Doctest – Does it Work ?

If you already know doctest, you may skip this section.

Testing this module with doctest is pretty straightforward: create a test.py file with a function test_primes with no implementation but a doctest for the primes function:

def test_primes():
    """
    >>> from primes import primes
    >>> n = 50
    >>> primes(n)
    [2, 3, 5, 7, 11, 13, 17, 19, 23, 29, 31, 37, 41, 43, 47]
    """

You are pretty confident that if test_primes doctest works as advertised in this doctest, the primes function behaves correctly. Therefore, you may now test sieve against primes:

def test_sieve():
    """
    >>> from primes import primes, sieve
    >>> n = 10000
    >>> primes(n) == sieve(n)
    True
    """

Finally, add the following boilerplate a the end of your file:

if __name__ == "__main__":
    import doctest
    doctest.testmod()

You are ready to execute this test suite with:

$ python test.py

No error displayed ? The primes module has successfully passed all tests !

Docbench – Is is Fast ?

Create a new file name benchmark.py. Add functions that act as docbench holders for the function primes and sieve. We rely on the convention that only the time spent in the last statement of any docbench will be measured, the previous ones are considered setup code.

def benchmark_primes():
    """
    >>> from primes import primes
    >>> n = 10000
    >>> primes(n)
    """

def benchmark_sieve():
    """
    >>> from primes import sieve
    >>> n = 10000
    >>> sieve(n)
    """

Add the following boilerplate at the end of your file:

if __name__ == "__main__":
    import docbench
    docbench.benchmod()

Run your benchmark with:

$ python benchmark.py

You should end up with an output similar too:

Benchmark                  Time
-------------------------  ----------------
__main__.benchmark_primes  1.03
__main__.benchmark_sieve   0.00876

Indeed, sieve is quite faster than primes !