Skip to content

current-coverage.md

Latest commit

 

History

History
91 lines (72 loc) · 3.27 KB

current-coverage.md

File metadata and controls

91 lines (72 loc) · 3.27 KB
Raw

Current test coverage

The single canonical, always-current source for test counts and coverage (per CLAUDE.md, quality-checks.md, and the developer testing.md). Other docs reference this file instead of duplicating a number that would drift.

Counts are collected with --collect-only (no run needed); coverage percentages come from the CI coverage.yml night shift, not from a local run (make test-coverage is heavy and thermally stressful, so it is opt-in only). Re-measure with the commands in the last section.

Test counts

Level Suite Count How measured (this session, v2.15.0)
Unit + integration Backend (backend/tests) 1858 cd backend && poetry run pytest --collect-only -q
Unit + integration Plugins (14 packages, sum below) 1137 per-plugin pytest --collect-only, backend's shared venv
Unit + component Frontend (Vitest) 10438 (988 files) cd frontend && bunx vitest run (tail total, CI push run)
Grand total measured this session 13433 1858 + 1137 + 10438

E2E (Playwright smoke + Dexie-mode + visual) is separate from make test and is not counted here; run cd e2e && npx playwright test.

Plugins, per package

Plugin Tests
assessment 110
session 237
tracking 64
tools 58
gamification 69
content-loader 339
anki 20
notebooklm 27
learning-repo 53
missions 41
ai-anthropic 35
ai-openai 32
ai-gemini 34
ai-perplexity 18
Total 1137

Delta since the last recorded baseline

The previous recorded baseline is the v2.14.0 snapshot (history/2026-09-17-coverage.md). v2.15.0 added parametric exercises, three extension types, the Settings reorganisation, the compact lesson summary and the pytest path-isolation guard (#3145).

Suite Baseline (v2.14.0) Current (v2.15.0) Delta
Backend 1824 1858 +34
Plugins 1130 1137 +7
Frontend (Vitest) 9775 10438 +663

Coverage percentages

Coverage runs on CI (.github/workflows/coverage.yml, daily 03:00 UTC

  • workflow_dispatch), NOT as part of the local workflow. Pull the latest HTML + coverage.xml reports without a local run:
gh run download --name backend-coverage
gh run download --name frontend-coverage
gh run download --name adaptive-learner-plugin-gamification-coverage   # etc. per plugin

Targets per module type live in quality-checks.md ("Coverage Targets per Module Type"). Codecov is intentionally not wired up.

Re-measuring (authoritative commands)

# Backend
cd backend && poetry run pytest --collect-only -q | tail -1

# Plugins (sum across all packages)
cd backend && for d in ../plugins/adaptive-learner-plugin-*/; do \
  poetry run pytest "$d/tests" --collect-only -q 2>/dev/null | tail -1; done

# Frontend (Vitest) - runs the suite; take the "Tests N passed" total
cd frontend && bunx vitest run 2>&1 | grep -E "Tests +[0-9]+"

Update this file when the counts move (this IS the canonical location - do not copy the numbers elsewhere). Per test-coverage-audits.md, a full audit additionally copies the previous version to docs/audits/history/YYYY-MM-DD-coverage.md before overwriting.