How to check if mock functions has been called?

Question

I'm writing unit tests for a simple function that writes bytes into s3:

import s3fs

def write_bytes_as_csv_to_s3(payload, bucket, key):
    fs = s3fs.S3FileSystem()
    fname = f"{bucket}/{key}"
    print(f"writing {len(payload)} bytes to {fname}")
    with fs.open(fname, "wb") as f:
        f.write(payload)
    return fname


def test_write_bytes_as_csv_to_s3(mocker):
    s3fs_mock = mocker.patch('s3fs.S3FileSystem')
    open_mock = mocker.MagicMock()
    # write_mock = mocker.MagicMock()
    # open_mock.write.return_value = write_mock
    s3fs_mock.open.invoke.return_value = open_mock
    result = write_bytes_as_csv_to_s3('awesome'.encode(), 'random', 'key')
    assert result == 'random/key'
    s3fs_mock.assert_called_once()
    open_mock.assert_called_once()
    # write_mock.assert_called_once()

How can I check if method open and write has been called once? Not sure how to set mocker to cover my case.

Answer 1

The unit-test you written above is perfect and mostly covered all the functionality of the methods which you want to test.

In pytest, there is a functionality to get the unittest coverage report which will show the lines covered by unittest.

Kindly install the pytest plugin html-report(if not installed) and execute the following document:-

py.test --cov=<filename to cover: unittest> --cov-report=html <testfile>

After that, you would likely found a html file in the current location o/r in the htmlconv / directory. And from that, you could easily figure it out about the line covered and also the percentage of the unittest test coverage.

Answer 2

The issue is understanding how each mock is created and what exactly it mocks. For example mocker.patch('s3fs.S3FileSystem') returns a mock of s3fs.S3FileSystem , not the instance returned by calling s3fs.S3FileSystem() . Then to mock with fs.open(fname, "wb") as f you need to mock what the __enter__ dunder method returns. Hopefully the following code makes the relations clear:

def test_write_bytes_as_csv_to_s3(mocker):
    # Mock of the open's context manager
    open_cm_mock = mocker.MagicMock()

    # Mock of the object returned by fs.open()
    open_mock = mocker.MagicMock()
    open_mock.__enter__.return_value = open_cm_mock

    # Mock of the new instance returned by s3fs.S3FileSystem()
    fs_mock = mocker.MagicMock()
    fs_mock.open.return_value = open_mock

    # Patching of s3fs.S3FileSystem 
    mocker.patch('s3fs.S3FileSystem').return_value = fs_mock

    # Running the tested code and making assertions
    result = write_bytes_as_csv_to_s3('awesome'.encode(), 'random', 'key')

    assert result == 'random/key'
    assert open_cm_mock.write.call_count == 1