produces a structured SmokeReport with p50/p95 latency, pass counts, and per-query diagnostics - app/src/screens/dev/AmicusSmokeScreen.tsx — internal utility UI: header, Run button, per-query rows that expand to show failures, citations, and response preview; Copy-JSON-to-clipboard Wired: - app/src/navigation/types.ts — new AmicusSmoke route on MoreStackParamList - app/src/navigation/MoreStack.tsx — lazy-registered only when the flag is on, so production builds never surface the screen Tests: 10 unit tests covering canned-query shape, citation extraction, and trailing JSON parsing. Full run (3,239 tests) stays green; tsc clean. Phase 1 exit criterion: this harness passing against staging proxy. https://claude.ai/code/session_01Pht3kzgdvkn81DDfL9SnFe
Test Results✅ All tests passed
Coverage
⏱️ Duration: 73.3s |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #1453. Depends on #1447, #1448, #1450, #1451, #1452.
Summary
gap_signal.gap=true), meta-FAQ, character study, archaeology, and textual transmission.featureFlags.AMICUS_SMOKE_TEST(requires__DEV__andEXPO_PUBLIC_AMICUS_SMOKE=true).New modules
app/src/constants/featureFlags.ts— the flag registry (first occupant isAMICUS_SMOKE_TEST).app/src/services/amicus/__smoke__/canned_queries.json— 10 queries w/ expected citations + source types.app/src/services/amicus/__smoke__/runSmokeTest.ts— harness: retrieves, streams chat, extracts citations, parses gap_signal, builds SmokeReport with p50/p95 latency.app/src/screens/dev/AmicusSmokeScreen.tsx— minimal utility UI: Run button, expandable per-query rows, Copy-JSON-to-clipboard.Wiring
app/src/navigation/types.ts— addsAmicusSmoketoMoreStackParamList.app/src/navigation/MoreStack.tsx— lazy-registered only when the flag is on.Test plan
npx tsc --noEmitcleannpx jest— 3,239 tests pass (10 new: canned-query shape, citation extraction, trailing-JSON parsing)How to run locally
cd app && EXPO_PUBLIC_AMICUS_SMOKE=true EXPO_PUBLIC_AMICUS_DEV_TOKEN=<receipt> EXPO_PUBLIC_AMICUS_PROXY_URL=https://ai-staging.contentcompanionstudy.com npm run devOut of scope
https://claude.ai/code/session_01Pht3kzgdvkn81DDfL9SnFe