Description
In this collection, We will cover these two: unittest - part of the Python library, similar to JUnit 3
DocTest - test by example, part of the Python library
And Other testing frameworks: Py.Test - very simple "assert" syntax.
– can also run unittest style tests
Mock objects - create "fake" external components
Study Set Content:
Unit Testing in Python
James Brucker
Python Testing Frameworks
We will cover these two:
unittest
- part of the Python library, similar to JUnit 3
DocTest
- test by example, part of the Python library
Other testing frameworks:
Py.Test
- very simple "assert" syntax.
–
can also run unittest style tests
Mock objects
- create "fake" external components
https://wiki.python.org/moin/PythonTestingToolsTaxonomy
unittest example
import unittest
class
TestBuiltins
(
unittest.TestCase
):
"""Test some python built-in methods"""
def
test
_len
(self):
self.
assertEqual
(5, len("hello"))
self.
assertEqual
(3, len(['a','b','c']))
# edge case
self.
assertEqual
(0, len(""))
def
test
_str_upper
(self):
self.
assertTrue
( "ABC".isupper() )
self.
assertFalse
( "ABc".isupper() )
s = ""
# edge case
self.assertFalse( s.isupper() )
class extends
TestCase
Run tests from the command line
cmd>
python -m unittest test_module
cmd>
python -m unittest module.TestClass
cmd>
python -m unittest tests/test_module.py
Run all tests or just specific test. Three ways:
Other Ways to Run Tests
import unittest
class
TestBuiltins
(
unittest.TestCase
):
"""Test some python built-in method"""
def
test_len
(self):
self.assertEqual(5, len("hello"))
self.assertEqual(3, len(['a','b','c']))
if __name__ == "__main__":
unittest.main()
1. Let the IDE run them for you.
2. Use a test script or build tool.
3. Add a "main" script to end of your Test class...
Exercise: Try it Yourself
Test math.sqrt() and math.pow().
import unittest
import math
class
MathTest
(
unittest.TestCase
):
def
test_sqrt
(self):
self.assertEqual(5, math.sqrt(25))
self.assertEqual(0, math.sqrt(0))
#edge case
def
test_pow
(self):
#
TODO
Write some tests of
math.pow(x,n)
Exercise: Run Your Tests
Run on the command line:
cmd>
python -m unittest test_math
..
----------------------------------------------------
Ran 2 tests in 0.001s
Run with verbose (-v) output
cmd>
python -m unittest -v test_math.py
test_sqrt (test_math.MathTest) ... ok
test_pow (test_math.MathTest) ... ok
----------------------------------------------------
Ran 2 tests in 0.001s
Write two
Failing
Tests
import unittest
import math
class
MathTest
(
unittest.TestCase
):
# This answer is WRONG. Test should fail.
def
test_wrong_sqrt
(self):
self.assertEqual(1, math.sqrt(25))
# sqrt of negative number is not allowed.
def
test_sqrt_of_negative
(self):
self.assertEqual(4, math.sqrt(-16))
Exercise: Run the Tests
Run on the command line:
cmd>
python -m unittest math_test.py
..EF
===================================================
ERROR: test_sqrt_of_negative (math_test.MathTest)
---------------------------------------------------
Traceback (most recent call last):
File "test_math.py", line 10, in test_sqrt_negative
self.assertEqual(4, math.sqrt(-16))
ValueError
: math domain error
===================================================
FAIL: test_wrong_sqrt (test_math.MathTest)
Trackback (most recent call last):
AssertionError
: 1 != 5.0
Test Results
The test summary prints:
Ran 4 tests in 0.001s
FAILED (
failures
=1,
errors
=1)
How are "
failure
" and "
error
" different?
Failure means __________________________
Error means ____________________________
Tests Outcomes
Success
: passes all "assert"
Failure
: fails an "assert" but code runs OK
Error
: error while running test, such as exception raised
What Can You assert?
assertTrue
( gcd(-3,-5) > 0 )
assertFalse
( "hello".isupper() )
assertEqual
( 2*2, 4)
assertNotEqual
( "a", "b")
assertIsNone
(a)
# test "a is None"
assertIsNotNone
(a)
# test "a is not None"
assertIn
( a, list)
# test "a in list"
assertIsInstance
(3, int)
# test isinstance(a,b)
assertListEqual
( list1, list2 )
# all elments equal
Many more!
See "unittest" in the Python Library docs.
Skip a Test or Fail a Test
import unittest
class MyTest(unittest.TestCase):
@unittest.skip("Not done yet")
def test_add_fractions(self):
pass
def test_fraction_constructor(self):
self.fail
(
"Write this test!"
)
Test for Exception
def test_sqrt_of_negative( self ):
"""sqrt of a negative number should throw
ValueError.
"""
self.
assert????
( math.sqrt(-1) )
What if your code should throw an exception?
Test for Exception
def test_sqrt_of_negative(self):
with self.assertRaises
(ValueError):
math.sqrt(-1)
assertRaises
expects a block of code to raise an
exception:
What to Name Your Tests?
1.
Test methods
begin with
test_
and use
snake_case
.
def
test_
sqrt(self)
def
test_
sqrt_of_negative(self)
2.
Test class
name
either
starts
with Test (Python style) or
ends
with "Test" (JUnit style) and uses CamelCase.
class
Test
Math(unittest.TestCase)
class Math
Test
(unittest.TestCase)
What to Name Your Tests?
3.
Test filename
should start with
test_
& use snake
case
test_
math.py
test_
list_util.py or
test_
listutil.py
Note:
if test file
ends
with _test like
math
_test
.py
then
Python's "test discovery" feature (used by Django)
won't run the tests unless you write:
python -m unittest discover
-p "*_test.py"
Exercise: Test Driven Development
Write some tests for this function
before
you write the
function body. Just return 0:
def average( lst ):
"""Return average of a list of numbers"""
return 0
Exercise: Define Test Cases
1.
Typical case
: list contains a few numbers
2.
Edge cases
: a) list with only 1 number,
b) list with many values all the same,
c) list containing some 0 values (changes average).
3.
Illegal case
: empty list
What should happen in this case??
**
**
Hint: Python has a builtin
max(list)
function.
TDD forces you to think about
what the code
should do
.
Write the Tests
File:
test_average.py
import unittest
from listutil import average
class TestAverage(unittest.TestCase):
def test_average_singleton_list(self):
self.assertEqual( 5, average([5]) )
def test_list_with_many_values(self):
# test average of many values
def test_average_of_empty_list(self):
# test that average([]) throws exception
10 minutes
