You never know…

Small update to my previous post (link) about pytest. With help of Oleg Eterevsky I found there is a better way to have tests in separate directory, without manging import path.

Answer is simple: Just add __init__.py to the tests directory.

Directory structure

/app/
/app/app
/app/app/mymodule.py
/app/tests
/app/tests/__init_.py <- zero byte python magic
/app/tests/test_mymodule.py

After that tests in the ‘tests’ directory can use normal import:

from app import mymodule

It works like magic. Still, to have proper code coverage reports I need to specify filename to track. This done in the old way:

#!/usr/bin/python
import os
import inspect
import sys
import pytest
@pytest.fixture
def config():
    from dibctl import config
    return config
def test_fake(config):
    assert config
if __name__ == "__main__":
    ourfilename = os.path.abspath(inspect.getfile(inspect.currentframe()))
    currentdir = os.path.dirname(ourfilename)
    parentdir = os.path.dirname(currentdir)
    file_to_test = os.path.join(
        parentdir,
        os.path.basename(parentdir),
        os.path.basename(ourfilename).replace("test_", '')
    )
    pytest.main([
     "-vv",
     "--cov", file_to_test,
     "--cov-report", "term-missing"
     ] + sys.argv)

But this is much nicer than original version, because all than noise around paths related only to coverage reporting, not the tests itself.