SigHej/Docs/TESTING.md

Testing & Code Quality — Social Proximity

Scope

These rules apply to the Python backend (backend/). The Flutter app and Next.js admin have separate tooling.

Quality analysis runs via DevOpsMCP — a set of code analysis tools accessible through the MCP server at https://devops-mcp.i80.dk.

---

When to Run What

Tool	Trigger	Threshold
check_pep_compliance	Before every commit	Zero violations
check_personal_standards	Before every commit	Zero error-level findings
analyze_complexity	When writing functions > 20 lines	Max cyclomatic complexity: 8
add_type_hints	When adding new functions	Min type hint coverage: 80%
analyze_missing_docstrings	New classes and public functions	No missing docstrings on public API
analyze_yaml + analyze_github_actions	When editing Gitea CI/CD workflow files	Zero security violations

---

Coding Standards (Python backend)

Enforced by check_personal_standards

• No subprocess with shell=True — use SDK clients instead
• No bare except: — always catch specific exceptions
• No % or .format() string formatting — use f-strings
• No mutable default arguments — use None and set inside function body
• No hardcoded secrets — use environment variables via pydantic-settings

Enforced by analyze_complexity

• Max cyclomatic complexity per function: 8
• Functions exceeding this must be refactored (extract helper functions)

Enforced by add_type_hints

• All function parameters must have type annotations
• All function return types must be annotated
• Use Optional[T] or T | None — not bare None returns without annotation

---

Workflow

1. Write code
2. Run check_personal_standards(file) → fix all 'error' findings
3. Run analyze_complexity(file) → refactor if any function > 8
4. Run add_type_hints(file, apply_changes=False) → review suggestions, apply manually
5. Run check_pep_compliance(file) → fix formatting
6. Commit

---

Test Strategy (backend)

Tests live in backend/tests/ and use pytest.

Layer	What to test	Example
Unit	Service logic in isolation	test_match_service.py — interest overlap calculation
Integration	API endpoints with real Redis/DB	test_session.py — POST /session creates Redis key
Contract	Request/response shapes match Pydantic models	Via pytest + httpx TestClient

Run tests:

cd backend
pytest tests/ -v

---

Flutter App Testing

Layer	Tool	Notes
Unit	flutter test	Test service logic, models
Widget	flutter test + WidgetTester	Test UI components in isolation
Integration	flutter drive	Requires physical device or emulator

BLE scanning cannot be tested in emulators — use a physical device for BLE integration tests.

---

DevOpsMCP Quick Reference

# Upload file and run analysis
session = DevOpsMCP-start_project_session("sighej-backend", [
    {"file_path": "main.py", "content": "<content>"}
])

DevOpsMCP-check_personal_standards(file_path="main.py")
DevOpsMCP-analyze_complexity(file_path="main.py")
DevOpsMCP-add_type_hints(file_path="main.py", apply_changes=False)
DevOpsMCP-check_pep_compliance(file_path="main.py")
DevOpsMCP-generate_tests(file_path="main.py")