Up To Schedule - Back To Make Incremental Changes II - Forward To Make Changes from Anywhere (GitHub)
Based on materials by Katy Huff, Rachel Slaybaugh, and Anthony Scopatz
Software testing is a process by which one or more expected behaviors and results from a piece of software are exercised and confirmed. Well chosen tests will confirm expected code behavior for the extreme boundaries of the input domains, output ranges, parametric combinations, and other behavioral edge cases.
Unless you write flawless, bug-free, perfectly accurate, fully precise, and predictable code every time, you must test your code in order to trust it enough to answer in the affirmative to at least a few of the following questions:
- Does your code work?
- Always?
- Does it do what you think it does? (Patriot Missile Failure)
- Does it continue to work after changes are made?
- Does it continue to work after system configurations or libraries are upgraded?
- Does it respond properly for a full range of input parameters?
- What about edge and corner cases?
- What's the limit on that input parameter?
- How will it affect your publications?
Verification is the process of asking, "Have we built the software correctly?" That is, is the code bug free, precise, accurate, and repeatable?
Validation is the process of asking, "Have we built the right software?" That is, is the code designed in such a way as to produce the answers we are interested in, data we want, etc.
Uncertainty Quantification is the process of asking, "Given that our algorithm may not be deterministic, was our execution within acceptable error bounds?" This is particularly important for anything which uses random numbers, eg Monte Carlo methods.
Let's return the the simplestats module that we began earlier:
cd ~/simplestats
It has a file stats.py with the following averaging function:
def mean(vals):
"""Calculate the arithmetic mean of a list of numbers in vals"""
total = sum(vals)
length = len(vals)
return total/lengthMake a branch in this repository for adding tests:
git checkout -b add_tests
or
git branch add_tests
git checkout add_tests
The simplest way to add a test is to add a function that calls this function with arguments for which we already know the answer.
def mean(vals):
"""Calculate the arithmetic mean of a list of numbers in vals"""
total = sum(vals)
length = len(vals)
return total/length
def test_mean():
"""Test some standard behavior of the mean() function."""
assert(mean([2, 4]) == 3)The assert command will make your program stop if the condition is not true
and is common when writing tests.
You can try this test in iPython:
In [1]: import stats as s
In [2]: s.test_mean()
Practice using git: Commit this addition to the repository
git add stats.py
git commit -m "Added the first test"
Let's add one more test:
def test_float_mean():
"""Test some standard behavior when the result is not an integer."""
assert(mean([1, 2]) == 1.5)and try it with:
In [4]: reload(s)
In [5]: s.test_float_mean()
The newest test fails, but we don't get much explanation of why it fails.
Practice using git: Commit this change to the repository
git add stats.py
git commit -m "Added a floating point test, but it fails"
It is more common to place tests in a different file so that they don't
clutter the module that does the real work. Let's move our tests to a new
file called test_stats.py. Now, our stats.py file contains only:
def mean(vals):
"""Calculate the arithmetic mean of a list of numbers in vals"""
total = sum(vals)
length = len(vals)
return total/lengthand our test_stats.py file contains:
from stats import mean
def test_mean():
"""Test some standard behavior of the mean() function."""
assert(mean([2, 4]) == 3)
def test_float_mean():
"""Test some standard behavior when the result is not an integer."""
assert(mean([1, 2]) == 1.5)To make it even easier to test, we can add some lines at the bottom of
test_stats.py to run each of our tests:
test_mean()
test_float_mean()and then run this from the command-line:
python test_stats.py
The same tests pass and fail, but still not much explanation.
Practice using git: Commit this change to the repository
git add stats.py test_stats.py
git commit -m "Moved tests to a separate file to declutter module."
We could start adding some lines to give us more information about each test and why it might fail, but that could get tedious as we write basically the same things over and over for each test. Since we don't want to repeat ourselves, we might write some functions to keep track of the expected result, and report when it doesn't match the observed results. However, that seems like something that many people need, so maybe someone else did that already, and we don't want to repeat others, either.
The testing framework we'll discuss today is called nose. However, there are
several other testing frameworks available in most language. Most notably there
is JUnit in Java which can arguably attributed to
inventing the testing framework. Google also provides a test
framework for C++ applications (note, there's
also CTest). There
is at least one testing framework for R:
testthat.
Nose tests are files that begin with Test-, Test_, test-, or
test_. Specifically, these satisfy the testMatch regular expression
[Tt]est[-_]. (You can also teach nose to find tests by declaring them
in the unittest.TestCase subclasses that you create in your code. You
can also create test functions which are not unittest.TestCase
subclasses if they are named with the configured testMatch regular
expression.)
To write a nose test, we use the same assertions as we've used so far.
assert should_be_true()
assert not should_not_be_true()Since the nose package finds the tests and runs them automatically, we don't
need to include lines that run the tests. We can remove the lines that call
the tests and just use our existing test_stat.py file like this:
nosetests test_stat.py
Or from IPython:
In [1]: !nosetests test_stats.py
We get a little more information, but still not that helpful.
Practice using git: Commit this change to the repository
git add test_stats.py
git commit -m "Introduced nose testing"
However, nose itself defines number of convenient assert functions which can be used to test more specific aspects of the code base.
from nose.tools import assert_equal, assert_almost_equal, assert_true, \
assert_false, assert_raises, assert_is_instance
assert_equal(a, b)
assert_almost_equal(a, b)
assert_true(a)
assert_false(a)
assert_raises(exception, func, *args, **kwargs)
assert_is_instance(a, b)
# and many more!Be careful with division in Python, there's a gotcha!
In [1]: 3 / 2
Out[1]: 1
In [2]: 3.0 / 2
Out[2]: 1.5
You can explicitly change the type of an object or variable in Python:
In [1]: float(3)
Out[1]: 3.0
In [2]: num = 4
In [3]: float(num)
Out[3]: 4.0
- Change the two tests to use
assert_equaland run with nosetests. - Fix the
mean()function to resolve the test
Notice how much useful information you get from nose tests:
- some .... to indicate progress
- details about the failed test including the values that were not equal
- the total number of tests that were completed
- the time it took to run those tests
Practice using git: Commit this change to the repository
git add test_stats.py
git commit -m "Using nose tools to get even better output"
One of the challenges of testing is to determine what the edge cases might be. Here are some cases that we might try for the mean function
- different lengths of lists:
- [2, 4, 6]
- [2, 4, 6, 8]
- negative numbers:
- [-4, -2]
- [-2, 2, 4]
- floating point numbers:
- [2.0, 4.0, 6.0]
- [2.5, 4.5, 6.0]
- Add tests for each of by adding either new functions or new lines to the existing function. Note: some of them may fail!
- What is necessary to fix the failing tests? Try using
assert_almost_equaland make it pass the test.
Practice using git: Commit this change to the repository
git add test_stats.py
git commit -m "Added many more tests"
What happens if someone tries to use this function with strings?
mean(['hello','world'])Some mistakes don't just give a wrong answer, but fail to even finish. Python provides a mechanism to deal with this called exceptions.
In this example, python automatically raises a TypeError exception when we try to take the sum of a string.
In this case, we can add a test for the expected behavior: raising a TypeError
exception, by using the nose tool assert_raises:
def test_string_mean():
assert_raises(TypeError, mean, ['hello','world'])Practice using git: Commit this change to the repository
git add test_stats.py
git commit -m "Added a test for exceptions when passing in non-numeric results."
We can provide some extra information to the user by catching the TypeError exception:
def mean(vals):
"""Calculate the arithmetic mean of a list of numbers in vals"""
try:
total = sum(vals)
length = len(vals)
except TypeError:
raise TypeError("The list contains non-numeric elements")
return total/lengthPractice using git: Commit this change to the repository
git add stats.py
git commit -m "Added extra error message for TypeError."
The three right answers are:
- ALWAYS!
- EARLY!
- OFTEN!
The longer answer is that testing either before or after your software is written will improve your code, but testing after your program is used for something important is too late.
If we have a robust set of tests, we can run them before adding something new and after adding something new. If the tests give the same results (as appropriate), we can have some assurance that we didn't wreck anything. The same idea applies to making changes in your system configuration, updating support codes, etc.
Another important feature of testing is that it helps you remember what all the parts of your code do. If you are working on a large project over three years and you end up with 200 classes, it may be hard to remember what the widget class does in detail. If you have a test that checks all of the widget's functionality, you can look at the test to remember what it's supposed to do.
In a collaborative coding environment, where many developers contribute to the same code base, developers should be responsible individually for testing the functions they create and collectively for testing the code as a whole.
Professionals often test their code, and take pride in test coverage, the percent of their functions that they feel confident are comprehensively tested.
Test driven development (TDD) is a philosophy whereby the developer creates code by writing the tests first. That is to say you write the tests before writing the associated code!
This is an iterative process whereby you write a test then write the minimum amount code to make the test pass. If a new feature is needed, another test is written and the code is expanded to meet this new use case. This continues until the code does what is needed.
TDD operates on the YAGNI principle (You Ain't Gonna Need It). People who diligently follow TDD swear by its effectiveness. This development style was put forth most strongly by Kent Beck in 2002.
Say you want to write a std() function which computes the
Standard Deviation. You
would - of course - start by writing the test, possibly testing a single set
of numbers, by adding this to test_stats.py:
from nose.tools import assert_equal, assert_almost_equal
def test_std1():
obs = std([0.0, 2.0])
exp = 1.0
assert_equal(obs, exp)You would then go ahead and write the actual function:
def std(vals):
# you snarky so-and-so
return 1.0Practice using git: Since this works, commit the change to the repository
git add stats.py test_stats.py
git commit -m "Added a test for std() and then a function that passes the test."
And that is it, right?! Well, not quite. This implementation fails for most other values. Adding tests we see that:
def test_std1():
obs = std([0.0, 2.0])
exp = 1.0
assert_equal(obs, exp)
def test_std2():
obs = std([])
exp = 0.0
assert_equal(obs, exp)
def test_std3():
obs = std([0.0, 4.0])
exp = 2.0
assert_equal(obs, exp)These extra tests now require that we bother to implement at least a slightly more reasonable function:
def std(vals):
# a little better
if len(vals) == 0:
return 0.0
return vals[-1] / 2.0Practice using git: Since this works again, commit the change to the repository
git add stats.py test_stats.py
git commit -m "Added more tests for std() and updated function so that is passes all tests."
However, this function still fails whenever vals has more than two elements or the first element is not zero. Time for more tests!
def test_std1():
obs = std([0.0, 2.0])
exp = 1.0
assert_equal(obs, exp)
def test_std2():
obs = std([])
exp = 0.0
assert_equal(obs, exp)
def test_std3():
obs = std([0.0, 4.0])
exp = 2.0
assert_equal(obs, exp)
def test_std4():
obs = std([1.0, 3.0])
exp = 1.0
assert_equal(obs, exp)
def test_std5():
obs = std([1.0, 1.0, 1.0])
exp = 0.0
assert_equal(obs, exp)At this point, we had better go ahead and try do the right thing...
def std(vals):
# finally, some math
n = len(vals)
if n == 0:
return 0.0
mu = sum(vals) / n
var = 0.0
for val in vals:
var = var + (val - mu)**2
return (var / n)**0.5Practice using git: Since this works again, commit the change to the repository
git add stats.py test_stats.py
git commit -m "Added more tests for std() and updated function so that is passes all tests."
Here it becomes very tempting to take an extended coffee break or possibly a power lunch. But then you remember those pesky infinite values! Perhaps the right thing to do here is to just be undefined. Infinity in Python may be represented by any literal float greater than or equal to 1e309.
def test_std1():
obs = std([0.0, 2.0])
exp = 1.0
assert_equal(obs, exp)
def test_std2():
obs = std([])
exp = 0.0
assert_equal(obs, exp)
def test_std3():
obs = std([0.0, 4.0])
exp = 2.0
assert_equal(obs, exp)
def test_std4():
obs = std([1.0, 3.0])
exp = 1.0
assert_equal(obs, exp)
def test_std5():
obs = std([1.0, 1.0, 1.0])
exp = 0.0
assert_equal(obs, exp)
def test_std6():
obs = std([1e500])
exp = NotImplemented
assert_equal(obs, exp)
def test_std7():
obs = std([0.0, 1e4242])
exp = NotImplemented
assert_equal(obs, exp)This means that it is time to add the appropriate case to the function itself:
def std(vals):
# sequence and you shall find
n = len(vals)
if n == 0:
return 0.0
mu = sum(vals) / n
if mu == 1e500:
return NotImplemented
var = 0.0
for val in vals:
var = var + (val - mu)**2
return (var / n)**0.5Practice using git: Since this works again, commit the change to the repository
git add stats.py test_stats.py
git commit -m "Added tests for infinity in std() and updated function so that is passes all tests."
Can you think of other tests to make for the std() function? I promise there
are at least two.
Implement one new test in test_stats.py, run nosetests, and if it fails, implement a more robust function for that case.
And thus - finally - we have a robust function together with working tests!
Try your new test-driven development chops by implementing the var()
function, noting that the variance is the square of the standard devation.
The type of tests that are written is determined by the testing framework you adopt. Don't worry, there are a lot of choices.
Exceptions: Exceptions can be thought of as type of runtime test. They alert the user to exceptional behavior in the code. Often, exceptions are related to functions that depend on input that is unknown at compile time. Checks that occur within the code to handle exceptional behavior that results from this type of input are called Exceptions.
Unit Tests: Unit tests are a type of test which test the fundamental units of a program's functionality. Often, this is on the class or function level of detail. However what defines a code unit is not formally defined.
To test functions and classes, the interfaces (API) - rather than the implementation - should be tested. Treating the implementation as a black box, we can probe the expected behavior with boundary cases for the inputs.
System Tests: System level tests are intended to test the code as a whole. As opposed to unit tests, system tests ask for the behavior as a whole. This sort of testing involves comparison with other validated codes, analytical solutions, etc.
Regression Tests: A regression test ensures that new code does change anything. If you change the default answer, for example, or add a new question, you'll need to make sure that missing entries are still found and fixed.
Integration Tests: Integration tests query the ability of the code to integrate well with the system configuration and third party libraries and modules. This type of test is essential for codes that depend on libraries which might be updated independently of your code or when your code might be used by a number of users who may have various versions of libraries.
Test Suites: Putting a series of unit tests into a collection of modules creates a test suite. Typically the suite as a whole is executed (rather than each test individually) when verifying that the code base still functions after changes have been made.
Behavior: The behavior you want to test. For example, you might want to test the fun() function.
Expected Result: This might be a single number, a range of numbers, a new fully defined object, a system state, an exception, etc. When we run the fun() function, we expect to generate some fun. If we don't generate any fun, the fun() function should fail its test. Alternatively, if it does create some fun, the fun() function should pass this test. The expected result should known a priori. For numerical functions, this is result is ideally analytically determined even if the function being tested isn't.
Assertions: Require that some conditional be true. If the conditional is false, the test fails.
Fixtures: Sometimes you have to do some legwork to create the objects that are necessary to run one or many tests. These objects are called fixtures as they are not really part of the test themselves but rather involve getting the computer into the appropriate state.
For example, since fun varies a lot between people, the fun() function is a method of the Person class. In order to check the fun function, then, we need to create an appropriate Person object on which to run fun().
Setup and teardown: Creating fixtures is often done in a call to a setup function. Deleting them and other cleanup is done in a teardown function.
The Big Picture: Putting all this together, the testing algorithm is often:
setup()
test()
teardown()But, sometimes it's the case that your tests change the fixtures. If so, it's better for the setup() and teardown() functions to occur on either side of each test. In that case, the testing algorithm should be:
setup()
test1()
teardown()
setup()
test2()
teardown()
setup()
test3()
teardown()Up To Schedule - Back To Make Incremental Changes II - Forward To Make Changes from Anywhere (GitHub)