Question

尝试在执行map_async()操作的函数周围编写一些单元测试。更具体地说，我想确认一些文件在其中一个进程发生异常时被清除。下面提供了有意图的伪代码示例。

foo.py

def write_chunk(chunk):
    ... create file from chunk
    return created_filename

class Foo:
    def write_parallel(chunks):
        filenames = set()
        try:
            pool = Pool(processes=2)
            pool.map_async(write_chunk, chunks, callback=filenames.add)
        except Exception:
            //handle exception
        finally:
            cleanup_files(filenames)

test_foo.py

@patch("foo.write_chunk")
def test_write_parallel_exception_cleanup(self, mock_write_chunk):
    def mock_side_effect(chunk):
        if "chunk_1" == chunk:
            raise Exception
        else:
            return chunk
    mock_write_chunk.side_effect = mock_side_effect

    foo = Foo()
    foo.write_parallel({"chunk_1", "chunk_2"})
    //assert "chunk_2" cleaned up and exception is thrown.

但是，当我去执行测试时，我得到以下PicklingError：PicklingError: Can't pickle <class 'mock.MagicMock'>: it's not the same object as mock.MagicMock。

如何用我自己的模拟函数执行替换映射函数所需的结果？

Answer 1

因为问题源于尝试Mock和Pickle这个函数，我决定将功能拉出到一个单独的函数，模拟该函数，同时允许原始函数被pickle。见下文：

foo.py

def write_chunk(chunk):
    return write_chunk_wrapped(chunk)

def write_chunk_wrapped(chunk)
    ... create file from chunk
    return created_filename

class Foo:
    def write_parallel(chunks):
        filenames = set()
        try:
            pool = Pool(processes=2)
            pool.map_async(write_chunk, chunks, callback=filenames.add)
        except Exception:
            //handle exception
        finally:
            cleanup_files(filenames)

test_foo.py

@patch("foo.write_chunk_wrapped")
def test_write_parallel_exception_cleanup(self, mock_write_chunk_wrapped):
    def mock_side_effect(chunk):
        if "chunk_1" == chunk:
            raise Exception
        else:
            return chunk
    mock_write_chunk_wrapped.side_effect = mock_side_effect

    foo = Foo()
    foo.write_parallel({"chunk_1", "chunk_2"})
    //assert "chunk_2" cleaned up and exception is thrown.

模拟map_async函数参数产生PicklingError

1 个答案: