Ask AI

Asset sensors#

Asset sensors allow you to instigate runs when materializations occur.


Relevant APIs#

NameDescription
RunRequestThe sensor evaluation function can yield one or more run requests. Each run request creates a job run.
SkipReasonIf a sensor evaluation doesn't yield any run requests, it can instead yield a skip reason to log why the evaluation was skipped or why there were no events to be processed.
@asset_sensorThe decorator used to define an asset sensor. The decorated function is an evaluation function that takes in a SensorEvaluationContext and an asset materialization event. The decorator returns an AssetSensorDefinition
AssetSensorDefinitionA special sensor definition class for asset sensors. You almost never want to use initialize this class directly. Instead, you should use the @asset_sensor which returns a AssetSensorDefinition
---

Defining an asset sensor#

An asset sensor checks for new AssetMaterialization events for a particular asset key. This can be used to kick off a job that computes downstream assets or notifies appropriate stakeholders.

One benefit of this pattern is that it enables cross-job and even cross-code-location dependencies. Each job run instigated by an asset sensor is agnostic to the job that caused it.

Dagster provides a special asset sensor definition format for sensors that fire a single RunRequest based on a single asset materialization. Here is an example of a sensor that generates a RunRequest for every materialization for the asset key my_table:

from dagster import (
    AssetKey,
    EventLogEntry,
    RunConfig,
    SensorEvaluationContext,
    asset_sensor,
)


@asset_sensor(asset_key=AssetKey("my_table"), job=my_job)
def my_asset_sensor(context: SensorEvaluationContext, asset_event: EventLogEntry):
    assert asset_event.dagster_event and asset_event.dagster_event.asset_key
    yield RunRequest(
        run_key=context.cursor,
        run_config=RunConfig(
            ops={
                "read_materialization": ReadMaterializationConfig(
                    asset_key=list(asset_event.dagster_event.asset_key.path)
                )
            }
        ),
    )

Testing an asset sensor#

To write a unit test for an asset sensor, you first need to construct a context object using a test instance that can log AssetMaterialization events.

You can then directly invoke your asset sensor with that context and evaluate the results.

from dagster import DagsterInstance, build_sensor_context, materialize


def test_my_asset_sensor():
    @asset
    def my_table():
        return 1

    instance = DagsterInstance.ephemeral()
    ctx = build_sensor_context(instance)

    result = list(my_asset_sensor(ctx))
    assert len(result) == 1
    assert isinstance(result[0], SkipReason)

    materialize([my_table], instance=instance)

    result = list(my_asset_sensor(ctx))
    assert len(result) == 1
    assert isinstance(result[0], RunRequest)