Testing Patterns

Deterministic Simulations

Use constant arrival times and fixed seeds for reproducible tests:

from happysimulator import Source, ConstantArrivalTimeProvider

# Constant rate for predictable timing
source = Source.constant(rate=10, target=server)

# Or use the full constructor
source = Source(
    name="Traffic",
    event_provider=provider,
    arrival_time_provider=ConstantArrivalTimeProvider(interval=0.1),
)

For stochastic components, set seeds:

import random
random.seed(42)
# or use seed= parameter on distributions

Test Structure

tests/
├── unit/           # fast, isolated tests
├── integration/    # full simulation tests
└── conftest.py     # shared fixtures

Integration Tests

Integration tests need a Simulation to inject clocks into entities:

from happysimulator import Simulation

def test_server_processes_request():
    sink = Sink()
    server = MyServer("server", downstream=sink)
    source = Source.constant(rate=1, target=server)

    sim = Simulation(
        entities=[source, server, sink],
        duration=10,
    )
    sim.run()

    assert sink.events_received > 0

Fixtures

test_output_dir

Per-test directory for artifacts (plots, logs):

def test_with_output(test_output_dir):
    # test_output_dir is a Path to a unique directory
    plot_path = test_output_dir / "latency.png"

Visualization Tests

import pytest

def test_plot():
    plt = pytest.importorskip("matplotlib")
    import matplotlib
    matplotlib.use("Agg")  # non-interactive backend
    # ... create plots ...

Common Pitfalls

Problem	Fix
Non-deterministic tests	Use ConstantArrivalTimeProvider, set random.seed(42)
self.now is None	Register entity in Simulation(entities=[...])
Queue grows forever	Ensure arrival rate < service rate
Source timing surprises	Source.constant(rate=1) first event at t=1.0, not t=0

Next Steps

• Core Concepts — Event, Entity, Simulation fundamentals
• Logging — enable debug logging for test failures