Alternative syntax for comparing/asserting expressions in Python. Supports pluggable matchers for custom comparisons.
Compare is a compact utility that provides an alternative, expressive syntax for comparing data values. Have you been looking for an escape from the stale XUnit style asserts that plague the omni-present unittest? You may have found just what you need. I invite you to take a look at this little utility. If it fits your style, you may use it as a drop-in replacement for the “self.assert…” style of doing assertions in python.
The compare API exposes the expect construct which allows you to compare values with readable and extensible syntax. It was designed to be a stand-alone alternative assertion syntax. As such you may use it as-is with your favorite testing/specification framework.
Project source: https://github.com/rudylattae/compare
PyPI page: http://pypi.python.org/pypi/compare
The core implementation of compare is a single file module with no additional requirements beyond the Python Standard Library.
The simplest and recommended way to install compare is with Pip. You may install the latest stable release from PyPI with pip:
> pip install compare
If you do not have pip, you may use easy_install:
> easy_install compare
Alternatively, you may download the source package from the compare page on PyPI, extract it and install it using:
> python setup.py install
If you wish, you may grab the in development (cutting-edge but unstable) version compare.py from the project repository and put it into your project directory.
When you install the package, you get the “expect” starter, a simple function that allows you to compare two values and fail if the outcome does not meet your expectation. This starter has extensible matchers that enable you to describe the expected outcome in a pythonic BDD manner.
Compare shines brightest when you are crafting executable specifications for your software. It helps you maintain your flow of thought without succumbing to test-focused non-pythonic distrations like “self.assertEqual(s)…”, “self.assertTrue”, etc.
Here is a trivial example of the readability you gain when you employ the “expect” construct in your specs.
> cat hello.py:
greeting = 'Hello you'
> cat hello_specs.py:
from compare import expect import hello expect(hello.greeting).to_equal('Hello you')
If you define an expectation that is not met, you will get an “Unmet Expectation” error which inherits from the python AssertionError so it is compatible with the usual unittest tools. Here is an example of such an error:
>>> from compare import expect >>> opts = ['foo', 'bar', 'baz'] >>> expect(opts).to_contain('BAT') Traceback (most recent call last): ... UnmetExpectation: Expected ['foo', 'bar', 'baz'] to contain 'BAT'
The expect syntax does not yet have a clean way to negate a matcher. This feature is planned for the next release. An example of the anticipated usage:
expect(['a', 'c', 'd']).NOT.to_contain('b')
The to_return matcher does not accept any parameters to pass to the callable.
Matchers do not accept custom fail messages.
I welcome any questions or feedback about bugs and suggestions on how to improve compare. Let me know what you think about compare. I am on twitter @RudyLattae . I appreciate constructive criticsms or high fives :)
Do you have suggestions for improvement? Then please create an issue with details of what you would like to see. I’ll take a look at it and work with you to either kill the idea or implement it.