dynaconf: [bug] Testing and mocking docs are wrong

I think the easiest way is to dump a settings file to a temporary location and load it. Something like:

@pytest.fixture
def mock_settings_path(tmpdir):
    settings_path = tmpdir.mkdir("test_package").join("settings.py")
    settings_path.write(
        textwrap.dedent(
            f"""
                SETTING_ONE = True
            """
        )
    )
    return str(settings_path) 

And you can use the fixture as follows:

def test_settings_file(mock_settings_path):
    settings = LazySettings(settings_file=mock_settings_path)
    assert SETTING_ONE == True

The beauty of this approach is that you can actually add a Mock to your mock settings file, e.g.

# in your test module
MOCK_SETTING_TWO = MagicMock()

# in your mock settings file
f"""
from {__name__} import MOCK_SETTING_TWO
SETTING_TWO = MOCK_SETTING_TWO
"""

Nice insight! @rochacbruno So actually there’s no way to mock settings? I was digging in the tests but didn’t find anything to load a json or a dict and being able to mock tests.

I wrote something like this. Not optimum but for now it works ok for my purposes. Maybe somebody finds it useful.

    from dynaconf import LazySettings
    from dynaconf.loaders.json_loader import load

    mocked_settings = LazySettings(environments=True, ENV_FOR_DYNACONF="DEFAULT")

    data = """
    {
        "default" : {
            "SETTING_ONE": false,
            "SETTING_TWO": "test"
        }
    }
    """
    load(mocked_settings, filename=data)

Thanks!

Another option would be to use Monkeypatch

def __init__(self, *args, **kwargs):
        self._loaded_files = []
+       self.current_env = kwargs.get('env', 'development')
        super(DynaconfDict, self).__init__(*args, **kwargs)

@mazzi In fact, MonkeyPatching Dynaconf settings in Pytest doesn’t work, unfortunately. Setting values don’t change that way.

We have the same issue here with additional accessors as as_dict or as_int here:

AttributeError: 'DynaconfDict' object has no attribute 'as_int'