Rounds Test

The rounds project uses fakes over mocks (fake implementations of ports in tests/fakes/) to ensure tests are maintainable and reflect real adapter behavior.

Implementation

# Change to the rounds package directory (where pyproject.toml lives)
cd /home/austinsand/workspace/orchestrator/rounds/rounds

# Run pytest with the specified test path (or all tests if no path given)
if [ -n "$TEST_PATH" ]; then
    pytest -v "tests/$TEST_PATH"
else
    pytest -v
fi

Configuration details (from rounds/pyproject.toml):

• Test discovery pattern: test_*.py files (line 54)
• Test class pattern: Test* classes (line 55)
• Test function pattern: test_* functions (line 56)
• Async mode: Automatic asyncio loop handling (line 52)
• Test paths: tests/ directory (line 53)

Examples

Example 1: Run all tests

/rounds-test

Output: Runs the entire test suite including:

• tests/core/ - Domain logic unit tests
• tests/adapters/ - Adapter integration tests
• tests/fakes/ - Fake implementation validation
• tests/integration/ - End-to-end workflow tests
• Root level tests: test_composition_root.py, test_new_implementations.py, test_workflows.py

Example 2: Run specific test file

/rounds-test test_composition_root.py

Output: Runs only the composition root tests (13,693 bytes of dependency injection validation)

Example 3: Run core domain tests

/rounds-test core

Output: Runs unit tests for:

• core/models.py - Immutable domain entities (Signature, Diagnosis, ErrorEvent)
• core/ports.py - Abstract port interfaces
• core/fingerprint.py - Error fingerprinting logic
• core/triage.py - Error classification
• core/investigator.py - Investigation orchestration

Example 4: Run adapter tests

/rounds-test adapters

Output: Runs integration tests for concrete adapter implementations:

• adapters/store/sqlite.py - SQLite persistence layer
• adapters/diagnosis/claude_code.py - Claude-powered diagnosis engine
• adapters/telemetry/ - Trace query implementations (SigNoz, Jaeger, etc.)

Example 5: Run specific test function

/rounds-test test_workflows.py::test_poll_cycle

Output: Runs only the poll cycle workflow test (complete error detection → fingerprinting → diagnosis flow)

Test Strategy Overview

The rounds project follows hexagonal architecture testing principles:

1. Domain layer tests (tests/core/) use fakes from tests/fakes/ to validate business logic without external dependencies
2. Adapter tests (tests/adapters/) verify concrete implementations against port contracts
3. Integration tests (tests/integration/) validate full system composition in main.py

Key test files (from rounds/tests/):

• test_composition_root.py (13KB) - Dependency wiring validation
• test_new_implementations.py (25KB) - New feature integration tests
• test_workflows.py (16KB) - End-to-end diagnostic workflows

Fake implementations (from tests/fakes/):

• fakes/store.py - In-memory signature repository
• fakes/telemetry.py - Synthetic error event generator
• fakes/diagnosis.py - Deterministic diagnosis engine

This skill was automatically generated from the rounds project structure and pyproject.toml configuration.