props 0.0.3

QuickCheck for Python

Props

Property-based testing for Python à la QuickCheck.

for_all

for_all takes a list of generators (see below) and a property. It then tests the property for arbitrary values of the generators.

Here’s an example testing the commutative and associative properties of ints:

for_all(int, int)(lambda a, b: a + b == b + a)
for_all(int, int)(lambda a, b: a * b == b * a)
for_all(int, int, int)(lambda a, b, c: c * (a + b) == a * c + b * c)

Properties

Properties are functions which take instances of generators and return True if their condition is met:

def prop_associative(a, b, c):
    return a + (b + c) == (a + b) + c

for_all(int, int, int)(prop_associative)
for_all(float, float, float)(prop_associative)  # Warning: float isn't actually associative!

Properties can also fail early by raising AssertionError:

def prop_list_append_pop(list, element):
    if element not in list:
        list.append(element)
        assert element in list
        list.pop()
        return element not in list
    return element in list

for_all(list, int)(prop_list_append_pop)

Generators

Note: These are not the same as Python generators. We should rename them. Generaters? Blech.

A generator is a specification of a set of possible Python objects. A generator is either:

• One of the following built-in types:
  • None, bool, int, float, long, complex, str, tuple, set, list, or dict
• A class that implements the ArbitraryInterface
• Or constructed using the generator combinators.

Combinators

• maybe_a
  • Generates either an arbitrary value of the specified generator or None.
• maybe_an
  • An alias for maybe_a. Provided for syntactic convenience.
• one_of
  • Generates an arbitrary value of one of the specified generators.
• tuple_of
  • Generates a tuple by generating values for each of the specified generators.
• set_of
  • Generates a homogeneous set of the specified generator. You can generate non-homogeneous sets using set.
• list_of
  • Generates a homogeneous list of the specified generator. You can generate non-homogeneous lists using list.
• dict_of
  • Generates a homogeneous dict of the specified generators using kwargs. You can generate non-homogeneous dicts using dict.

arbitrary

arbitrary takes a generator and returns a single instance of the generator.

ArbitraryInterface

We provide a mixin with one classmethod, arbitrary, which raises NotImplementedError. To implement generators for your own classes, please inherit from ArbitraryInterface and provide an implementation for arbitrary.

Here’s an example implementation of a Binary Tree class:

class BinaryTree(ArbitraryInterface):
    ...
    @classmethod
    def arbitrary(cls):
        return arbitrary(one_of(Leaf, Node))

class Leaf(BinaryTree):
    ...
    @classmethod
    def arbitrary(cls):
        return cls(...)  # an instance of Leaf.

class Node(BinaryTree):
    ...
    @classmethod
    def arbitrary(cls):
        return cls(
            ...
            # This is equivalent:
            arbitrary(BinaryTree),
            # to this:
            BinaryTree.arbitrary()
        )  # an instance of Node with two subtrees.

AbstractTestArbitraryInterface

We also provide an AbstractTestArbitraryInterface with you can mixin to your test cases for each class that implements ArbitraryInterface to ensure the arbitrary method is implemented:

class TestBinaryTree(AbstractTestArbitraryInterface,
                     TestCase):
    def setUp(self):
        self.cls = BinaryTree

To Do

• all built in types: http://docs.python.org/2/library/stdtypes.html
• ranges
• import some faker generators for more semantic random values