I have a main() function in python that gets command line arguments. Is there a way for me to write pytest tests for this function and define the arguments in the code?
e.g.
def main():
# argparse code
args, other = arg_parser.parse_known_args()
return args.first_arg
def test_main():
res = main() # call with first_arg = "abc"
assert(res == "abc")
To add to the previous answers, instead of modifying sys.argv It is safer to use a context manager which can cover up and protect the underlying object. An example would be
with unittest.mock.patch('sys.argv', ['program_name', '--option1', 'inputFile']):
main()
This works only with python3. For python2 the Mock library does the trick.
I found this solution on a different stackoverflow post here.
sys.argv gets ignored by argparse. So something like ['junk', '--option1', 'inputFile'].parse_args takes a argv parameter. The docs uses this repeatedly in it's examples
parser = argparse.ArgumentParser()
parser.add_argument('--foo', action='store_true')
parser.add_argument('bar')
parser.parse_known_args(['--foo', '--badger', 'BAR', 'spam'])
where the string list replicates sys.argv[1:] that it would get from the commandline. If the argument is None (or omitted) the parser uses sys.argv[1:].
So if
def main(argv=None):
# argparse code
args, other = arg_parser.parse_known_args(argv)
return args.first_arg
You could test with
main(['foo', '-f','v'])
The unittesting file for argparse.py uses both this approach, and your's of modifying sys.argv directly.
https://docs.python.org/3/library/argparse.html#beyond-sys-argv
https://docs.python.org/3/library/argparse.html#partial-parsing
The best solution I found so far is this
def test_main():
sys.argv = ["some_name", "abc"]
res = main()
and for flags:
sys.argv.append("-f")
sys.argv.append("v")
For an entrypoint.py this might be particularly simple to add an extra argument. In this case --disable-warnings
import sys
import pytest
if __name__ == "__main__":
exit_code = pytest.main(sys.argv[1:] + ["--disable-warnings"])
