ensure: Literate assertions in Python

ensure is a set of simple assertion helpers that let you write more expressive, literate, concise, and readable Pythonic code for validating conditions. It’s inspired by should.js, expect.js, and builds on top of the unittest/JUnit assert helpers.

If you use Python 3, you can use ensure to enforce your function signature annotations: see PEP 3107 and the @ensure_annotations decorator below.

Because ensure is fast, is a standalone library (not part of a test framework), doesn’t monkey-patch anything or use DSLs, and doesn’t use the assert statement (which is liable to be turned off with the -O flag), it can be used to validate conditions in production code, not just for testing (though it certainly works as a BDD test utility library).

Aside from better looking code, a big reason to use ensure is that it provides more consistent, readable, and informative error messages when things go wrong. See Motivation and Goals for more.

Installation

pip install ensure

Synopsis

from ensure import ensure

ensure(1).is_an(int)
ensure({1: {2: 3}}).equals({1: {2: 3}}).also.contains(1)
ensure({1: "a"}).has_key(1).whose_value.has_length(1)
ensure.each_of([{1: 2}, {3: 4}]).is_a(dict).of(int).to(int)
ensure(int).called_with("1100101", base=2).returns(101)
ensure(dict).called_with(1, 2).raises(TypeError)
check(1).is_a(float).or_raise(Exception, "An error happened: {msg}. See http://example.com for more information.")

In Python 3:

from ensure import ensure_annotations

@ensure_annotations
def f(x: int, y: float) -> float:
    return x+y

See More examples below.

Notes

The ensure module exports the Ensure class and its convenience instance ensure. Instances of the class are callable, and the call will reset the contents that the instance is inspecting, so you can reuse it for many checks (as seen above).

The class raises EnsureError (a subclass of AssertionError) by default.

There are several ways to chain clauses, depending on the grammatical context: .also, .which, and .whose_value are available per examples below.

Raising custom exceptions

You can pass a callable or exception class as the error_factory keyword argument to Ensure(), or you can use the Check class or its convenience instance check(). This class behaves like Ensure, but does not raise errors immediately. It saves them and chains the methods otherwise(), or_raise() and or_call() to the end of the clauses.

from ensure import check

check("w00t").is_an(int).or_raise(Exception)
check(1).is_a(float).or_raise(Exception, "An error happened: {msg}. See http://example.com for more information.")
check("w00t").is_an(int).or_raise(MyException, 1, 2, x=3, y=4)

def build_fancy_exception(original_exception):
    return MyException(original_exception)

check("w00t").is_an(int).otherwise(build_fancy_exception)
check("w00t").is_an(int).or_call(build_fancy_exception, *args, **kwargs)

More examples

ensure({1: {2: 3}}).is_not_equal_to({1: {2: 4}})
ensure(True).does_not_equal(False)
ensure(1).is_in(range(10))
ensure(True).is_a(bool)
ensure(True).is_(True)
ensure(True).is_not(False)

ensure(["train", "boat"]).contains_one_of(["train"])
ensure(range(8)).contains(5)
ensure(["spam"]).contains_none_of(["eggs", "ham"])
ensure("abcdef").contains_some_of("abcxyz")
ensure("abcdef").contains_one_or_more_of("abcxyz")
ensure("abcdef").contains_all_of("acf")
ensure("abcd").contains_only("dcba")
ensure("abc").does_not_contain("xyz")
ensure([1, 2, 3]).contains_no(float)
ensure(1).is_in(range(10))
ensure("z").is_not_in("abc")
ensure(None).is_not_in([])
ensure(dict).has_attribute('__contains__').which.is_callable()
ensure({1: "a", 2: "b", 3: "c"}).has_keys([1, 2])
ensure({1: "a", 2: "b"}).has_only_keys([1, 2])

ensure(1).is_true()
ensure(0).is_false()
ensure(None).is_none()
ensure(1).is_not_none()
ensure("").is_empty()
ensure([1, 2]).is_nonempty().also.has_length(2)
ensure(1.1).is_a(float).which.equals(1.10)
ensure(KeyError()).is_an(Exception)
ensure({x: str(x) for x in range(5)}).is_a_nonempty(dict).of(int).to(str)
ensure({}).is_an_empty(dict)
ensure(None).is_not_a(list)

import re
ensure("abc").matches("A", flags=re.IGNORECASE)
ensure([1, 2, 3]).is_an_iterable_of(int)
ensure([1, 2, 3]).is_a_list_of(int)
ensure({1, 2, 3}).is_a_set_of(int)
ensure({1: 2, 3: 4}).is_a_mapping_of(int).to(int)
ensure({1: 2, 3: 4}).is_a_dict_of(int).to(int)
ensure({1: 2, 3: 4}).is_a(dict).of(int).to(int)
ensure(10**100).is_numeric()
ensure(lambda: 1).is_callable()
ensure("abc").has_length(3)
ensure("abc").has_length(min=3, max=8)
ensure(1).is_greater_than(0)
ensure(1).exceeds(0)
ensure(0).is_less_than(1)
ensure(1).is_greater_than_or_equal_to(1)
ensure(0).is_less_than_or_equal_to(0)
ensure(1).is_positive()
ensure(1.1).is_a_positive(float)
ensure(-1).is_negative()
ensure(-1).is_a_negative(int)
ensure(0).is_nonnegative()
ensure(0).is_a_nonnegative(int)
ensure([1,2,3]).is_sorted()

ensure("{x} {y}".format).called_with(x=1, y=2).equals("1 2")
ensure(int).called_with("1100101", base=2).returns(101)
ensure("{x} {y}".format).with_args(x=1, y=2).is_a(str)
with ensure().raises(ZeroDivisionError):
    1/0
with ensure().raises_regex(NameError, "'w00t' is not defined"):
    w00t

See complete API documentation.

Enforcing function annotations

Use the @ensure_annotations decorator to enforce function signature annotations:

from ensure import ensure_annotations

@ensure_annotations
def f(x: int, y: float) -> float:
    return x+y

f(1, 2.3)

>>> 3.3

f(1, 2)

>>> ensure.EnsureError: Argument y to <function f at 0x109b7c710> does not match annotation type <class 'float'>

Compare this runtime type checking to compile-time checking in Mypy and type hinting in PEP 484/Python 3.5+.

Motivation and goals

Many BDD assertion libraries suffer from an excess of magic, or end up having to construct statements that don’t parse as English easily. ensure is deliberately kept simple to avoid succumbing to either issue. The source is easy to read and extend.

Work remains to make error messages raised by ensure even more readable, informative, and consistent. Going forward, ability to introspect exceptions to extract structured error information will be a major development focus. You will be in control of how much information is presented in each error, which context it’s thrown from, and what introspection capabilities the exception object will have.

The original use case for ensure is as an I/O validation helper for API endpoints, where the client needs to be sent a very clear message about what went wrong, some structured information (such as an HTTP error code and machine-readable reference to a failing element) may need to be added, and some information may need to be hidden from the client. To further improve on that, we will work on better error translation, marshalling, message formatting, and schema validation helpers.

Authors

• Andrey Kislyuk

• Harrison Metzger

Links

• Project home page (GitHub)

• Documentation

• Package distribution (PyPI)

Bugs

Please report bugs, issues, feature requests, etc. on GitHub.

License

Licensed under the terms of the Apache License, Version 2.0.

API documentation

class ensure.Check(*args, **kwargs)

Like Ensure, but if a check fails, saves the error instead of raising it immediately. The error can then be acted upon using or_raise() or or_call().

.each_of() is not supported by the Check inspector; all other methods are supported.

or_call(_callable, *args, **kwargs)

Calls _callable with supplied args and kwargs if a predicate fails.

or_raise(error_factory, message=None, *args, **kwargs)

Raises an exception produced by error_factory if a predicate fails.

Parameters:

• error_factory – Class or callable (e.g. Exception) which will be invoked to produce the resulting exception. You can define a custom callable here; it will be given the underlying predicate’s exception (AssertError) as the first argument, followed by any arguments passed to or_raise.

• message – String to be formatted and passed as the first argument to error_factory. If this is given, subsequent arguments passed to or_raise will be used to format contents of the string, and will not be passed to error_factory. The keyword argument error will be set to the underlying predicate’s exception.

otherwise(error_factory, message=None, *args, **kwargs)

Raises an exception produced by error_factory if a predicate fails.

Parameters:

• error_factory – Class or callable (e.g. Exception) which will be invoked to produce the resulting exception. You can define a custom callable here; it will be given the underlying predicate’s exception (AssertError) as the first argument, followed by any arguments passed to or_raise.

• message – String to be formatted and passed as the first argument to error_factory. If this is given, subsequent arguments passed to or_raise will be used to format contents of the string, and will not be passed to error_factory. The keyword argument error will be set to the underlying predicate’s exception.

class ensure.Ensure(subject=None, error_factory=<class 'ensure.main.EnsureError'>, catch=<class 'Exception'>)

Constructs a root-level inspector, which can perform a variety of checks (predicates) on subjects passed to it. If the checks do not pass, by default EnsureError is raised. This can be configured by passing the error_factory keyword to the constructor.

Subjects can be passed to the inspector at construction time or by calling the resulting object (each call resets the subject):

Ensure(1).is_an(int)

e = Ensure()
e(1).is_an(int)

Some predicates return child inspectors which can be chained into a series of predicates, for example:

ensure({1: {2: "a"}}).has_key(1).whose_value.is_a(dict).of(int).to(str)

called_with(*args, **kwargs)

Before evaluating subsequent predicates, calls subject with given arguments (but unlike a direct call, catches and transforms any exceptions that arise during the call).

contains(element)

Ensures subject contains other.

contains_all_of(elements)

Ensures subject contains all of elements, which must be an iterable.

contains_key(key)

Ensures subject is a collections.Mapping and contains key.

contains_no(prototype)

Ensures no item of subject is of class prototype.

contains_none_of(elements)

Ensures subject contains none of elements, which must be an iterable.

contains_one_of(elements)

Ensures subject contains exactly one of elements, which must be an iterable.

contains_one_or_more_of(elements)

Ensures subject contains at least one of elements, which must be an iterable.

contains_only(elements)

Ensures subject contains all of elements, which must be an iterable, and no other items.

contains_some_of(elements)

Ensures subject contains at least one of elements, which must be an iterable.

does_not_contain(element)

Ensures subject does not contain element.

does_not_equal(other)

Ensures subject is not equal to other.

classmethod each_of(iterable)

Applies subsequent predicate(s) to all items in iterable.

equals(other)

Ensures subject is equal to other.

exceeds(other)

Ensures subject is greater than other.

has_attribute(attr)

Ensures subject has an attribute attr.

has_key(key)

Ensures subject is a collections.Mapping and contains key.

has_keys(keys)

Ensures subject is a collections.Mapping and contains keys, which must be an iterable.

has_length(length=None, min=None, max=None)

Ensures subject has length length (if given), length at least min (if given), and length at most max (if given).

has_only_keys(keys)

Ensures subject is a collections.Mapping and contains keys, and no other keys.

hasattr(attr)

Ensures subject has an attribute attr.

is_(other)

Ensures subject is other (object identity check).

is_a(prototype)

Ensures subject is an object of class prototype.

is_a_dict_of(prototype)

Ensures subject is a dict containing only objects of class prototype.

is_a_list_of(prototype)

Ensures subject is a list containing only objects of class prototype.

is_a_mapping_of(prototype)

Ensures subject is a collections.Mapping containing only objects of class prototype.

is_a_negative(prototype)

Ensures subject is less than 0 and is an object of class prototype.

is_a_nonempty(prototype)

Ensures subject is an object of class prototype and has non-zero length.

is_a_nonnegative(prototype)

Ensures subject is greater than or equal to 0 and is an object of class prototype.

is_a_positive(prototype)

Ensures subject is greater than 0 and is an object of class prototype.

is_a_set_of(prototype)

Ensures subject is a set containing only objects of class prototype.

is_an(prototype)

Ensures subject is an object of class prototype.

is_an_empty(prototype)

Ensures subject is an object of class prototype and has zero length.

is_an_instance_of(prototype)

Ensures subject is an object of class prototype.

is_an_iterable_of(prototype)

Ensures subject is an iterable containing only objects of class prototype.

is_callable()

Ensures subject is a callable.

is_empty()

Ensures subject has length zero.

is_false()

Ensures subject is False.

is_greater_than(other)

Ensures subject is greater than other.

is_greater_than_or_equal_to(other)

Ensures subject is greater than or equal to other.

is_in(iterable)

Ensures subject is contained in iterable.

is_less_than(other)

Ensures subject is less than other.

is_less_than_or_equal_to(other)

Ensures subject is less than or equal to other.

is_negative()

Ensures subject is less than 0.

is_none()

Ensures subject is None.

property is_none_or

Ensures subject is either None, or satisfies subsequent (chained) conditions:

Ensure(None).is_none_or.is_an(int)

is_nonempty()

Ensures subject has non-zero length.

is_nonnegative()

Ensures subject is greater than or equal to 0.

is_not(other)

Ensures subject is not other (object identity check).

is_not_a(prototype)

Ensures subject is not an object of class prototype.

is_not_equal_to(other)

Ensures subject is not equal to other.

is_not_in(iterable)

Ensures subject is not contained in iterable.

is_not_none()

Ensures subject is not None.

is_numeric()

Ensures subject is an int, float, or long.

is_positive()

Ensures subject is greater than 0.

is_true()

Ensures subject is True.

matches(pattern, flags=0)

Ensures subject matches regular expression pattern.

not_a(prototype)

Ensures subject is not an object of class prototype.

not_in(iterable)

Ensures subject is not contained in iterable.

raises(expected_exception)

Ensures preceding predicates (specifically, called_with()) result in expected_exception being raised.

raises_regex(expected_exception, expected_regexp)

Ensures preceding predicates (specifically, called_with()) result in expected_exception being raised, and the string representation of expected_exception must match regular expression expected_regexp.

with_args(*args, **kwargs)

Before evaluating subsequent predicates, calls subject with given arguments (but unlike a direct call, catches and transforms any exceptions that arise during the call).

exception ensure.EnsureError

class ensure.IntegerByteString

class ensure.IntegerString

class ensure.NumericByteString

class ensure.NumericString

ensure.ensure_annotations(f)

Decorator to be used on functions with annotations. Runs type checks to enforce annotations. Raises EnsureError if any argument passed to f is not of the type specified by the annotation. Also raises EnsureError if the return value of f is not of the type specified by the annotation. Examples:

from ensure import ensure_annotations

@ensure_annotations
def f(x: int, y: float) -> float:
    return x+y

print(f(1, y=2.2))

>>> 3.2

print(f(1, y=2))

>>> ensure.EnsureError: Argument y of type <class 'int'> to
<function f at 0x109b7c710> does not match annotation type <class 'float'>

Warning

The current implementation of the decorator might trigger a _pickle.PicklingError when a decorated function is run through multiprocessing.

ensure.ensure_raises(expected_exception, *args, **kwargs)

Fail unless an exception of class expected_exception is raised by the callable when invoked with specified positional and keyword arguments. If a different type of exception is raised, it will not be caught, and the test case will be deemed to have suffered an error, exactly as for an unexpected exception.

If called with the callable and arguments omitted, will return a context object used like this:

with self.assertRaises(SomeException):
    do_something()

An optional keyword argument ‘msg’ can be provided when assertRaises is used as a context object.

The context manager keeps a reference to the exception as the ‘exception’ attribute. This allows you to inspect the exception after the assertion:

with self.assertRaises(SomeException) as cm:
    do_something()
the_exception = cm.exception
self.assertEqual(the_exception.error_code, 3)

ensure.ensure_raises_regex(expected_exception, expected_regex, *args, **kwargs)

Asserts that the message in a raised exception matches a regex.

Args:

expected_exception: Exception class expected to be raised. expected_regex: Regex (re.Pattern object or string) expected

to be found in error message.

args: Function to be called and extra positional args. kwargs: Extra kwargs. msg: Optional message used in case of failure. Can only be used

when assertRaisesRegex is used as a context manager.

ensure.ensure_annotations(f)

Decorator to be used on functions with annotations. Runs type checks to enforce annotations. Raises EnsureError if any argument passed to f is not of the type specified by the annotation. Also raises EnsureError if the return value of f is not of the type specified by the annotation. Examples:

from ensure import ensure_annotations

@ensure_annotations
def f(x: int, y: float) -> float:
    return x+y

print(f(1, y=2.2))

>>> 3.2

print(f(1, y=2))

>>> ensure.EnsureError: Argument y of type <class 'int'> to
<function f at 0x109b7c710> does not match annotation type <class 'float'>

Warning

The current implementation of the decorator might trigger a _pickle.PicklingError when a decorated function is run through multiprocessing.

Table of Contents

• Index

• Module Index

• Search Page