OpenSMI Client

OpenSMI Client is the client half of the OpenSMI framework: a Python library for connecting to any SMI-compliant OPC UA server — generically discovering its machines, components and skills, and interacting with them without needing a generated, server-specific client.

Since SMI defines a standardized information model, OpenSMI Client doesn't need to know your server's specific machine/skill types ahead of time — it browses the address space at runtime and exposes what it finds through a consistent Python API.

Browsing is optional, though: if you already know the relevant node IDs (e.g. from the server's configuration, or cached from a previous run), you can point the client directly at them and skip discovery entirely, or mix both approaches — browse once to locate a machine, then operate on it and its sub-objects using known nodes from there on.

Requirements

• Python 3.11+
• An SMI-compliant OPC UA server to connect to (e.g. one built with OpenSMI-Server)

Installation

pip install opensmi-client

Quick Start

import asyncio

from opensmi.core import SkillState
from opensmi.client import RemoteServer


async def main() -> None:
    # remember: do not commit the actual password to version control...
    async with RemoteServer("opc.tcp://localhost:4841/", username="operator", password="operator") as server:
        machine = server.machines["ExampleMachine"]

        await machine.lock.init_lock()

        skill = machine.skill_set["ExampleSkill"]
        # reset the Skill if necessary
        if await skill.read_current_state() != SkillState.READY:
            await skill.reset(wait_for=SkillState.READY)

        # bulk write the parameters
        await skill.parameter_set.write_all({"x": 5, "y": 10})

        # start the skill and wait for completion
        await skill.start(wait_for=SkillState.COMPLETED)

        # retrieve the results before resetting
        print("results:", await skill.final_result_data.read_all())

        await skill.reset(wait_for=SkillState.READY)


if __name__ == "__main__":
    asyncio.run(main())

This connects to the ExampleMachine from the server quick start, locks it, drives its ExampleSkill through a full parameter-write → start → completion → read-result → reset cycle, and prints the computed sum.

Project Status

The client API is stable and follows semantic versioning. Changes to the underlying OPC UA information model are incorporated as needed but kept backwards compatible, and do not change the Python API — code written against OpenSMI Client should keep working across information-model revisions.

License

The library itself is licensed under the MIT License.

Example code contained in the examples directory is dedicated to the public domain under the CC0 1.0 Universal license.

---

This text was drafted with AI assistance (Claude, Anthropic) and reviewed/edited by the author before publication.