Testing with pymodbus Simulator

[DostalJ]

(Dear developers, thank you very much for your work. I enjoy using the package!)

Recently, on a new project, we started using pymodbus.simulator with two objectives:

1. Allow the team to develop locally without access to PLC running Modbus Server.
2. Automate testing of our python code using pytest.

What we do now

When we start developing, we run a simple bash command:

./venv/bin/pymodbus.simulator --json_file ./simulator_setup.json --log debug

which starts a configured simulator that mocks the real Modbus Server the code is supposed to interact with. This works great and achieves our objective – but when running tests, this approach is not ideal.

Before automatically testing our code (using pytest), we manually start the Simulator and run the tests. The tests could look like this:

class ExampleDeviceModbusClient:
    def __init__(self, host: str, port: int):
        self.client = ModbusTcpClient(host, port)

    def check_connection(self) -> bool:
        """
        Check if the Modbus device is reachable.
        Returns True if reachable, False otherwise.
        """
        try:
            # Simple diagnostic request
            result = self.client.read_device_information()
            return not result.isError()
        except ConnectionException:
            return False


@pytest.fixture
def device():
    return ExampleDeviceModbusClient("0.0.0.0", port=503)


def test_connection(device: ExampleDeviceModbusClient):
    assert device.check_connection() in [True, False]

ExampleDeviceModbusClient is a minimal class that illustrates object we work with, device() is a pytest fixture and test_connection() is simple test for purposes of this example

What we'd like to do

The workflow above works, but is not ideal. The optimal solution would be to start Simulator only before test and close it after tests are done. That would allow us to automate the testing in some CI system, like Github Actions. I have tried some ways, but failed. My guess is that I don't understand how Simulator is started directly from python.

My attempt (not working):

import multiprocessing

import pytest
from pymodbus.client import ModbusTcpClient
from pymodbus.exceptions import ConnectionException
from pymodbus.server import ModbusSimulatorServer


async def run_simulator():
    task = ModbusSimulatorServer(json_file="../simulator_setup.json")
    await task.run_forever(only_start=True)


@pytest.fixture(autouse=True)
def run_around_tests():
    # Code that will run before your test, for example:
    p = multiprocessing.Process(target=run_simulator)
    # A test function will be run at this point
    yield
    # Code that will run after your test, for example:
    p.close()


class ExampleDeviceModbusClient:
    def __init__(self, host: str, port: int):
        self.client = ModbusTcpClient(host, port)

    def check_connection(self) -> bool:
        """
        Check if the Modbus device is reachable.
        Returns True if reachable, False otherwise.
        """
        try:
            # Simple diagnostic request
            result = self.client.read_device_information()
            return not result.isError()
        except ConnectionException:
            return False


@pytest.fixture
def device():
    return ExampleDeviceModbusClient("0.0.0.0", port=503)


def test_connection(device: ExampleDeviceModbusClient):
    assert device.check_connection() in [True, False]

Do you have any experience or success with automatic testing using pymodbus.simulator? How could I make this work? Would you reccommend any other approach?

[janiversen]

There are several easy ways to do that.

The best way is to use pytest functionality like e.g.:

https://stackoverflow.com/questions/7786648/how-to-call-setup-once-for-all-tests-and-teardown-after-all-are-finished

We do it slightly different with a fixture, that yield to the test...but this is because we want the server to start/stop with each single test.

see test/conftest system_health_check