Devices

A device is a persistent process that exposes methods to tasks. It can control physical equipment or hold virtual state across protocol runs. EOS creates device processes when loading a lab. For objects that only need exclusive allocation, use Resources.

The GC Sampling task uses a gas chromatograph and a robot for sample injection.

Device Implementation

Devices live in the devices directory of an EOS package. Each device has its own subdirectory (e.g., devices/magnetic_mixer) containing a device.py.

A device is a class that subclasses Device with a type. The docstring is the device description, and a nested Config model declares its initialization parameters:

device.py

from typing import Any

from eos import Device

from my_package.common.device_client import DeviceClient


class MagneticMixer(Device, type="magnetic_mixer"):
    """Magnetic mixer for mixing the contents of a container."""

    class Config(Device.Config):
        port: int = 5004

    async def _initialize(self, config: Config) -> None:
        self.client = DeviceClient(config.port)
        self.client.open_connection()

    async def _cleanup(self) -> None:
        self.client.close_connection()

    async def _report(self) -> dict[str, Any]:
        return {"port": self.client.port}

    def mix(self, container: str, mixing_time: int, mixing_speed: int) -> None:
        self.client.send_command("mix", {"container": container, "time": mixing_time, "speed": mixing_speed})

One implementation can back several devices of the same type. A lab sets each device’s init_parameters, which EOS validates against Config when the lab loads and again when the device starts. Fields without a default are required.

The lifecycle methods are optional:

• _initialize opens connections, using the validated Config.

• _cleanup releases them.

• _report returns current state.

Public methods such as mix are what tasks call.

Shared Device Code

A Device subclass without a type is not registered, so it can serve as a base class. Subclasses inherit its methods and type, and extend its Config:

class BridgeDevice(Device):
    class Config(Device.Config):
        host: str = "127.0.0.1"
        port: int = 8765

    async def _initialize(self, config: Config) -> None:
        self.bridge = Bridge(config.host, config.port)


class Hotplate(BridgeDevice, type="hotplate"):
    """Hotplate reached through the bridge."""

    class Config(BridgeDevice.Config):
        station: str

    async def _initialize(self, config: Config) -> None:
        self.station = config.station
        await super()._initialize(config)

Devices on Other Computers

EOS ships device code to the computer running the device, so the package does not need to be installed there. Third-party libraries it uses must be installed on that computer. Import libraries that only exist there, such as Windows-only SDKs, inside _initialize so the orchestrator can still load the device.