Implementation:Apache Airflow Core Pyproject

Knowledge Sources	Apache_Airflow
Domains	Build_System, Packaging
Last Updated	2026-02-08 21:00 GMT

Overview

Package configuration for apache-airflow-core defining the build system (hatchling), 80+ runtime dependencies, optional extras, version 3.2.0, and Python >=3.10 requirement.

Description

The airflow-core/pyproject.toml file is a 325-line TOML configuration that defines the apache-airflow-core package -- the core scheduling, API server, and execution engine for Apache Airflow. Key aspects:

• Build system: Uses Hatchling (v1.27.0) as the build backend with pinned build dependencies including GitPython, packaging, pathspec, pluggy, and trove-classifiers.
• Python version: Requires >=3.10,!=3.14 (Python 3.14 proactively excluded pending dependency support).
• Version: 3.2.0, automatically synchronized from src/airflow/__init__.py via a prek hook.
• Runtime dependencies: Over 80 direct dependencies including:
  • Web framework: fastapi[standard-no-fastapi-cloud-cli], uvicorn, starlette
  • Database: sqlalchemy[asyncio]>=2.0.36, alembic>=1.13.1, aiosqlite
  • Scheduling: croniter, pendulum>=3.1.0
  • Serialization: pydantic>=2.11.0, dill, jsonschema
  • Observability: opentelemetry-api, opentelemetry-exporter-otlp, structlog>=25.4.0
  • Pre-installed providers: common-compat, common-io, common-sql, smtp, standard
  • Task SDK: apache-airflow-task-sdk<1.3.0,>=1.2.0
• Optional extras: async (eventlet, gevent, greenlet), graphviz, kerberos, memray, gunicorn, otel, statsd, and all (combines graphviz, gunicorn, kerberos, otel, statsd).
• Entry point: airflow = "airflow.__main__:main" as the CLI command.
• Build targets: Configures sdist and wheel targets with shared distribution force-includes for configuration, logging, secrets, timezones, and other shared modules.

Usage

# Install apache-airflow-core with all extras
pip install apache-airflow-core[all]

# Install with specific extras
pip install apache-airflow-core[async,kerberos]

# Install base package only
pip install apache-airflow-core

Code Reference

Source Location

• Repository: Apache_Airflow
• File: airflow-core/pyproject.toml
• Lines: 325

Signature

[build-system]
requires = [
    "GitPython==3.1.45",
    "hatchling==1.27.0",
    "packaging==25.0",
    # ...
]
build-backend = "hatchling.build"

[project]
name = "apache-airflow-core"
description = "Core packages for Apache Airflow, schedule and API server"
license = "Apache-2.0"
requires-python = ">=3.10,!=3.14"
version = "3.2.0"

dependencies = [
    "fastapi[standard-no-fastapi-cloud-cli]>=0.128.1,<0.128.4",
    "sqlalchemy[asyncio]>=2.0.36",
    "alembic>=1.13.1, <2.0",
    "pendulum>=3.1.0",
    "pydantic>=2.11.0",
    "apache-airflow-task-sdk<1.3.0,>=1.2.0",
    # ... 80+ dependencies
]

[project.optional-dependencies]
"async" = ["eventlet>=0.37.0", "gevent>=25.4.1", "greenlet>=3.1.0", "greenback>=1.2.1"]
"graphviz" = ["graphviz>=0.20; sys_platform != 'darwin'"]
"kerberos" = ["pykerberos>=1.1.13", "requests-kerberos>=0.14.0", "thrift-sasl>=0.4.2"]
"all" = ["apache-airflow-core[graphviz,gunicorn,kerberos,otel,statsd]"]

[project.scripts]
airflow = "airflow.__main__:main"

[tool.hatch.version]
path = "src/airflow/__init__.py"

[tool.hatch.build.targets.sdist.force-include]
"../shared/configuration/src/airflow_shared/configuration" = "src/airflow/_shared/configuration"
"../shared/logging/src/airflow_shared/logging" = "src/airflow/_shared/logging"
# ... other shared distributions

I/O Contract

Inputs

Name	Type	Required	Description
Source code	Python modules	Yes	Python source files under airflow-core/src/airflow/
Shared distributions	Python packages	Yes	Shared library code force-included from ../shared/ directories
UI assets	Built JavaScript/CSS	No	Pre-compiled UI assets in src/airflow/ui/dist/
Version file	src/airflow/__init__.py	Yes	Contains __version__ used as package version

Outputs

Name	Type	Description
sdist	.tar.gz	Source distribution archive including shared distributions
wheel	.whl	Built distribution with compiled UI assets and shared code
airflow CLI	Console script	Entry point for the airflow command-line tool

Usage Examples

Install With Extras

# Install core with all optional extras
pip install apache-airflow-core[all]

# Install with specific extras for async support
pip install apache-airflow-core[async]

# Install with Kerberos authentication support
pip install apache-airflow-core[kerberos]

Development Installation

# Install in development mode with uv
cd airflow-core
uv pip install -e ".[all]"

Optional Extras Summary

Extra	Dependencies	Description
async	eventlet, gevent, greenlet, greenback	Async execution support
graphviz	graphviz (non-macOS)	DAG visualization rendering
kerberos	pykerberos, requests-kerberos, thrift-sasl	Kerberos authentication
memray	memray	Memory profiling
gunicorn	gunicorn	WSGI HTTP server
otel	opentelemetry-exporter-prometheus	Prometheus metrics export
statsd	statsd	StatsD metrics client
all	(combines above except async and memray)	All standard extras

Related Pages

• Environment:Apache_Airflow_Python_Runtime_Environment