Implementation:Astronomer Astronomer cosmos Get Airflow Task

Knowledge Sources	Astronomer_Astronomer_cosmos
Domains	Orchestration, Task_Conversion
Last Updated	2026-02-07 17:00 GMT

Overview

get_airflow_task converts a Cosmos Task entity into a native Airflow BaseOperator, bridging the Cosmos graph model with the Airflow execution engine.

Description

The get_airflow_task function is the primary translation layer between the Cosmos internal task representation and Apache Airflow's operator model. It accepts a Cosmos Task data class, along with the target DAG and an optional TaskGroup, and returns a fully configured BaseOperator instance that Airflow can schedule and execute.

Internally the function inspects the operator_class attribute on the Task entity, instantiates the corresponding Airflow operator, and wires up any arguments and extra context that were attached to the task during graph construction. When a TaskGroup is provided the resulting operator is registered within that group, enabling hierarchical DAG visualisation in the Airflow UI.

This function sits at line 25 of a compact 55-line module, reflecting its focused responsibility as a thin adapter rather than a complex transformation pipeline.

Usage

Use get_airflow_task whenever you need to materialise a Cosmos DAG graph into runnable Airflow tasks. This is typically called during the DAG rendering phase, after the Cosmos graph of Task and Group entities has been fully constructed. It is essential for any custom rendering strategy that converts Cosmos entities into Airflow-native objects.

Code Reference

Source Location

• Repository: Astronomer_Astronomer_cosmos
• File: cosmos/core/airflow.py
• Lines: 25 onward (module is 55 lines total)

Signature

def get_airflow_task(
    task: Task,
    dag: DAG,
    task_group: TaskGroup | None = None,
) -> BaseOperator:
    ...

Import

from cosmos.core.airflow import get_airflow_task

I/O Contract

Inputs

Name	Type	Required	Description
task	Task	Yes	A Cosmos Task entity containing operator_class, arguments, and extra_context
dag	DAG	Yes	The Airflow DAG instance to which the resulting operator will be attached
task_group	TaskGroup or None	No	An optional Airflow TaskGroup for nesting the operator within a visual group

Outputs

Name	Type	Description
operator	BaseOperator	A fully configured Airflow operator instance ready for scheduling and execution

Usage Examples

from airflow import DAG
from cosmos.core.graph.entities import Task
from cosmos.core.airflow import get_airflow_task

with DAG(dag_id="example_cosmos_dag") as dag:
    cosmos_task = Task(
        id="run_model",
        operator_class="cosmos.operators.local.DbtRunLocalOperator",
        arguments={"models": "my_model"},
    )

    airflow_operator = get_airflow_task(task=cosmos_task, dag=dag)

from airflow.utils.task_group import TaskGroup
from cosmos.core.airflow import get_airflow_task

with DAG(dag_id="grouped_dag") as dag:
    with TaskGroup(group_id="dbt_models") as tg:
        for task_entity in cosmos_group.entities:
            get_airflow_task(task=task_entity, dag=dag, task_group=tg)

Related Pages

• Environment:Astronomer_Astronomer_cosmos_Python_Airflow_Runtime
• Astronomer_Astronomer_cosmos_Graph_Entities -- defines the Task data class consumed by this function