cattrs 1.2.0

Composable complex class support for attrs.

cattrs

---

cattrs is an open source Python library for structuring and unstructuring data. cattrs works best with attrs classes and the usual Python collections, but other kinds of classes are supported by manually registering converters.

Python has a rich set of powerful, easy to use, built-in data types like dictionaries, lists and tuples. These data types are also the lingua franca of most data serialization libraries, for formats like json, msgpack, yaml or toml.

Data types like this, and mappings like dict s in particular, represent unstructured data. Your data is, in all likelihood, structured: not all combinations of field names are values are valid inputs to your programs. In Python, structured data is better represented with classes and enumerations. attrs is an excellent library for declaratively describing the structure of your data, and validating it.

When you’re handed unstructured data (by your network, file system, database…), cattrs helps to convert this data into structured data. When you have to convert your structured data into data types other libraries can handle, cattrs turns your classes and enumerations into dictionaries, integers and strings.

Here’s a simple taste. The list containing a float, an int and a string gets converted into a tuple of three ints.

>>> import cattr
>>> from typing import Tuple
>>>
>>> cattr.structure([1.0, 2, "3"], Tuple[int, int, int])
(1, 2, 3)

cattrs works well with attrs classes out of the box.

>>> import attr, cattr
>>>
>>> @attr.s(slots=True, frozen=True)  # It works with normal classes too.
... class C:
...     a = attr.ib()
...     b = attr.ib()
...
>>> instance = C(1, 'a')
>>> cattr.unstructure(instance)
{'a': 1, 'b': 'a'}
>>> cattr.structure({'a': 1, 'b': 'a'}, C)
C(a=1, b='a')

Here’s a much more complex example, involving attrs classes with type metadata.

>>> from enum import unique, Enum
>>> from typing import List, Optional, Sequence, Union
>>> from cattr import structure, unstructure
>>> import attr
>>>
>>> @unique
... class CatBreed(Enum):
...     SIAMESE = "siamese"
...     MAINE_COON = "maine_coon"
...     SACRED_BIRMAN = "birman"
...
>>> @attr.s
... class Cat:
...     breed: CatBreed = attr.ib()
...     names: Sequence[str] = attr.ib()
...
>>> @attr.s
... class DogMicrochip:
...     chip_id = attr.ib()
...     time_chipped: float = attr.ib()
...
>>> @attr.s
... class Dog:
...     cuteness: int = attr.ib()
...     chip: Optional[DogMicrochip] = attr.ib()
...
>>> p = unstructure([Dog(cuteness=1, chip=DogMicrochip(chip_id=1, time_chipped=10.0)),
...                  Cat(breed=CatBreed.MAINE_COON, names=('Fluffly', 'Fluffer'))])
...
>>> print(p)
[{'cuteness': 1, 'chip': {'chip_id': 1, 'time_chipped': 10.0}}, {'breed': 'maine_coon', 'names': ('Fluffly', 'Fluffer')}]
>>> print(structure(p, List[Union[Dog, Cat]]))
[Dog(cuteness=1, chip=DogMicrochip(chip_id=1, time_chipped=10.0)), Cat(breed=<CatBreed.MAINE_COON: 'maine_coon'>, names=['Fluffly', 'Fluffer'])]

Consider unstructured data a low-level representation that needs to be converted to structured data to be handled, and use structure. When you’re done, unstructure the data to its unstructured form and pass it along to another library or module. Use attrs type metadata to add type metadata to attributes, so cattrs will know how to structure and destructure them.

• Free software: MIT license

• Documentation: https://cattrs.readthedocs.io.

• Python versions supported: 3.7 and up. (Older Python versions, like 2.7, 3.5 and 3.6 are supported by older versions; see the changelog.)

Features

• Converts structured data into unstructured data, recursively:

  • attrs classes are converted into dictionaries in a way similar to attr.asdict, or into tuples in a way similar to attr.astuple.

  • Enumeration instances are converted to their values.

  • Other types are let through without conversion. This includes types such as integers, dictionaries, lists and instances of non-attrs classes.

  • Custom converters for any type can be registered using register_unstructure_hook.

• Converts unstructured data into structured data, recursively, according to your specification given as a type. The following types are supported:

  • typing.Optional[T].

  • typing.List[T], typing.MutableSequence[T], typing.Sequence[T] (converts to a list).

  • typing.Tuple (both variants, Tuple[T, ...] and Tuple[X, Y, Z]).

  • typing.MutableSet[T], typing.Set[T] (converts to a set).

  • typing.FrozenSet[T] (converts to a frozenset).

  • typing.Dict[K, V], typing.MutableMapping[K, V], typing.Mapping[K, V] (converts to a dict).

  • attrs classes with simple attributes and the usual __init__.

    • Simple attributes are attributes that can be assigned unstructured data, like numbers, strings, and collections of unstructured data.

  • All attrs classes with the usual __init__, if their complex attributes have type metadata.

  • typing.Union s of supported attrs classes, given that all of the classes have a unique field.

  • typing.Union s of anything, given that you provide a disambiguation function for it.

  • Custom converters for any type can be registered using register_structure_hook.

Credits

Major credits to Hynek Schlawack for creating attrs and its predecessor, characteristic.

cattrs is tested with Hypothesis, by David R. MacIver.

cattrs is benchmarked using perf, by Victor Stinner.

This package was created with Cookiecutter and the audreyr/cookiecutter-pypackage project template.

History

1.2.0 (2021-01-31)

• converter.unstructure now supports an optional parameter, unstructure_as, which can be used to unstructure something as a different type. Useful for unions.

• Improve support for union un/structuring hooks. Flesh out docs for advanced union handling. (#115)

• Fix GenConverter behavior with inheritance hierarchies of attrs classes. (#117) (#116)

• Refactor GenConverter.un/structure_attrs_fromdict into GenConverter.gen_un/structure_attrs_fromdict to allow calling back to Converter.un/structure_attrs_fromdict without sideeffects. (#118)

1.1.2 (2020-11-29)

• The default disambiguator will not consider non-required fields any more. (#108)

• Fix a couple type annotations. (#107) (#105)

• Fix a GenConverter unstructuring issue and tests.

1.1.1 (2020-10-30)

• Add metadata for supported Python versions. (#103)

1.1.0 (2020-10-29)

• Python 2, 3.5 and 3.6 support removal. If you need it, use a version below 1.1.0.

• Python 3.9 support, including support for built-in generic types (list[int] vs typing.List[int]).

• cattrs now includes functions to generate specialized structuring and unstructuring hooks. Specialized hooks are faster and support overrides (omit_if_default and rename). See the cattr.gen module.

• cattrs now includes a converter variant, cattr.GenConverter, that automatically generates specialized hooks for attrs classes. This converter will become the default in the future.

• Generating specialized structuring hooks now invokes attr.resolve_types on a class if the class makes use of the new PEP 563 annotations.

• cattrs now depends on attrs >= 20.1.0, because of attr.resolve_types.

• Specialized hooks now support generic classes. The default converter will generate and use a specialized hook upon encountering a generic class.

1.0.0 (2019-12-27)

• attrs classes with private attributes can now be structured by default.

• Structuring from dictionaries is now more lenient: extra keys are ignored.

• cattrs has improved type annotations for use with Mypy.

• Unstructuring sets and frozensets now works properly.

0.9.1 (2019-10-26)

• Python 3.8 support.

0.9.0 (2018-07-22)

• Python 3.7 support.

0.8.1 (2018-06-19)

• The disambiguation function generator now supports unions of attrs classes and NoneType.

0.8.0 (2018-04-14)

• Distribution fix.

0.7.0 (2018-04-12)

• Removed the undocumented Converter.unstruct_strat property setter.

• Removed the ability to set the Converter.structure_attrs instance field.

  As an alternative, create a new Converter::

  
  

  .. code-block:: python

  
  

  >>> converter = cattr.Converter(unstruct_strat=cattr.UnstructureStrategy.AS_TUPLE)

• Some micro-optimizations were applied; a structure(unstructure(obj)) roundtrip is now up to 2 times faster.

0.6.0 (2017-12-25)

• Packaging fixes. (#17)

0.5.0 (2017-12-11)

• structure/unstructure now supports using functions as well as classes for deciding the appropriate function.

• added Converter.register_structure_hook_func, to register a function instead of a class for determining handler func.

• added Converter.register_unstructure_hook_func, to register a function instead of a class for determining handler func.

• vendored typing is no longer needed, nor provided.

• Attributes with default values can now be structured if they are missing in the input. (#15)

• Optional attributes can no longer be structured if they are missing in the input.

  In other words, this no longer works:

  
  

  .. code-block:: python

  
  

  @attr.s

  class A:

  a: Optional[int] = attr.ib()

  
  

  >>> cattr.structure({}, A)

  
  

• cattr.typed removed since the functionality is now present in attrs itself. Replace instances of cattr.typed(type) with attr.ib(type=type).

0.4.0 (2017-07-17)

• Converter.loads is now Converter.structure, and Converter.dumps is now Converter.unstructure.

• Python 2.7 is supported.

• Moved cattr.typing to cattr.vendor.typing to support different vendored versions of typing.py for Python 2 and Python 3.

• Type metadata can be added to attrs classes using cattr.typed.

0.3.0 (2017-03-18)

• Python 3.4 is no longer supported.

• Introduced cattr.typing for use with Python versions 3.5.2 and 3.6.0.

• Minor changes to work with newer versions of typing.

  • Bare Optionals are not supported any more (use Optional[Any]).

• Attempting to load unrecognized classes will result in a ValueError, and a helpful message to register a loads hook.

• Loading attrs classes is now documented.

• The global converter is now documented.

• cattr.loads_attrs_fromtuple and cattr.loads_attrs_fromdict are now exposed.

0.2.0 (2016-10-02)

• Tests and documentation.

0.1.0 (2016-08-13)

• First release on PyPI.