Introduction to PyContracts¶
PyContracts is a Python package that allows to declare constraints on function parameters and return values. It supports a basic type system, variables binding, arithmetic constraints, and has several specialized contracts (notably for Numpy arrays).
As a quick intro, please see this presentation about PyContracts.
Why: The purpose of PyContracts is not to turn Python into a statically-typed language
(albeit you can be as strict as you wish), but, rather, to avoid the time-consuming and
obfuscating checking of various preconditions. In fact, more than the type constraints, I found
useful the ability to impose value and size constraints. For example, “I need a list of at least
3 positive numbers” can be expressed as
list[>=3](number, >0)). If you find that
PyContracts is overkill for you, you might want to try a simpler alternative, such as
typecheck. If you find that PyContracts is not enough for you, you probably want to be
using Haskell instead of Python.
Specifying contracts: Contracts can be specified in three ways:
Using the ``@contract`` decorator:
@contract(a='int,>0', b='list[N],N>0', returns='list[N]') def my_function(a, b): ...
Using annotations (for Python 3):
@contract def my_function(a : 'int,>0', b : 'list[N],N>0') -> 'list[N]': # Requires b to be a nonempty list, and the return # value to have the same length. ...
Using docstrings, with the
@contract def my_function(a, b): """ Function description. :type a: int,>0 :type b: list[N],N>0 :rtype: list[N] """ ...
Deployment: In production, all checks can be disabled using the function
contracts.disable_all(), so the performance hit is 0.
Extensions: You can extend PyContracts with new contracts types:
new_contract('valid_name', lambda s: isinstance(s, str) and len(s)>0) @contract(names='dict(int: (valid_name, int))') def process_accounting(records): ...
Any Python type is a contract:
@contract(a=int, # simple contract b='int,>0' # more complicated ) def f(a, b): ...
ContractsMeta is a metaclass,
like ABCMeta, which propagates contracts to the subclasses:
from contracts import contract, ContractsMeta, with_metaclass class Base(with_metaclass(ContractsMeta, object)): @abstractmethod @contract(probability='float,>=0,<=1') def sample(self, probability): pass class Derived(Base): # The contract above is automatically enforced, # without this class having to know about PyContracts at all! def sample(self, probability): ....
Numpy: There is special support for Numpy:
@contract(image='array[HxWx3](uint8),H>10,W>10') def recolor(image): ...
Status: The syntax is stable and it won’t be changed. PyContracts is very well tested on Python 2.x.
Status on Python 3.x: We reached feature parity! Everything works on Python 3 now.
- Chris Beaumont (Harvard-Smithsonian Center for Astrophysics):
$varsyntax; kwargs/args for extensions.
- Brett Graham (Rowland Institute at Harvard University):
attr(name:type)syntax for checking types of attributes.
- William Furr: bug reports and performance improvements
- Karol Kuczmarski (Google Zurich): implementation of “string” and “unicode” contracts
- Maarten Derickx (Leiden U.): documentation fixes
- Calen Pennington (EdX): disabling checks inside check() function.
- Adam Palay (EdX): implementation of environment variable enabling/disabling override.
- Ryan Heimbuch: bug reports
- Bernhard Biskup: bug reports
- asharp: bug fixes
- Dennis Kempin (Google mothership): Sphinx-style constraints specs
- Andy Hayden: Python 3 support
(Please let me know if I forgot anybody.)