Mocking s3 in tests properly with moto

[Ouradze]

Hi, I am using django-storages to upload some data. I would like to know if there is a proper way to mock s3 using moto in the tests. At the moment, I managed to mock the upload with boto3 only but when using it with the storage as I have no endpoint defined, it fails. I am not sure if I defined my fixtures with pytest correctly. If anyone has already tried that, I would be thankful for the help. If my question is unclear, please tell me.

Thanks in advance.

conftest.py

@pytest.fixture
def aws_credentials():
    """S3 mock client"""
    os.environ.setdefault("AWS_ACCESS_KEY_ID", "testing")
    os.environ.setdefault("AWS_SECRET_ACCESS_KEY", "testing")


@pytest.fixture
def s3_client(aws_credentials):
    """S3 mock client"""
    with mock_s3():
        conn = boto3.client("s3", region_name="us-east-1")
        yield conn

storage.py

class DatasetUploadStorage(S3Boto3Storage):
    location = settings.DATASET_UPLOAD_PATH
    object_parameters = {"ACL": "public-read"}

utils.py

def upload_file(filename, storage, file_obj):
    storage.save(filename, file_obj)
    file_url = storage.url(filename)
    return file_url

test.py

@contextmanager
def s3_setup(s3_client):
    s3_client.create_bucket(Bucket="bucket")
    yield


def test_upload_file(s3_client):
    dataset_storage = DatasetUploadStorage()
    with open("sample_dataset.csv") as fp:
        with s3_setup(s3_client):
            file_url = upload_file("dataset.csv", dataset_storage, fp)
    # some assert :)

[jschneier]

An example was merged in #911