Run your dbt Core projects as Apache Airflow® DAGs and Task Groups with a few lines of code. Benefits include:

Run dbt projects against Airflow connections instead of dbt profiles
Native support for installing and running dbt in a virtual environment to avoid dependency conflicts with Airflow
Run tests immediately after a model is done to catch issues early
Utilize Airflow's data-aware scheduling to run models immediately after upstream ingestion
Turn each dbt model into a task/task group complete with retries, alerting, etc.

Quickstart

Check out the Getting Started guide on our docs. See more examples at /dev/dags and at the cosmos-demo repo.

Example Usage

You can render a Cosmos Airflow DAG using the DbtDag class. Here's an example with the jaffle_shop project:

This will generate an Airflow DAG that looks like this:

We follow Semantic Versioning for releases. Check CHANGELOG.rst for the latest changes.

All contributions, bug reports, bug fixes, documentation improvements, enhancements are welcome.

A detailed overview an how to contribute can be found in the Contributing Guide.

As contributors and maintainers to this project, you are expected to abide by the Contributor Code of Conduct.