attrs

attrs is a Python library that provides a lightweight, declarative way to define classes with attributes. It focuses on reducing boilerplate by automatically generating common methods and by offering attribute-level features such as validation, conversion, and metadata. The project predates Python’s standard library dataclasses and remains popular for its rich customization options.

A central concept is the attr.s decorator used to define a class and attr.ib for individual attributes. Attributes can have default values, default factories, converters that transform input, and validators that enforce constraints. The auto_attribs feature allows defining attributes with Python type hints, while slot and frozen options enable memory efficiency and immutability. Generated methods typically include __init__, __repr__, and __eq__, with optional ordering and custom hashing behavior.

Validators and converters provide a flexible mechanism for data integrity. Built-in validators cover common checks (e.g., instance_of, in_, matches). Converters transform incoming values (for example, parsing strings to numbers) before storage. Attrs also supports metadata and metadata-driven tooling, enabling integration with other libraries and tooling.

Additional utilities include asdict and astuple for deep serialization, and evolve for creating modified copies of existing instances. The library emphasizes readability and explicitness, making class definitions clearer and easier to maintain.

Patterns and ecosystem: attrs is widely used in Python projects as an alternative to dataclasses when advanced attribute handling is needed. It remains compatible with multiple Python versions and can interoperate with typing and other data-validation libraries.