ss-tools/test.coverage.md at c30cca78f3e1b8fc46f3e8abfafca863f0d464a1

Files

busya d1d2a0f92e feat: tiered test infrastructure with Makefile + smart selector + OpenCode commands

Root Makefile with timeout-protected test targets:
  - Tier 1 (<30s):   make test, make test-unit, make test-frontend
  - Tier 2 (smart):  make test-related F=file.py (via @RELATION BINDS_TO)
  - Tier 3 (<5min):  make test-integration (Docker, --run-integration)
  - Coverage:        make coverage (backend + frontend)
  - Lint:            make lint (ruff + eslint)

Smart test selector (scripts/find-related-tests.py):
  - Extracts module names from #region anchors, class/function defs
  - Searches 400 BINDS_TO entries across all test files
  - Confidence scoring: exact > case-insensitive > substring > heuristic
  - Fallback: filename-based fuzzy matching

OpenCode commands:
  - /test.all      — full suite + coverage
  - /test.unit     — fast unit tests (<30s)
  - /test.related  — smart selection by file
  - /test.coverage — coverage reports with thresholds

Speckit workflow updated:
  - speckit.test.md:   raw pytest → make targets with timeout safety
  - speckit.plan.md:   quickstart uses make targets
  - speckit.tasks.md:  verification uses make targets
  - speckit.implement.md: default stack uses make targets

Frontend: added 'coverage' script to package.json

2026-07-21 18:24:53 +03:00

2.5 KiB

Raw Blame History

description, handoffs, tools

description

handoffs

tools

Generate coverage reports for both backend and frontend, and verify against thresholds.

label	agent	prompt	condition
Improve Coverage	qa-tester	Coverage is below threshold. Identify uncovered critical paths and propose additional tests. Current uncovered areas: $ARGUMENTS	Coverage below threshold

bash, read, grep

User Input

$ARGUMENTS

Goal

Generate test coverage reports for backend (pytest-cov) and frontend (vitest v8), and verify that coverage meets project thresholds.

Required Skills

MANDATORY USE skill({name="semantics-testing"}) — coverage conventions.

Execution Steps

1. Generate coverage reports

make coverage

2. Review backend coverage

# Detailed terminal output with uncovered lines
cd backend && source .venv/bin/activate && python -m pytest tests/ --ignore=tests/integration/ --cov=src --cov-report=term-missing 2>&1 | tail -50

Key metrics to extract:

Overall statement coverage (%)
Files with <80% coverage (list top 5 offenders)
Files with 0% coverage (untested)

3. Review frontend coverage

# Frontend coverage (also available via make coverage-frontend)
cd frontend && npx vitest run --coverage 2>&1 | tail -50

Frontend thresholds in vitest.config.js:

Statements: 98%
Lines: 98%
Functions: 95%
Branches: 80%

4. Check against thresholds

If any threshold is not met, identify the specific files/modules dragging coverage down.

5. (Optional) Open HTML reports

ls backend/htmlcov/index.html && echo "Backend report: backend/htmlcov/index.html"
ls frontend/coverage/index.html && echo "Frontend report: frontend/coverage/index.html"

Output Format

## Coverage Report

### Backend (pytest-cov)
- Statement Coverage: XX% 
- Files below 80%: N (list top 3-5)
- Untested files: N (list top 3-5)

### Frontend (vitest v8)
- Statement Coverage: XX% (threshold: 98%) [PASS/FAIL]
- Line Coverage: XX% (threshold: 98%) [PASS/FAIL]  
- Function Coverage: XX% (threshold: 95%) [PASS/FAIL]
- Branch Coverage: XX% (threshold: 80%) [PASS/FAIL]

### HTML Reports
- Backend: backend/htmlcov/index.html
- Frontend: frontend/coverage/index.html

### Overall: [ALL_THRESHOLDS_MET / BELOW_THRESHOLD]

Constraints

Coverage is generated from unit tests ONLY (no Docker integration tests).
If vitest coverage fails with "threshold not met", report which files are below threshold.
Do NOT modify source code to artificially increase coverage.

2.5 KiB Raw Blame History