Python library that makes testing refactored code super simple. Inspired by scientist from github. The goal of seamus is to make testing of refactored code easy. Let's say that you decide to refactor a method in your code. But if your code is running in production environment, it is too risky to roll out the refactored code directly. seamus comes to rescue here. You can use it with plenty of options offered to strategically test your refactored code in background leaving your end users unaffected. Once you are sure that the refactored code is behaving as expected, then you can roll it out.
- Python >= 3.2
seamus can be installed via pip
$ pip install seamusIt can also be installed from source:
$ git clone https://github.com/nerandell/seamus
$ cd seamus && python setup.py installHere is a simple example :
from seamus import Seamus
class SeamusExample:
def add_numbers(self, num1, num2):
seamus = Seamus()
seamus.use(self._original_func, num1, num2)
seamus.test(self._refactored_func, num1, num2)
result = seamus.run()
return result
def _original_func(self, num1, num2):
return num1 + 2 * num2 - num2
def _refactored_func(self, num1, num2):
return num1 + num2It is that simple. The result returned is always the actual result. However the refactored code is run by seamus in background and variety of inferences can be made on the basis of information obtained.
However there is still a lot of boilerplate code here. To make things simpler, you can also use the decorator:
from seamus import seamus
class SeamusExample:
def add_numbers(self, num1, num2):
result = self._original_func(num1, num2)
return result
@seamus(refactored_func=_refactored_func)
def _original_func(self, num1, num2):
return num1 + 2 * num2 - num2
def _refactored_func(self, num1, num2):
return num1 + num2This makes the code concise. However currenly, seamus doesn't do a lot with the result. To make some use of the result, you have to override the publish method provided by Seamus class.
from seamus import Seamus
class ExtendedSeamus(Seamus):
def publish(self, is_equal):
print(is_equal)
class SeamusExample:
def add_numbers(self, num1, num2):
seamus = ExtendedSeamus()
seamus.use(self._original_func, num1, num2)
seamus.test(self._refactored_func, num1, num2)
result = seamus.run()
return result
def _original_func(self, num1, num2):
return num1 + 2 * num2 - num2
def _refactored_func(self, num1, num2):
return num1 + num2You can also use the decorator with factory argument. factory can be just about
any callable which returns an extended Seamus class or atleast quack like Seamus class (at your own risk).
from seamus import Seamus
class ExtendedSeamus(Seamus):
def publish(self, is_equal):
print(is_equal)
class SeamusExample:
def add_numbers(self, num1, num2):
result = self._original_func(num1, num2)
return result
@seamus(refactored_func=_refactored_func, factory=ExtendedSeamus)
def _original_func(self, num1, num2):
return num1 + 2 * num2 - num2
def _refactored_func(self, num1, num2):
return num1 + num2By default, to compare the result returned by actual and the refactored function, seamus uses == operator.
However you can easily override it by passing comparator as an argument and it can return a true or a false value based on your own logic.
from seamus import Seamus
class SeamusExample:
def add_numbers(self, num1, num2):
result = self._original_func(num1, num2)
return result
@seamus(refactored_func=_refactored_func, comparator=lambda x, y: x + 1 == y)
def _original_func(self, num1, num2):
return num1 + 2 * num2 - num2
def _refactored_func(self, num1, num2):
return num1 + num2By default, both the functions are run everytime, but what if you have to run the refactored version only a few times? You can do that by providing a strategy.
from seamus import Seamus
class SeamusExample:
def add_numbers(self, num1, num2):
result = self._original_func(num1, num2)
return result
@seamus(refactored_func=_refactored_func, run_strategy=lambda: random() > 0.5)
def _original_func(self, num1, num2):
return num1 + 2 * num2 - num2
def _refactored_func(self, num1, num2):
return num1 + num2In the above exmaple, the refactored version will only run 50% of the time.
Documentation is available here
seamus is offered under the MIT license.
The latest developer version is available in a github repository: https://github.com/nerandell/seamus
The name is derived from Pink Floyd's song "Seamus" from the 1971 album Meddle.