opensemantic.lab

Python models and controllers for the world.opensemantic.lab page package.

Builds on opensemantic.base (DataToolController, TimeSeriesDatabaseController, archiving, typed read/write).

Overview

• Auto-generated Pydantic models (v1 and v2): OpcUaServer, OpcUaDataChannel, LaboratoryProcess, Sample, Material, etc.
• OpcUaServer controller - OPC UA client/server lifecycle with channel subscriptions and data archiving

OpcUaServer controller

Extends DataToolController with OPC UA protocol logic:

from opensemantic import compute_scoped_uuid
from opensemantic.lab.v1 import OpcUaServer, OpcUaDataChannel, OpcUaClientMode
from opensemantic.characteristics.quantitative.v1 import Temperature, Time, TimeUnit
from opensemantic.core.v1 import Label
from uuid import NAMESPACE_URL, uuid5

SERVER_UUID = uuid5(NAMESPACE_URL, "my-server")

server = OpcUaServer(
    uuid=SERVER_UUID,
    name="my_server",
    label=[Label(text="My Server")],
    url="opc.tcp://localhost:4840",
    data_channels=[
        OpcUaDataChannel(
            uuid=compute_scoped_uuid(SERVER_UUID, "ns=2;s=Temperature"),
            node_id="ns=2;s=Temperature",
            name="temperature",
            label=[Label(text="Temperature")],
            opcua_data_type="Float",
            client_mode=OpcUaClientMode.Subscription,
            sampling_interval=Time(value=100, unit=TimeUnit.milli_second),
            refresh_interval=Time(value=500, unit=TimeUnit.milli_second),
            characteristic=Temperature.get_cls_iri(),
        ),
    ],
    auto_archive=True,
)

OPC UA-specific methods

• run_as_server(params) - start an OPC UA server with value callbacks
• run_as_client(params) - connect, subscribe to channels, auto-archive
• read_channel(params) / write_channel(params) - direct channel I/O
• browse(params) - browse the OPC UA address space
• stop() - flush archive buffer and disconnect

Channel UUIDs

OpcUaDataChannel requires an explicit UUID. Use compute_scoped_uuid(server_uuid, node_id) from opensemantic to avoid collisions across servers.

Controller fields (runtime only)

url, archive_database, auto_archive, reset_opcua_connection_on_error are controller-only fields. They are stripped from to_json() / to_jsonld() serialization. Set them as runtime config when constructing the controller.

Inherited features

All DataTool features from opensemantic.base work automatically: auto-archive from storage_locations, typed read/write, subobject ID prefixing, channel characteristic warnings. See the opensemantic.base README for details.

LiveDataToolView (Live Dashboard UI)

Extends the base DataToolView with real-time OPC UA streaming via Bokeh.

Features:

• Archive tab with all base DataToolView features (stacked plots, unit switching, log console)
• Live tab with Bokeh streaming plots updated via OPC UA subscriptions
• Per-channel unit conversion for live data
• Configurable history window, buffer size, and update interval via JsonEditor

Interactive demo: archive view and real-time OPC UA streaming

Archive view with accumulated OPC UA data

Real-time OPC UA streaming with stacked Bokeh plots

from opensemantic.lab.view import LiveDataToolView
from opensemantic.base.view._config import LiveDashboardConfig, PlotConfig

view = LiveDataToolView(
    controllers=[ctrl],
    config=LiveDashboardConfig(lang="en", plot=PlotConfig(auto_fetch=True)),
    title="Live Dashboard",
)
view.servable()

See examples/live_dashboard.py for a full working example with an embedded OPC UA server.

Examples

See examples/ for complete runnable examples:

• opc_ua_server.py - define server, store to oold backend, run
• opc_ua_client.py - load from backend, connect, print archived data
• opcua_archiving.py - full workflow with auto-archiving and typed read/write
• live_dashboard.py - live dashboard with embedded OPC UA server

Installation

pip install opensemantic.lab             # models only
pip install opensemantic.lab[controller] # + asyncua, opensemantic.base[controller]
pip install opensemantic.lab[view]       # + opensemantic.base[view]

Testing

pytest tests/test_controller.py

Tests include OPC UA server/client integration (uses asyncua's built-in test server, no external services needed).

For PostgREST-based archiving tests, see the opensemantic.base testing docs - requires a running pgstack instance with TimescaleDB.