Skip to content

Repository navigation

TESTING.md

Latest commit

 

History

History
110 lines (81 loc) · 4.57 KB

TESTING.md

File metadata and controls

110 lines (81 loc) · 4.57 KB
Raw

Testing

Sift uses automated tests suitable for a product moving toward production use in professional environments: fast unit and API tests on every change, plus a GitHub Actions CI workflow on every push and pull request.

Test inventory

Suite File(s) Count What is verified
Backend API backend/src/__tests__/app.test.ts 13 All HTTP endpoints: GET /health, settings CRUD, POST /api/ingest, GET /api/documents, GET /api/documents/:id, DELETE /api/documents/:id, POST /api/scan
Backend ingest backend/src/__tests__/ingest.test.ts 5 LLM success path, LLM fallback (heuristic summary + error_message), file-not-found, directory rejection, raw_preview truncation
FHIR pipeline backend/src/pipelines/__tests__/fhir.test.ts 2 Invalid JSON handling, patient name extraction from Bundle
HL7 pipeline backend/src/pipelines/__tests__/hl7.test.ts 3 looksLikeHl7 detection, segment parsing, PID extraction
Frontend API client frontend/src/api.test.ts 8 All api.* methods (health, getSettings, saveSettings, listDocuments, getDocument, ingestPath, deleteDocument), error propagation
Markdown renderer frontend/src/MarkdownText.test.tsx 15 Bold, italic, h1/h2/h3 headings, */- unordered lists, 1. ordered lists, blank-line list flushing, inline bold inside list items, realistic LLM synthesis output
Tauri runtime helper frontend/src/tauri-env.test.ts 1 isTauriRuntime() returns false in jsdom
Total 47

Stack

Layer Framework Notes
Backend Jest 29 + ts-jest + Supertest In-memory SQLite (openMemoryDatabase()); LLM mocked via jest.fn() on global.fetch
Frontend Vitest 3 + jsdom + @testing-library/react fetch mocked with vi.stubGlobal; Tauri APIs not available in jsdom
Tauri host cargo check (CI only) Verifies Rust compiles; full driver/UI tests are a later increment

Running tests

Everything at once (from repo root)

npm test

Runs backend Jest, then frontend Vitest.

Backend only

cd backend
npm test              # Jest + coverage + --forceExit
npm run test:watch    # Jest watch mode

Frontend only

cd frontend
npm test              # vitest run
npm run test:watch    # vitest interactive

Typecheck (no test run)

# Backend
cd backend && npx tsc -p tsconfig.build.json --noEmit

# Frontend
cd frontend && npx tsc --noEmit

# Or via root shortcut
npm run lint:backend

GitHub Actions CI

Workflow: .github/workflows/ci.yml

Three parallel jobs run on every push and pull request:

Job Steps
Backend npm ci → tsc --noEmit → jest --coverage --forceExit → upload backend/coverage/ artifact (14 days)
Frontend npm ci → tsc --noEmit → vitest --coverage → upload frontend/coverage/ artifact (14 days) → vite build
Tauri host Install Linux WebKit/GTK deps → dtolnay/rust-toolchain@stable → cargo check --locked

Coverage artifacts are available as downloadable zip files from the Actions tab on any completed workflow run.

Note: Commit package-lock.json files for both backend/ and frontend/ so npm ci is reproducible in CI.

Adding new tests

  • Backend endpoint → add a case in backend/src/__tests__/app.test.ts using the existing createApp(db) + openMemoryDatabase() pattern.
  • Ingest path → add to backend/src/__tests__/ingest.test.ts; mock global.fetch for LLM responses.
  • Pipeline → add to backend/src/pipelines/__tests__/<name>.test.ts.
  • Frontend component → add a *.test.tsx file alongside the component; use render() + cleanup() from @testing-library/react and query within the returned container, not document, to avoid cross-test contamination.
  • Frontend API method → add a case in frontend/src/api.test.ts following the vi.stubGlobal("fetch", mockFetch(...)) pattern.

Roadmap

  • Raise coverage on pdf.ts (PDF fixture) and llm.ts (timeout / non-200 error paths).
  • React Testing Library component tests for the SettingsForm and document delete flow.
  • Windows runner job for tauri build on release tags (add cross-env CI=false as documented in Build and release).
  • Coverage gate: fail CI if overall coverage drops below a defined threshold.

Related docs