Reference

Components

Package	Role
airflow-pydantic	DAG/task models, rendering, and Airflow compatibility imports
airflow-config	Hydra/YAML composition and DAG generation
airflow-common	Shared operators and task helpers
airflow-balancer	Host/port selection and configuration viewer
airflow-ha	HA cycles, retries, and execution limits
airflow-priority	DAG-run listener routing to alert backends
airflow-supervisor	Supervisor workload integration
airflow-nomad	Nomad jobs, watchdogs, and allocation logs
airflow-cron	Cron configuration converted to Airflow DAGs
supervisor-pydantic	Supervisor configuration and RPC client
nomad-pydantic	Nomad job models and client tools
cron-pydantic	Cron configuration models

The airflow_laminar module exposes __version__. Import tasks and models from the component packages.

Dependency bounds

Dependency	Bound
airflow-balancer	>=0.7.11,<0.8
airflow-common	>=0.7.10,<0.8
airflow-config	>=1.12.5,<1.13
airflow-cron	>=0.1.1,<0.2
airflow-ha	>=1.6.5,<1.7
airflow-nomad	>=0.2,<0.3
airflow-priority	>=1.6.1,<1.7
airflow-pydantic	>=1.7.0,<1.8
airflow-supervisor	>=1.10.5,<1.12
cron-pydantic	>=0.1.0,<0.2
nomad-pydantic	>=0.2.0,<0.3
supervisor-pydantic	>=1.4.3,<1.6

Compatibility

airflow-laminar requires Python >=3.11 and installs all components listed above. Install Apache Airflow through a component’s airflow or airflow3 extra, or use an existing Airflow environment.

Environment	Installation path	Verification
Configuration loading and source generation	airflow-laminar	Models support generation without a running scheduler or managed workload.
Airflow 2	airflow-laminar plus a component’s airflow extra	Examples checked on Python 3.11 / Airflow 2.11.2.
Airflow 3	airflow-laminar plus a component’s airflow3 extra	Examples checked on Python 3.11 / Airflow 3.3.2.

The Balancer component extras select Airflow >=2.8,<3 or >=3,<3.4 and install SSH/standard providers. Component packages declare their provider requirements and supported versions. The checks above cover installation and DAG construction; running workloads requires the services listed below.

Runtime requirements

Integration	External runtime
Configuration/model rendering	Python packages; a running Airflow scheduler is not needed.
Cron conversion	Airflow executes commands on its task worker. No system cron daemon is needed for these generated DAGs.
Local Supervisor	Supervisor tools/processes on the managed host, with access to the configured working directory and RPC endpoint.
SSH Supervisor	SSH access and Supervisor on the remote managed host.
Nomad	Nomad CLI and a reachable Nomad agent/cluster with an eligible client and the job’s task driver.
Priority delivery	The chosen backend SDK and receiver credentials.

A Nomad development agent runs client and server roles in one process. The Nomad tutorial includes that setup and the task driver used by its example.

Entry points

The meta-package has no command-line entry points. Component commands include airflow-config and airflow-balancer-viewer; each SDK documents its own commands.