Python packages

This page lists every Python package installed where your pipeline tasks run, with its exact version. Tasks run in the Airflow runtime, which has its own version: a new Databaas version only changes your packages when it ships a new runtime. Pick a runtime to see its list, or find yours under which runtime your Databaas version runs.

The list applies to every Python task: a plain @task, and operators such as PythonOperator. A @task.virtualenv task sees the same packages when it keeps system_site_packages=True, but a package it lists in requirements replaces the one below. See add a package that is not preinstalled.

Databaas runtime

The packages Databaas installs and pins on purpose. They stay available from one runtime version to the next; their versions change only when Databaas upgrades them. See libraries you can use for what each one is for.

Apache Airflow

Airflow itself and its provider packages.

Other packages

Installed because a package above needs them, or as part of the Python environment Airflow ships with. You can import them, but nothing guarantees them: a later runtime version can change their version or drop them when the package that needs them changes. If your pipeline relies on one directly, ask your Databaas team to add it to the Databaas runtime, or pin it in a @task.virtualenv task.

Which runtime your Databaas version runs

Each Databaas version runs one Airflow runtime; several versions can run the same one. Select a runtime to see its packages. A runtime marked unversioned predates runtime version numbers and is named after the build it came from.