<\/span># file: fact.c<\/span>\n\nlong<\/span> <\/span>fact<\/span>(<\/span>long<\/span> <\/span>n<\/span>)<\/span> <\/span>{<\/span><\/span>\n <\/span>if<\/span> <\/span>(<\/span>n<\/span> <\/span><<\/span> <\/span>1<\/span>)<\/span><\/span>\n <\/span>return<\/span> <\/span>1<\/span>;<\/span><\/span>\n <\/span>return<\/span> <\/span>n<\/span> <\/span>*<\/span> <\/span>fact<\/span>(<\/span>n<\/span> <\/span>-<\/span> <\/span>1<\/span>);<\/span><\/span>\n}<\/span><\/span>\n<\/pre><\/div>\n\n\nwhich we want to compile as a shared object, e.g. with<\/p>\n
<\/span>gcc -shared -o fact.so fact.c\n<\/pre><\/div>\n\n\nHow do we test the fact<\/code> function from Python? Easy peasy!<\/p>\n<\/span># file: fact.py<\/span>\n\nfrom<\/span> ctypes<\/span> import<\/span> CDLL<\/span>\n\nlibfact<\/span> =<\/span> CDLL<\/span>(<\/span>".\/fact.so"<\/span>)<\/span>\n\nassert<\/span> libfact<\/span>.<\/span>fact<\/span>(<\/span>6<\/span>)<\/span> ==<\/span> 720<\/span>\nassert<\/span> libfact<\/span>.<\/span>fact<\/span>(<\/span>0<\/span>)<\/span> ==<\/span> 1<\/span>\nassert<\/span> libfact<\/span>.<\/span>fact<\/span>(<\/span>-<\/span>42<\/span>)<\/span> ==<\/span> 1<\/span>\n<\/pre><\/div>\n\n\nAssuming we are in the directory where both fact.so<\/code> and fact.py<\/code> reside, we\ncan test the fact<\/code> function inside fact.c<\/code> simply with<\/p>\n<\/span>python3 fact.py\n<\/pre><\/div>\n\n\nIf the test succeeds, the script's return code will be 0.<\/p>\n
\nCongratulations! You have now tested some C code with Python! \ud83c\udf89<\/p>\n<\/blockquote>\n
Enter pytest<\/h1>\n
We are not here to just play around with bare assert<\/code>s. I promised you the full\npower of Python and pytest<\/code>, so we can't settle with just this simple\nexample. Let's add pytest<\/code> to our test dependencies and do this instead<\/p>\n<\/span># file: test_fact.py<\/span>\n\nfrom<\/span> ctypes<\/span> import<\/span> CDLL<\/span>\n\nimport<\/span> pytest<\/span>\n\n@pytest<\/span>.<\/span>fixture<\/span>\ndef<\/span> libfact<\/span>():<\/span>\n yield<\/span> CDLL<\/span>(<\/span>".\/fact.so"<\/span>)<\/span>\n\n\ndef<\/span> test_fact<\/span>(<\/span>libfact<\/span>):<\/span>\n assert<\/span> libfact<\/span>.<\/span>fact<\/span>(<\/span>6<\/span>)<\/span> ==<\/span> 720<\/span>\n assert<\/span> libfact<\/span>.<\/span>fact<\/span>(<\/span>0<\/span>)<\/span> ==<\/span> 1<\/span>\n assert<\/span> libfact<\/span>.<\/span>fact<\/span>(<\/span>-<\/span>42<\/span>)<\/span> ==<\/span> 1<\/span>\n<\/pre><\/div>\n\n\nNow run pytest<\/code> to get<\/p>\n<\/span>$ pytest\n=========================== test session starts ============================\nplatform linux -- Python 3.10.2, pytest-7.0.0, pluggy-1.0.0\nrootdir: \/tmp\ncollected 1 item\n\ntest_fact.py . [100%]\n\n============================ 1 passed in 0.00s =============================\n<\/pre><\/div>\n\n\nThat's some more informative output than what a plain Python test script would\ngive us! How about starting to leverage some of the other pytest<\/code> features,\nlike parametrised tests? Let's rewrite our test case like so<\/p>\n<\/span># file: test_fact.py<\/span>\n\nfrom<\/span> ctypes<\/span> import<\/span> CDLL<\/span>\n\nimport<\/span> pytest<\/span>\n\n\n@pytest<\/span>.<\/span>fixture<\/span>\ndef<\/span> libfact<\/span>():<\/span>\n yield<\/span> CDLL<\/span>(<\/span>".\/fact.so"<\/span>)<\/span>\n\n\n@pytest<\/span>.<\/span>mark<\/span>.<\/span>parametrize<\/span>(<\/span>"n,e"<\/span>,<\/span> [(<\/span>6<\/span>,<\/span> 720<\/span>),<\/span> (<\/span>0<\/span>,<\/span> 1<\/span>),<\/span> (<\/span>-<\/span>42<\/span>,<\/span> 1<\/span>)])<\/span>\ndef<\/span> test_fact<\/span>(<\/span>libfact<\/span>,<\/span> n<\/span>,<\/span> e<\/span>):<\/span>\n assert<\/span> libfact<\/span>.<\/span>fact<\/span>(<\/span>n<\/span>)<\/span>