Sync quickstart¶

The async API is canonical, but a sync facade lives at :class:dtollib.sync.Dtol for scripts, notebooks, and REPL use.

Read two thermocouples¶

from dtollib import (
    TaskSpec,
    ThermocoupleInput,
    ThermocoupleType,
)
from dtollib.sync import Dtol


spec = TaskSpec(
    name="surface_temperatures",
    channels=[
        ThermocoupleInput(
            physical_channel=0,
            name="surface_tc_K",
            thermocouple_type=ThermocoupleType.K,
            min_val_degc=-50.0,
            max_val_degc=200.0,
        ),
        ThermocoupleInput(
            physical_channel=1,
            name="back_tc_K",
            thermocouple_type=ThermocoupleType.K,
            min_val_degc=-50.0,
            max_val_degc=200.0,
        ),
    ],
)


with Dtol.open_device(spec) as session:
    reading = session.poll()
    print(reading.values)
    # → {"surface_tc_K": 23.41, "back_tc_K": 22.97}

What the sync facade is¶

Dtol.open_device(...) returns a :class:SyncDtolSession, which is a blocking wrapper around the async :class:DtolSession. Internally, the wrapper spins up an :class:anyio.from_thread.BlockingPortal so every synchronous call dispatches through the same async code path — no parallel implementation.

That means:

• The sync API has the same behaviour as the async API. If await session.poll() returns a sentinel-filled DaqReading, so does session.poll() in the sync facade.
• The sync session is not reusable — exiting its with block closes the portal. Open a fresh session for each run.

When to use it¶

Use the sync facade when:

• You're writing a notebook or REPL session and don't want to wrap every call in anyio.run(...).
• You're integrating into a sync codebase and async-all-the-way isn't available.

Use the async API when:

• You're composing acquisition with other async I/O (Alicat MFCs, Sartorius balances, network sinks).
• You want per-task concurrency via DtolManager.poll(names) — the async manager parallelises cross-board polls.